United States
Environmental Protection
Agency
Office of
Emergency and
Remedial Response
EPA/ROD/R04-88/034
April 1988
Superfund
Record of Decision
Brown Wood Preserving, FL
-------
REPORT DOCUMENTATION i »• REPORT NO 2-
pAGE i EPA/ROD/R04-86/034
4. Title and Subtitle
SUPERFUND RECO.RD Or DECISION
Brown Wood Preserving, ?L
"irsc Remeaial Ace ion - Final
.. Author(s)
9. Performing Organization Name and Address
12. Sponsoring Organization Name and Addresi
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
3.
S.
6.
a.
10
11.
(C)
(G)
13
Recipient's Accession No.
Report Date
04/08/88
Performing Organization Rept. No.
Proiect/Task/Work Unit No.
Contract(C) or Grant(G) No.
Type of Report & Period Covered
800/000
14.
'15. Supplementary Notes
16. Abstract (Limit: 200 words)
Tne 55-acre Brown Wood Preserving sit-? is located approximately two miles west of the
City of Live OaK, Suwanee County, Florida. The site is located in karst terrain in
wnich sinkholes are a common geological feature. The areas surrounding the site are
considered rural and light agricultural. There are four private wells located along the
site periphery that ootain water from an aquifer 20-100 feet below the site. The public
water .supply wells for the City of Live Oak are located less than two miles away. The
ite contains a former wood preserving plant facility, which pressure treated timber
•products with creosote and some pentacnloropnenol (PC?) for thirty years between
1948 and iy7d. During this time, several different companies operated the facility. In
addition, the facility was rebui.lt following a fire in February 1974. Sludge and
contaminated soils have been identified tn the immediate vicinity of the plant site and
an upgradient lagoon. This three-acre lagoon drains approximately 74-acres and contains
water provided above approximately 3,000 yd^ of creosote sludge and contaminated
soil. In addition, small amounts of solidified creosote and PCP are contained in onsite
storage tanks and retorts. In 193.1, EPA was notified by one of the former facility
owners that hazardous waste may have been handled at the site. In July 1932, the
Florida Department of Environmental Regulation (FDSR) inspected the site and detected a
(See Attached Sheet)
17. Document Analysis a. Descriptors
Record of Decision
Brown Wood Preserving, FL
First Remedial Action - Final
Contaminated Media: sediments, sludge, soil, sw
Key Contaminants: organics (creosote, PAHs)
b. "Identlfiers/Open-Ended Terms
c. COSATI Field/Group
^ailability Statement
19. Security Class (This Report)
None
20. Security Class (This Page)
None
21. No. of Pages
347
22. Price
4NSI-Z39.18)
See Instructions on Reverse
OPTIONAL FORM 272 (4-77)
(Formerly NTIS-35)
Department of Commerce
-------
EPA/ROO/R04-88/034
Brown Wood Preserving, ?L
F i rst Re.lie dial Action -Final
15. ABSTRACT (continued)
number of organic compounds. A action, completed in February 1988, resulted in the
removal of approximately 200,000 gallons of lagoon water and 15,000 tons of contaminated
lagoon sludge and soil. The primary contaminants of concern affecting the soil,
sediments, sludge, and waste water are creosote constituents including PAHs.
The selected remedial action for this site includes: removal and treatment, if
necessary, of lagoon water with discharge to a POTW; excavation, treatment, and offsite
disposal of approximately 1,500 tons of the most severely contaminated soil and sludge;
onsite b iodegradat ion of approximately 10,000 tons of the remaining soils i.n a 14-acre
treatment area constructed with a liner and an internal drainage and spray irrigation
system; covering of the treatment area with clean fill after b io retried lat Ion; and ground
water monitor;ng. The estimated present worth cost for this remedial action Is
$2,740,UOU.
-------
Pecord of Decision.
for
the
Brown Wood Preserving Site
Live Oak, Florida
-------
CONTENTS OF RECORD OF DECISION
I. DECLARATION
A. Site Name and Location
B. Statement of Basis and Purpose
C. Description of the Selected Remedy
D. Descriptions of Consistency with CERCLA
as amended by SARA and the NQP
II. DECISION SUMMARY
A. Site Name, Location and Description
B. Site History
C. Enforcement History
D. Ccmunity Relations History
E. Alternatives Evaluation
F. Selected Remedy Description and Rationale
for Selection
III. RESPONSIVENESS SUMMARY
A. Background on Comunity Involvement
B. Summary of Public Garments and Agency Responses
C. Explanation of Differences Between Proposed Plan
and Selected Remedy
D. Connunity Relations conducted at the Site prior
to and during the Public Comment Period
-------
ADDENDA
ACRONYMS
AMAX/Brown AMAX Environmental Services, Inc. and the James
Graham Brown Foundation, Inc.
CFRCLA/SARA The amended Superfund Law
EPA or the Agency U.S. Environmental Protection Agency
FDER or the State Florida Department'of Environmental Regulation
LDR RCRA Land Disposal Restriction Regulations ("Land Ban")
NCP The National Contingency Plan
ORD Office of Research and Development
OSWER Office of Solid Waste and Emergency Response
OTA Office of Technology Assessment
FIGURES
1.0 Geologic Map for Suwannee County, Florida page 4
2.0 Vicinity Map - Live Oak, Florida pages 7-8
3.0 Map - Facility Layout pages 9-10
TABLES
1.0 Summary of Site Investigations page 12
2.0 Summary of Technology Screening page 17
3.0 Estimated Volumes of Contaminated
Materials Prior to Removal Activity page 25
4.0 Summary of Contaminant Concentrations page 26
5.0 Chronology of Environmental Sampling page 27
6.0 Indicator Chemicals page 28
7.0 Results of Detailed Screening of
Alternatives - Cost Estimates page 31
ATTACHMENTS
A Memorandum on the Extent of Removal Activity
which occurred December 1987 through February 1988
B Risk Assessment and addenda
C State letter of concurrence
D Letter from DOI re Natural Resources Survey
E Transcription re the Public Meeting on October 9, 1987
-------
RECORD OF DECISION
REMEDIAL ALTERNATIVE SELECTION
I. DECLARATION
A. SITE NAME AND LOCATION
The Brown Wood Preserving Site is located at the intersection of
Sawmill Road and Goldkist Road, west of the City of Live Oak,
Suwannee County, Florida. The Site is situated in the northwest
quarter of Section 22, Township 2 South, Range 13 East.
B. STATEMENT OF BASIS AND PURPOSE
This decision document represents the selected remedial action for
this Site developed in accordance with the Comprehensive Environmental
Response, Compensation and Liability Act of 1980 (SARA), and to the
extent practicable, the National Contingency Plan (40 CFR, Part 300),
(NCP). I have determined that the below-mentioned description of
the Selected Remedy for the Brown Wood Preserving Site is a cost-
effective remedy and provides adeouate protection of the public
health, welfare, and the environment. These activities will be
considered part of the approved action and eligible for Trust Fund
monies.
The State of Florida has been consulted and agrees with the approved
remedy.
This Record-of-Decision is based upon the administrative record
which describes site-specific conditions and the analysis of
effectiveness and cost of the remedial alternatives for the
Brown Wood Preserving Site. The attached index identifies the
items which comprise the administrative record upon which the
selection of the remedial action is based.
C. DESCRIPTION OF THE SELECTED REMEDY
e Removal of selected contaminated materials to an EPA-approved
hazardous waste facility.
0 Onsite biodegradation of remaining, less severely
contaminated soils in a treatment area that is constructed
with a liner and an internal drainage and spray irrigation
system. After the bioremediation is complete the land
treatment area will be covered with clean fill.
- 1 -
-------
0 Operation and maintenance ( O&M ) activities will include:
- Semi-annual groundwater monitoring for five (5) years
( See Section 121(c), CERCLA/SARA.) after the
completion of the construction activities.
- Activities necessary to the proper functioning of the
bioremediation process and the attendant systems.
- Additional O&M activities may be identified during the
remedial design.
D. DECLARATIONS OF CONSISTENCY WITH CERCLA
AS AMENDED BY SAPA AND THE NCP
The selected remedy is protective of human health and the environment
attains federal and state requirements that are applicable or relevant
and appropriate, (or that a waiver can be justified for whatever
ARARs which will not be met), and is cost-effective. This remedy
satisfies the preference for treatment that reduces toxicity, mobility,
or volume as a principal element. Finally, it is determined that this
remedy utilizes permanent solutions and alternative treatment (or
resource recovery) technologies to the maximum extent practicable.
I have also determined that the action being taken is appropriate wh
balanced against the availability of Trust Fund monies for use at ot
sites. In addition, the selected remedy is more cost-effective than
other remedial actions, and is necessary to protect the public health,
welfare or the environment. All off-site disposal shall be in
compliance with the existing policies of EPA.
If additional remedial actions are determined to be necessary, a
Record-of-Decision will be prepared for approval of the future L—Jia_
action.
Date . Greer C. Tidwell
Regional Administrator
- 2 -
-------
II. DECISION SUMMARY
A. SITE NAME, LOCATION AND DESCRIPTION
The Brown Wood Preserving Site is located at the intersection of
Sawmill Road and Gold Kist Road, aroroximately two (2) miles west
of the Citv of Live Oak, Suwannee County, Florida. The Site is
situated in the northwest Quarter of Section 22, Township 2 South,
Range 13 East (See Mans 1,2, and 3). The 55-acre Site is located
in karst terrain in which sinkholes are a common qeolcoical feature.
The Site is not located in a floodplain; the topography on-site
varies in elevation from 85 feet above mean sea level (amsl) to
111 amsl. The areas surrounding the plant site is considered rural
and light agricultural. A sawmill and a construction company are
located to the east and west, respectively, of the Site. There are
four (4) private wells located along the periphery of the Site.
The public water supply wells for the City of Live Oak are located
less than two (2) miles away. Domestic water in the vicinity of the
Site is produced by means of wells into the Floridan Aouifer.
The Site is underlain by karstic terrain consisting of limestone and
dolomite. This bedrock is reported to be over 2,500 feet thick with
the more permeable deposits contained within the upper 400 feet. The
Suwannee Limestone lies from approximately 20 to 100 feet below the
Site and is the uppermost unit of the Floridan Agui^er. Portions of
the Floridan Aguifer are honey-combed with dissolution chambers and
solution pores that allow for the rapid movement of groundwater. A
peanut-shaped three-acre lagoon draining approximately 74-acres
presently contains about 3,000 cubic yards of creosote sediments/
sludge. Several sets of railroad spur tracks, assorted building
foundations, and a few old, rusted steel structures exist on-site.
R. SITE HISTORY
The Site is the location of a plant that was used to treat lumber
products bv a number of different operators. Portions of the Site
have been and continue to be under different ownership.
The plant is no longer in operation; however, the facility was used
for pressure treating timber products with creosote and some
pentachlorophenol (PCP) for thirty (30) years from 1948 through
1978. During this time, the facility was operated by several
different conpanies.
The facility was constructed in 194R by W.F. Bancker fi Son and
operated until March 1, 1952 as Suwannee Wood Preserving, Inc.
At that time, the facility was sold to the Brown Wood Preserving
Co., Inc. which operated the plant until 1965. In 1965 the Brown
Wood Preserving Company, Inc. was merged into W.P. Brown & Sons
- 3 -
-------
I i-cr* • • '
I '-^.^—*. *JW*MHK UUE JTO«
"| *UIT.4«M«^».
I v\^ ."^
HAWTHONN FORMATION
GEOLOGIC MAP AND GEM^ALIZED EAST-WEST HYDROGEOLOGIC
--' • r*r\t i MTV «~j /-\«-
-------
Umber Company, Inc. which operated the plant until 1971. in
Auaust 1971, the lumber ccmpany was dissolved and the assets taken
over by the James Graham Brown Foundation, Inc. which operated the
Plant until the facility was closed due to a fire on February 7,
1973. The property was purchased on Anril 2, 1974 by Mr. W.F.
Belote who rebuilt the plant and ran a snail number of charqes
throuah one pressure cylinder.
On December 24, 1976, Amax Forest Products, Inc. purchased the plant
and operated it intermittently until June 22, 1978 when it was bouaht
by Live Oak Timber Treaters, Inc. who may have operated the facility
for a short time thereafter. The facility is currently owned by Mr.
J. Paul Crews.
The facility is a wood preserving plant consisting of two horizontal
retorts (cylinders), a series of storage tanks, and a boiler and
associated storage yards and a wastewater laaoon. The present facility
layout is shown on Figure 3.0 and described below. The plant layout
prior to the fire in 1973 is not definitely known and may have varied
somewhat from the existing facility layout.
Creosote was stored in two (2) 50,000 gallon storage tanks immediately
east of the retorts. Located east of the southern creosote tank is a
50,000 gallon tank used for storage of pentachloroohenol (PCP). North
of the PCP tank is a mixing tank and a 100,000 gallon tank used for
petroleum storage. PCP was received at the Site in solid form, mixed
with petroleum in the mixing tank and then stored in the PCP storage
tank.
Between the petroleum storage tank and the PCP tank is a water storage
tank. Process water was obtained from a well located northwest of the
water storage tank. North of the creosote storage tank is a concrete
slab on which pumns are located. At the south end of the retorts are
an oil/water separator and an evaporator.
Steam used to heat the creosote was generated in the boiler located
on the east side of the petroleum storage tank. The boiler was
fueled by wood chips and fuel oil.
Operations at the plant are believed to have occurred as follows.
Untreated timbers were received by rail and possibly by truck, then
stored in the area north of the plant until treated. The untreated
timbers were loaded onto small rail cars and moved into the retorts
for treafanent. Treated timbers were removed from the retorts and
allowed to dry along a track area immediately north of the retorts.
After drying treated timbers were moved to the timber storage area
north of the treatment plant. Timbers were shipped to buyers by
rail and truck from the storage area. The majority of products
produced at this Site were treated with creosote; however, at some
time during the plan operation, PCP was also used.
- 5 -
-------
Plant wastewater was discharged to the oil/water senarator and from there to
the laooon via a culvert and drainane ditch. Creosote recovered from the
oil/water separator was discharged to the evaporator located south of the
seoarator where it was heated by stean to evaoorate any remaining water.
The creosote was then returned to the storage tanks, or if unusable, it was
oumned to the off-specification storage tank located north of the plant.
Sludge and contaminated soils have been identified in the immediate vicinity
of the nlant site and the lagoon. Water is ponded in the lagoon above sludge
and contaminated soils. In addition, small amounts of solidified creosote
and PCP are contained in storage tanks and retorts remaining on the Site.
- 6 -
-------
Agrlculturaf
Cohseum
KDlOCC SuMvtT
Figure 2.0
!*
- 7 -
-------
Vicinity Map
Figure 2.0
- 8 -
-------
n
r
m
1
E
i
»
i
t!
w
!
2
MM* .'
. i
A
I
V
41
M
I"
j
MAP p«fFw«tD rr
LDWt CM64NCCM IMC.
Figure 3.0
- 9 -
-------
CONCRETE-
SLAB
with PUMPS
- 10 -
Facility Layout
Figure 3.0
-------
C. ENFORCEMENT HISTORY
In 1981, followina nassaae of CERCLA, one of the former owners of a
portion of the Site notified EPA under CERCLA Section 103, that
hazardous waste may have been handled at the Site.
On July 2, 1982, the FDER inspected the Site and took samples for
analysis. A number of orqanic compounds were detected by analysis
of those samples. In December 1982 EPA completed a Hazard Ranking
System scoring of the Site and, subsequently, placed the Site on the
National Priority List.
Contractors for both the EPA and the James Graham Brown Foundation
carried out investiqations at the Site in 1983. AMAX Environmental
Services, Inc., also indenendently reviewed the Site investigations
during that time.
On September 14, 1983, the EPA issued an administrative order
requiring that an RI/FS Work Plan be prepared. In response to
the order, AMAX Environmental Services, Inc. and the James Graham
Brown Foundation (AMAX/Brown) agreed to undertake an RI/FS study.
On March 22, 1984, an RI/FS Work Plan was submitted to EPA and the
FDER by AMAX/Brown. This Work Plan was Incorporated by reference
into the EPA Consent Order (for an RI/FS) dated April 20, 1984 and
the parallel FDER Consent Order filed on April 19, 1984.
\
At a meeting with the EPA and the FDER on February 22, 1985, AMAX/
Brown was asked to prepare a revised RI/FS Work Plan along with a
new schedule for the completion of the Vfork Plan. The Work Plan
was resubmitted and the tasks therein implemented. The RI/FS was
comoleted in August 1987.
A summary of the site investigations is presented in Table 1.0.
In October 1987 a meeting between EPA and AMAX/Brown occurred.
The discussion included the introduction by EPA of the possibility
for a removal of lagoon sludges by AMAX/Brown during the winter
dry season. In November 1987 AMAX/Brown stated their intention to
do a removal of the lagoon sludges. EPA approved of AMAX/Brown's
proposed activities and began negotiating a Consent Order while the
removal developed. The Consent Order was completed in January 1988
and the removal activities were completed in February 1988. AMAX/Brown
have been very cooperative in furthering the cleanup of the site.
- 11 -
-------
Gc-pany/Agency
Table 1.0
Summary of Site Investigations
Date Activities
Florida DER
NUS Corporation
EPA Emergency
Response Team
LETCO
LETCO
FTC&H/EEM
ReTeC
FTC&H/EEM
ERT
July 2, 1982
February 9, 1983
June 1983
September 1983
November 1984
March 1986
August 1986
March, 1987
July 1987
Private well sampling/
site inspection.
Private well, surface
water and soil sampling
Surface soil/waste
sampling 1
Soil borings/ground
water/surface water
sampling
Preliminary Site
Investigation Report 2
Draft RI Report 3
Lagoon test pits/
TCLP analyses
Final RI report 4
Risk Assessment ^
P.E. LaMoreaux and Associates Inc. "A Hydrogeologic Evaluat-
ion of the Impact of Past Waste Disposal Operations at an
Abandoned Wood Preserving Plant in Live Oak, Florida and
Recommendations for Remedial Action". November 15, 1983.
Law Engineering and Testing Company "Final Report of
Investigation for the Brown Wood Preserving Site, Live Oak,
Florida". November 20, 1984.
FTC&H and Environmental Engineering Management "Report on
the Remedial Investigation, Bro'wn Wood Preserving Site, Live
Oak, Florida". (First draft) March 6, 1986.
FTC&H and Environmental Engineering Management "Report on
the Remedial Investigation, Brown Wood Preserving Site, Live
Oak, Florida". (Final)
ERT "Risk Assessment to Accompany the Feasibility Study of
the Live Oak Superfund Site, Live Oak, Florida". July 1987.
- 12 -
-------
D. COMMUNITY RELATIONS HISTORY
Residents near the Site are Generally aware that the Site was a wood
orocessina facility sometime in the oast and that it is a hazardous
waste site.
The administrative record was installed in the Site repository in
Live Oak, Florida, on September 29, 1987. A notice reqardinq
the administrative record and a future public meeting was placed in
the local newspaper on October 1, 1987. The public cement
period beqan on November 2S, 1987 and ended on December 16, 1987.
The public meeting on the RI/FS results and the presentation of the
remedy of choice took place on December 9,1987 in Live Oak, Florida.
The meeting was attended by very few local citizens.
The EPA Remedial Project Manager received no comments from the public
on the proposed remedy or on any other facet of the project. However,
reports from the FHER's local liaison and from a local newspaper
reporter indicate that the community is pleased that EPA, the FDER,
and AMAX/Brown moved so rapidly to cleanup the site. There are no
environmental activist groups involved at the site.
- 13 -
-------
E. ALTERNATIVES EVALUATION
1. ALTERNATIVE DEVELOPMENT
The National Continaencv Plan (40 CFR 300.68) specifies that remedial
alternatives should be classified as either management of migration
(off-site miaration) or source control.
Management of miaration remedial action as specified in 40 CFR
300.68(e)(3) is necessary where hazardous substances have migrated
from the original source of contamination and rose a significant
threat to public health, welfare or the environment. Management
of miaration remedial actions have been eliminated bv the feasibility
study because the Remedial Investigation concluded that the contaminants
deposited at the Site have remained in place and do not pose an
immediate danger to human health, welfare or the environment.
Source controls as- defined in 40 CFR 300.68(e)(2) address situations
in which "a substantial concentration of hazardous substances remain
at or near the area where they were originally located and inadeouate
barriers exist to retard miaration of substances into the environment."
Source control remedial actions may include alternatives to contain
the hazardous substances in place or eliminate potential contamination
by transporting the hazardous substances to a new location. Based
upon the above definition, the purpose of source control remedial
actions is to prevent or minimize the miaration of hazardous substances
from the Brown Wood Preserving Site. In order to facilitate the
development of alternatives, technologies are arranged by general
category and applicability in Table 2.0. From the above list of
technically feasible remedial action technologies, seven (7) specific
alternatives were developed for the Brown Wood Site. These alternative
are described in Table 7.0.
In addition to the above reouirements for the development of alternatives
based on technical feasibility, the U.S. EPA Guidance on Feasibility
Studies under CERCLA (June 1985) states: "At least one alternative
for each of the following must, at a minimum, be evaluated within the
reouirements of the feasibility study guidance and presented to the
decisionmaker:
(a) Alternatives for treatment or disposal at an offsite facility
approved bv EPA (including RCRA, TSCA, CWA, CAA, MPRSA, and
SDWA approved facilities), as appropriate;
(b) Alternatives which attain applicable and relevant Federal public
health or environmental standards;
(c) As appropriate, alternatives which exceed applicable and relevant
public health or environmental standards;
- 14 -
-------
(d) Alternatives which do not attain anolicable or relevant public
health or environmental standards but will reduce the likelihood
of present or future threat frcn the hazardous substances. This
must include an alternative which closelv approaches the level of
protection provided by the anolicable or relevant standards and
meets CERCLA's objective of adequately nrotectina public health,
welfare and environment.
(e) A no-action alternative."
2. ALTERNATIVE SCREENING PROCESS
The purpose of the initial screening process is to identify, develop,
and incorporate complementary mitigating technologies into site-
specific alternatives. The National Oil and Hazardous Substances
Contingency Plan ( NCP ) Section 300.68(g)(h) outlines the process for
developing and screening remedial alternatives. The NCP states "a
limited number of alternatives should be developed for either source
control or offsite remedial action (or both) depending unon the type
of response that has been identified." Furthermore, "the alternatives
developed under CFR 300.68(g), Development of Alternatives, will be
subjected to an initial screening to narrow the list of potential
remedial actions for further detailed analysis." Three broad criteria
should be used in the initial screening of alternatives: D^cost;
2) effects of the alternatives; and 3) acceptable engineering practice.
In accordance with CFR 300.68{g) and (h) and U.S. EPA Guidance on
Feasibility Studies Under CERCLA, the initial screening process of
remedial action technologies was divided into six (6) steps:
0 Identification of Remedial Action Technologies based upon
General Response Actions:
0 Development of Technological Feasibility Criteria and Screening
(acceptable engineering practice);
0 Development of Remedial Action Alternatives;
0 Development of Environmental and Public Health Criteria
and Screening (acceptable engineering practice);
0 Other Criteria Screening; and
0 Cost Estimating and Screening.
The technologies/alternatives remaining after the initial screening
process were subjected to a detailed evaluation.
- 15 -
-------
3. ALTERNATIVE TECHNOLOGIES
Several alternative technologies were studied for possible
utilization as a remedy. The seven (7) technoloaies considered were:
in-situ containment, source removal, water treatment, hioloaical
treatment, thermal treatment, solvent extraction, and disposal.
These seven (7) technologies were further specified into twenty-four
(24) separate specific technological alternatives.
a. In-situ containment technologies were eliminated from consideration
because of the risk of a partial sinkhole collapse in the lagoon and
because the method would result in no reduction in the toxicity or
volume of the contaminants.
b. Several source removal technologies were evaluated. The technologies
evaluated in this category apply to sludge and heavily contaminated
soils found in the lagoon, the discharge ditch and the plant area.
Included in this category are inflow control, water removal, and
shallow excavation. There were no significant impediments to
implementing these technologies at the Site. Source removal will '__
a necessary component of most of the remedial action alternatives
evaluated.
c. The water treatment technologies evaluated are applicable to
contaminated water removed from the lagoon as part of the source
removal action. Technologies considered included: spray irrigation,
evaporation, treatment and discharge to a surface stream, and
pretreatment and discharge to the municipal sewer. The last option
is considered the most feasible option from a technical, economic
and administrative perspective.
d. Biological treatment technologies considered included: in-situ
treatment, biological batch reactors and land treatment. In-situ
treatment is technically feasible for low level contaminated soil.
However, the majority of the wastes reguiring treatment are highly
contaminated sludges. Therefore, in-situ treatment of highly
contaminated sludges was excluded, biological batch reactors are
potentially feasible for pretreatment of the highly contaminated
sludges. The residues from this process would still reguire final
treatment and disposal. Land treatment of low-level wastes is
considered feasible and was retained as an alternative technology.
e. Thermal treatment technologies included: examination of the Huber
System, the SHIRCO incinerator, the mobile rotary kiln, the mobile
circulating fluidized bed, the commercial incinerator, and the
industrial kiln. .These technologies are applicable to the sludges
and the contaminated soils. Technologies evaluated included both
mobile incinerators (for on-site waste incineration) and stationary,
off-site commercial incineration facilities. The SHIPOO system is
considered the most technically feasible option. This system will
be evaluated further in the review of the remedial action alternati'._s.
Several industrial kilns and commercial incinerators are technically
feasible. Of these options, the Marine Shale industrial kiln in
Louisiana is considered the most economically viable and was retained
for further evaluation.
- 16 -
-------
Table 2.0
SUMMARY OF TECHNOLOGY SCREENING
TECHNOLOGY APPLICABLE
Containment Technologies
o In-situ solidification No
o Vertical Seepage Cutoffs No
o Horizontal Barriers No
Source Removal
o Water Inflow Control Yes
o Water Removal Yes
o Excavation Yes
Water Treatment Technologies
o Pretreatment and Municipal Discharge Yes
o Treatment and Surface Water Discharge No
o Evaporation No
o Spray Irrigation Yes
Biological Treatment Technologies
o Land Treatment Yes
o Batch Reactors . Yes^-
o In-situ Treatment . No2
Thermal Treatment Technologies
o Huber System . No
o SHIRCO Yes
o Mobile Rotary Kiln No
o Mobile Circulating Fluidized Bed No
o Commercial Incinerator No
o Industrial Kiln Yes
Solvent Extraction
o B.E.S.T. Process No
o Critical Fluid Extraction System No
o PCS Soil Decontamination Process No
Disposal
o On-site Vault No
o Off-site Commercial Facilities Yes
1 Only for sludges
2 Applicable to low level contaminated soils
- 17 -
-------
f. Three solvent extraction processes were considered: the R.E.S.T.
process, the critical fluid extraction system, and the PCT soil
decontamination process. These technologies are applicable to
low-level sludqes as well as highly contaminated sludges. There
is a lack of operating history with these technologies. The most
advanced system for which there is a commercially available unit is
the B.E.S.T. process. However, this alternative was eliminated due
to (1) the extensive research and development work which would be
reouired, (2) the extensive utility requirements for the unit, (3)
the need to find a use for the oils recovered from the process, and
(4) the uncertain regulatory status of the system.
g. Two (2) disposal alternatives were examined: (1) on-site vaulting
and (2) disposal at off-site commercial facilities. Permanent
on-site vaulting was eliminated due to RCRA land disposal regulations,
the Site's location in karst terrain, and Florida's ban on the siting
of hazardous waste landfills within the State. Disposal at off-site
commercial facilities was retained for further evaluation.
4. SUMMARY OF ALTERNATIVES EVALUATIONS
As described in Table 2.0 and above, seven (7) general cleanup
technologies further divided into twenty-four (24) remedial
alternatives were initially screened with the intent to reduce
the number of alternatives to be evaluated in detail. This
initial screehina process involved the use of several criteria:
1) technical feasibility; 2) public health effects; 3) environmental
effects; 4) attainment of the legally applicable or relevant and
appropriate requirements of other Federal and State public
health or environmental laws, or provides the grounds for invoking
one of the six waivers provided for in SARA; 5) site-specific
application; and 6) cost.
Section 121 (b)(l) and (b)(2) of CERCLA/SARA says:
(1) Remedial actions in which treatment which permanently and significantly
reduces the volume, toxicity or mobility of the hazardous substances,
pollutants, and contaminants is a principal element, are to be preferred
over remedial actions not involving such treatment. The offsite transpon.
and disposal of hazardous substances or contaminated materials without
such treatment should be the least favored alternative remedial action
where practicable treatment technologies are available. The President
shall conduct an assessment or permanent solutions and alternative
treatment technologies or resource recovery technologies that, in whole
or in part, will result in a permanent and significant decrease in the
toxicity, mobility, or volume of the hazardous substance, pollutant, or
contaminant. In making such assessment, the President shall specifically
address the long-term effectiveness of various alternatives. In assessing
alternative remedial actions, the President shall, at a minimum, take into
account:
- 18 -
-------
(A) the lonq-term uncertainties associated with land disposal;
(3) the qoals, objectives, and requirements of the Solid Waste
Disoosal Act;
(C) the persistence, toxicity, nobility, and propensity to
bioaccumulate of such hazardous substances and their
constituents;
(D) short- and lonq-term potential for adverse health effects
from human exposure;
(E) lonq-term maintenance costs;
(F) the potential for future remedial action costs if the
alternative remedial action in question would fail; and
(G) the potential threat to human health and the environment
associated with excavation, transportation, and redisposal,
or containment.
The President shall select a remedial action that is protective of human
health and the environment, that is cost-effective, and that utilizes
permanent solutions and alternative treatment technoloqies or resource
recovery technoloqies to the maximum extent practicable. If the President
selects a remedial action not appropriate for a preference under this
subsection, the President shall publish an explanation as to why a remedial
action involving such reductions was not selected.
(2) The president may select an alternative remedial action meetinq the
objectives of this subsection whether or not such action has been achieved
in practice at any other facility or site that has similar characteristics.
In makinq such a selection, the President may take into account the deqree
of support for such remedial action by parties interested in such site.
Of the twenty-four (24) specific alternatives, fourteen (14) were
eliminated in the initial screeninq. The remaining ten (10) were
merged into seven (?) and the seven (7) specific alternatives were
subjected to detailed screeninq procedures (Table 2.0 describes
the results of the process.). A more detailed evaluation of the
seven (7) specific cleanup alternatives follows.
a. NO-ACTION
The no-action alternative implies leavinq the site in its present
condition without disturbinq the contaminated materials. This
alternative has the advantaqe that both the short-term and the
lonq-term risks of exposure are not impacted from their present
levels. The risk of spread" of contaminated materials is not
increased. A part of the no-action alternative may be continued
monitoring of the groundwater allowing identification of any
changes in site conditions or the migration of contaminated
materials off-site. Should changes be discovered which increase
the risks associated with this site, this alternative could be
reassessed and, if necessary, alternative actions could be taken.
- 19 -
-------
b. ON-SITE INCINERATION
This alternative includes excavation of all contaminated materials
at the site followed bv on-site incineration. The steos required
for this alternative include:
1. site preparation including mobilization and establishment of
on-site support facilities;
2. inflow control including construction of engineered perms
and drainage ditches;
3. water removal from the lagoon to site storage tanks and other
water disposal system;
4. discharge to the POTW of water from site storage tanks and
other water disposal system after treatment, if necessary;
5. excavation of lagoon sludges and contaminated soils and
short haul to incinerator stagina area;
6. incineration by use of the SHIRCO incinerator on-site;
7. site restoration including reqradinq to establish engineered
drainage patterns and appropriate revegetation. •
Institutional requirements for this alternative include obtaining
local approvals for discharge to the POTW and ccnpletina a pilot
test burn to demonstrate the applicability of the technology in
meeting RCRA and State air quality performance standards.
The impacts of this alternative include elevated risks for short-
term exposure to construction workers and process employees. The
impacts of these risks can be reduced, as discussed above, by
implementation of a site-specific health and safety plan. There
is no long-term risk associated with any contaminated materials
as they would be detoxified on-site and delistable.
Prior to implementation of this alternative, a pilot test burn
would be required to prove the viability of the treatment
technology and to define system operating parameters. Based upon
the results of the pilot test, burn performance objectives would be
established. Once these steps are completed, the alternative could
be implemented relatively quickly.
The time to accomplish the on-site incineration alternative is an
estimated minimum of seven (7) months.
No long-term operation, maintenance, and monitoring requirements
other than continued qroundwater monitoring are associated with
this alternative.
C. LAND TREATMENT
This alternative assumes that all the contaminated materials at the
site are excavated and land-treated in the former railroad tie
storage area. Prior to implementation of this alternative, an on-site
pilot scale demonstration would he necessary using the actual waste
and site soils. Following this demonstration, a detailed engineering
design for the facility would be prepared. The demonstration test
- 20 -
-------
would take anoroximately six months to one year to comnlete. An
additional six months would be necessary to prepare a treatment
demonstration work plan on the front end and a detailed desian on
the back end.
Implementation of this alternative would involve the followinq
steps:
.1. site preparation including mobilization, establishment of
on-site support facilities, and construction of the land
treatment area (which would include qradinq, berm and ditch
construction and construction of a site perimeter fence and
road);
2. inflow control;
3. lagoon water removal, treatment (if necessary), and disposal;
4. excavation and waste spreading according to the design plan
and schedule;
5. soil incorporation and cultivation usina a rototiller device;
6. soil treatment for at least one year;
7. closure, including providing a vegetative cover and post-closure
care (includina groundwater, soil, and soil core water monitoring)
for un to thirty (30) years unless the zone of incorporation
could be delisted.
The initial excavation and waste application period could occur
within three months of initiating site activities. The treatment
period would extend for one to two years.
The primary institutional requirement for this alternative involves
completing an extensive demonstration test which would simulate
full-scale operating conditions.
d. OFF-SITE INCINERATION
This alternative includes excavation of all contaminated materials
at the site and transportation to an industrial process kiln, where
the materials would be processed into construction aggregate. The
steps reouired with this alternative include:
1. site preparation including mobilization and establishment of
on-site support facilities;
2. inflow control;
3. lagoon water removal, treatment (if necessary), and disposal;
4. excavation of lagoon sludges and contaminated soils, stabilization
of at least a portion of the lagoon sludqe;
5. transportation of excavated materials to an industrial process
kiln where they would be processed by incineration into asphalt
aggregate; and
6. site restoration.
All transportation loads must be manifested and carried by licensed
hazardous waste haulers. Permits for the industrial process kiln
would be verified prior to initiating the process. However, the
- 21 -
-------
responsibility for obtaining and maintaining the necessary permits
for the incinerator is that of the owner/operator.
The impacts of this alternative include elevated risks for short-
term exposure to site workers. In addition, there is some risk to
the general population associated with transportation of the materials.
There is no long-term risk associated with any contaminated materials
as they would be detoxified and delisted.
It is estimated that this alternative would take at least seven (7)
months to implement completely. No long-term operation and maintenance
reouirements except groundwater monitoring are associated with this
alternative.
e. TREATMENT AND OFF-SITE DISPOSAL
Implementation of this technology as a remedial action alternative
would reguire the following steps:
1. site preparation;
2. inflow control;
3. lagoon water removal, treatment (if necessary), and disposal;
4. excavation, treatment (if necessary) by a mechanical or
pozzolanic stabilization process of lagoon sludges and
contaminated soils;
5. disposal of treated and untreated materials at an EPA-apnroved
disposal site; and
6. site restoration.
Institutional considerations for this alternative include obtaining
approval from local agencies for discharge to the local POTW. The
permanent status of the disposal site must be verified and commercial
disposers must accept these contaminated materials. The licenses of
the hazardous waste transporter must be verified and waste manifests
prepared prior to shipment.
For this alternative ther are elevated risks for short-term exposure
mainly to site workers. There is also a short-term risk to the general
population associated with hauling of the contaminated materials. At
least four (4) months would be required to complete the excavation,
removal, and restoration alternative.
This alternative would remove all contaminated materials from the site,
so no long-term operation and maintenance requirements except
qroundwater monitorinq would be incurred.
- 22 -
-------
f. TREATMENT AND DISPOSAL OF SLUDGES AND
LAND TREATMENT OF SOILS
This alternative assumes1'that the sludoes are treated on-site usina
either mechanical treatnent or stabilization and disposed off-site;
and the contaminated soils are land-treated on-site. This alternative
is a combination of the land treatment and off-site disposal alternatives
It offers more flexibility than either alternaitve as the most hiahly
contaminated sludges are immediately removed from the site. The soils
of low-level contamination are temporarily stored while a treatment
demonstration is completed.
The basic steps of this alternative are:
1. site preparation;
2. inflow control;
3. lagoon water removal, treatment (if necessary), and disposal;
4. sludoe excavation, treatment (where necessary), and removal
to an EPA-approved disposal facility;
5. soil excavation and storage in several lined and temporarily
capped cells;
fi. land treatment area construction on-site on approximately
fourteen (14) acres;
7. soil spreading over land treatment area;
8. soil incorporation with nutrients, etc.; s
9. treatment; and
10. closure.
Institutional considerations for this alternative include obtaining
approval from local agencies for discharge to the POTW and completing
a land treatment demonstration.
Assuming a one-year period for the demonstration and design phase,
the treatment could begin within one year of initiating sludge
removal activities. As the application would occur in a single
batch process, the entire waste application could be completed
within one month and the treatment period would last approximately
two (2) years.
Post-closure activities could last for thirty (30) years.
g. BIOLOGICAL TREATMENT OF SLUDGES AND LAND TREATMENT
This alternative assumes that the sludqes are pretreated biologically
using sequenced batch reactors followed by land treatment of the
biosludge and contaminated soils. This alternative would significantly
reduce the organic content of the sludges thereby decreasing land
area requirements for the land treatment alternative.
The pretreatment process would use three (3) polyethylene lined
reactors, each approximately 250,000 gallons in capacity and
equipped with three (3) 25-horsepower agitator/aerator units.
The system would be operated in the plug flow mode with an average
solids residence time of seven (7) days/reactor. The resulting
- 23 -
-------
solids from this process would be hauled to the land treatment
area for final treatment. Vendors of this technology suaqest that
uo to 90% reductions in total creosote constituents would be
produced by this process.
This alternative would take a minimum of nine (9) months to complete.
It would require completion of a land treatment demonstration and
laboratory treatability studies to determine the final design
parameters for the sludge biological pretreatment process. Ttie
basic stens and schedule identified in the land treatment alternative
apply to thios alternative as well except that the land area
reguirements for the pretreated sludges decrease to about six (6)
acres.
- 24 -
-------
Table 3.0 j
Estimated Volumes
of
Contaminated Materials
( Prior to removal which occurred Dec.'87 - Feb.'88 under EPA approval )
ocation Material Area Depth Volume
(ft*2) (ft) (yd-3)
.riant
Ditch
T igoon
T igoon
Notes:
Area Soil
Sludge
Sludge
Soil Soil
(1) Area calculated from
40600(1)
3000(1)
106938(2)
106938(2)
"Residue Area"
0.5(3)
3.5(4)
0-4.0(4)
0-3(5)
750-1100
400-600
3000
1000-6500
, Fishbeck-Thompson Car
& Huber, Jnc. "Amax/Brown Wood Preserving Site Hydrologic
Study", Sheet 10, Project No. F 84816A.
(2) Law Engineering Testing Co., "Final Report of Investigation
for the Brown Wood Preserving Site, Job No. GE4271, dated
November 1984.
(3) Assumed average depth based on analytical results sample
numbers SL-055, 058, 061, 062, 067, 300-323, 400, 401.
(4) Assumed depth based on Test Hole 12 by Ardaman and
Associates, 1985, and Test Pit 3 by ERT, 1985.
(5) See text for assumptions
- 25 -
-------
Table 4.0
Contaminant Concentra I'lont neiow oaap.»»
With Total Creosote Concentration* >IOOO pp»
I
K>
Total Creoaote Concentration
Sample Description above belou
February 1983 Soil
Staple* (Table A. 4)
June 20-24, 1983
Ditch Soil Staple*
(Table A.5>
June 20-24, 1583
La loon Soil Sample*
(Table A. SI
September 1983 Soil
Borlnfa (Table A. 7)
January 1985
Soil Teat Hole Samples
From Lagoon Vicinity
(Table A. 9)
Fro* Plant Sit* Vicinity
(Table A. 10)
August 1986 Sol Teat > ta
in Lagoon (Table A. 15)
Drainage Ditch
Surface Impoundment
A-01, 2 - 2.5 ft
B-OI
C-01, 2 ft
D-OI, 0.5 - 1.5 ft
D-02. 2 - 2.5 ft
E-01, 0.5 - 1.5 ft
F-01, 0.5 - 1.5 ft
C-01, 1.4 - 2.2 ft
H-OI , 0.5 - 1.5 ft
1-07. 2 - 2.5 ft
J-05. 3.6 ft
K-02. 0 - 1.6 ft
k-03, 0 - 0.5 ft
L-02A, 0 - 1 .0 ft
1 - 1 .8 ft
L-06A. 0 - 2.6 ft
2.6 - 3.1 ft
L-07, 0 - 1.5 ft
A-6, 0 - 1.0 ft
2 - 3 ft
Teal Hole 5, 0 - 1 ft
3 - 5 ft
Teat Dole 7, 0 - 1 ft
3 - 6 ft
Teat Hole 8. 0 - 1 ft
3 - 5 'ft
Teat Hole 9, 0 > 1 ft
3 - 5 ft
Teat Hole 10, 0 - 1 ft
3 - 5 ft
Teat Hole 12, 0 - 1 ft
5 - 7 ft
Teat Hole 13, 0 - 1 ft
5 - 7 ft
11 - 13 ft
Teat Hole 14, 0 - 1 ft
9 - 11 ft
T«al Pit 3. 0 - 0.5 ft
7e»t Pit 2, 1 - 3 ft
2)0300
398300
2548
25566
70776-166910
26672-179769
5139
2S63
1052
1357
4414
10750
5764-199030
27764
1)4153
364R
2667
1014
50.1
1005
71
1615
0.34
3360
1.05
16300
116.2
37390
4.31
5345
13.11
146389
166.9
16770
107.2
1966
2713
1009
101 1 10
218
(ppnl Carcinogenic PAH'a ( ppn )
•bove belou
17500
21300
781
6927
1875 -27400
7)9 -60879
1006
895
19
29
1426
1540
111 -24000
5171
15960
1132
372
269
594
690
0
3800
800
11600
900
918
3960
164
24
24
0
0
10.8
0
0.15
8.63
6.65
88. S
118.2
15.9
X Carcinogenic PAH'*
•bove below
BY
5X
3IX
27X
3- ISX
3-34X
20X
35X
2X
2X
32X
14X
2-12X
I9X
I2X
3IX
37X
27X
37X
:ix
OX
10X
ISX
8X
SX
34X
4X
ex
4BX
34X
OX
ox
9X
OX
IX
SX
4X
5X
I?X
7X
-------
Table 5.0
CHRONOLOGY OF ENVIRONMENTAL SAMPLING
LIVE OAK. FLORIDA
rucR
EPA
EPA
EPA
EPA
EPA/PELA
PELA
LETCO
EEM
Date
Surfer/82
2/8/83
2/8/83
2/9/83
2/83
6/83
6/24/83
9/83
9/84
Sampled
surface and
ground water
(private wells)
air
ground water
(private wells)
surface water
sediment
(ditch and lagopn)
borings
surface water
borings
surface soil
n
4
7
3
2
3
7
7
For
"purgeable organics'
r—1
ten
EEM
EEM
EEM
r-i
1/85
8/85
8/85
9/85
10/86
1/87
(plant area)
borings
(ditch and plant)
surface soil
(plant)
\
surface soil
(wood storage)
ground water
(monitoring wells)
ground water
(monitoring well)
ground water
(monitoring well and
private well)
10
26
16
metals
c.reo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents
•creo. constituents * phenol, dibenzofuran and PAH associated with creosote.
•pep • pentachlorophenol
77100 P-D897 - .... - ....
- 27 -
-------
Table 6.0
INDICATOR CHEMICALS
Compound
height of
Solubility
(b)
Log Kow'
Benzo (A) Anthracene B2
Benzo (b) Fluoranthene B2
Benzo (a) Pyrene B2
Chrysene B2
Dlbenzo (a,h) Anthracene B2
Indeno (1.2.3,cd) pyrene C
Fluoranthene not ranked
Pentachlorophenol D
14
0.8
3.8
2
0.5
0.2
260
14000
5.61
6.06
6.06
5.61
6.77
4.90
5.01
(a) EPA (1986). There are only a limited number of chemical compounds that
have been demonstrated unequivocally to be human carcinogens. However,
experimental and epidemiologic data are available that are suggestive of
the carcinogenic activity of certain compounds. The EPA
"weight-of-evioence" system for ranking from A to D (in decreasing
order) the level of certainty that a compound Is a human carcinogen is
explained in the text. There are no A or B-l level carcinogens present
at the Live Oak Site. Certain PAH compounds present at the site have
been rated at B-2 or C-level potential carcinogens. To obtain an
appropriately conservative risk assessment the assumption is made that
compounds are human carcinogens, even if there is limited certainty that
this effect is real. As such, potential carcinogens down to level C
weight of evidence have been selected. (EPA, 1986).
(b) Solubility (in ppb), at 25"C. Data for PAH from Craun and Middleton.
1984, data for pentachlorophenol from EPA, 1980c.
(c) Log octanol/water coefficient. Data for PAH from EPA, 1980b.
Data for pentachlorphenol EPA. 1980c.
- 28 -
-------
F. SELECTED REMEDY DESCRIPTION AND RATIONALE FOR SELECTION
1. EXTENT OF REMOVAL ACTIVITY
An EPA-approved removal occurred in December 1987 through March 1983.
This removal was undertaken by AMAX/Brown in accordance with an
Administrative Consent Order conpleted in January 1988. The removal
included:
a. Mobilization, including establishment of a site office and
on-site safety zones;
b. Removal of lagoon water to allow excavation of sludges;
c. Excavation, treatment, and disposal of approximately
15,000 tons of creosote contaminated lagoon sludge and
soil at the Chemical Waste Management, Inc. (CVM) secure
landfill near Emelle, Alabama;
d. Treatment of approximately 200,000 gallons of lagoon water
removed as part of the lagoon excavation;
e. Extensive sampling and analyses of soil and water at
appropriate locations and times; and
f. Dismantling, decontamination, and disposal of the old wood
preserving plant in an EPA-approved manner.
More detailed information on the removal is contained in Attachment A
( "LIVE"OAK INTERIM REMOVAL" Memorandum, John Ryan, February 11, 1988 )
v
2. SELECTED REMEDY
The selected remedy embodies the remaining work necessary to complete
the remediation of the site after considering the work accomplished
by the removal activities in December 1987 through March 1988.
The recommended alternative is basically Alternative f. : Treatment
and Disposal of Sludges and Land Treatment of Soils. However, there
are four modifications to Alternative f.: (1) if the land treatment
{ biodegradation ) does not attain the desired cleanup levels for the
appropriate organic contaminants within the time allowed, then an
alternative means of dealing with the contaminated soils, such as
removal, incineration, solidification, or vitrification, will be
determined by EPA at that time; (2) ground water monitoring will
continue for five years; (3) contaminated soils exceeding certain
higher contaminant levels may be stabilized and removed to an EPA-
approved hazardous waste facility ( Emelle, Alabama or Pinewood,
South Carolina ); and (4) contaminated soils exceeding certain lower
contaminant levels will be biodegraded in an onsite excavated area
that has been lined.
The remedy is consistent with 40 CFR Part 300.68 (J) in that the
above-modified Alternative f. is technically feasible, alleviates
all existing and potential health effects, presents no new public
health hazards and substantially reduces the threat to the surface
and ground water and the general environment.
- 29 -
-------
Preference is given to this option because of technical feasibility,
cost, site-specific permanence, and the existence of Land Disposal
Restriction variances which allow its implementation within certain
advantageous time frames. Modification of the processes involved
in this option may be required in order to satisfy design requirements
and site conditions.
Creosote, the wood preservative, consists of approximately two hundi.J
(200) different compounds. The below-mentioned compounds are six (6)
of those two hundred. These compounds were selected as indicators
because the EPA weight-of-evidence system indicates that they are
"possible" or "probable" human carcinogens.
Benzo [a] anthracene
Benzo [a] pyrene
Benzo [b] fluoranthene
Chrysene
Dibenzo [a,h] anthracene
Indeno [l,2,3,c,d] pyrene
Note: Fluoranthene and pentachlorophenol were also considered.
The selection of indicator parameters is based upon numerous previous
priority pollutant analyses conducted during the Remedial Investigation
phase. Although other types of contaminants were present onsite,
these compounds are among the most common and potentially the most
carcinogenic found at the Brown Wood Site.
Cleanup standards were based upon the results of a Risk Assessment
which focussed on attaining at least a 1 x 10-6 risk with regard to
ingestion of contaminated soil by a child. Cleanup standards were
described by means of the total concentration of the six indicator
parameters.
A detailed cost development and analysis of selected remedial alternatives
was done prior to the removal activity ( prior to December 1987 ) to
assure that the most cost-effective remedial action was chosen for
the Brown Wood Site. Cost estimates followed the procedures specifl.J
in 40 CFR 300.68 (8)(2MB), Guidance on Feasibility Studies under CEK< \,
and the Remedial Action Costing Procedures Manual.
Twenty-four (24) specific remedial action alternatives underwent t!._
evaluation process. Fourteen (14) were eliminated on the basis of
site-specific application, technical feasibility, public health and
welfare, and environmental evaluations. The ten (10) remaining were
merged into seven (7). A detailed cost analysis was performed for
each of the remaining seven (7) alternatives. These alternatives are
listed in Table 7.0.
- 30 -
-------
Results of the Detailed Screening of Alternatives
( Prior to the removal which was conducted Dec.'87 thru Feb.'88 )
Alternatives
Effectiveness
C~ Sludge Biological
Mo On-Site Lend Off-Site Off-Site Treatment Pretreat
Action Inciner. Treat. Inciner. Disposal I Land I Land
Treatment Treatment
Protective of
Public Health Below Average Average Average Average Average Average
Attains ARARs Below Above Average Above Average Average Average
Reduces TMV
Below Above Average Above Below Average Average
Minimal Envir.' Below Average Average Average Average Average Average
Impacts
Reliable
Above Average Average Below Average Average Average
Implementobility
Feasibility &
Availability Above Below
Average Above Above Above Average
Constructable Above Below Above Above Above Above Above
Timeliness
Below Average Average Above Above Above Average
0AM Requirement Below Average Average Above Above Average Average
Administrative Average Below Below Average Above Average Below
Feasibility
Costs (SOOO)
Capital Cost
Annual OiM Co«t 20
5,440 1,900 11,300 4,140 2,740 2,400
Total Present 190 5,440 1.900 11,300 4,140 2,740 2,400
Value Cost
- 31 -
-------
The final remedy consists of three (3) major tasks.
a. Site preparation which will include the following activities:
1) Clearing, grubbing, and grading the proposed biological
treatment area, where necessary;
2) Installing a drainage swale in the appropriate locations(s);
3) Installing a perimeter fence around the land treatment area; and
installing signs on the perimeter fence warning against
exposure to hazardous material.
b. Construction of the Treatment System for biodegradation which
will include the following tasks:
1) Site grading;
2) Installing a liner throughout the treatinent area;
3) Installing a drainage systen on top of the liner and underneath
the soils to be treated;
4) Installing a spray irrigation system within the land treatment
areas;
5) Constructing a stockpile/holding area for the soils to be treat.J;
6) Installing a water aeration system within the land treatment area;
7) Hooking up utilities; and
8) Spreading contaminated soil and consolidating the stockpile.
- 32 -
-------
c. Operation and Maintenance
1) Tilling, irrigation, and fertilization of the land treatment
area;
2) Maintenance of the water/leachate treatment system;
3) Monitoring of the land and water treatment systems;
4) Semi-annual monitoring of the ground water quality for five (5)
years after completion of construction activities; review of
site's condition after five (5) years ( See Section 121(c),
CERCLA/SARA. ); and
5) Maintenance of security fences and warning signs.
3. CLEANUP STANDARDS
The cleanup standards for the selected remedy are based upon the
Risk Assessment and addenda and were finalized by EPA after
discussions with the FDER and AMAX/Brown. These standards are
describeable by referring to one factor: the total concentration
of carcinogenic creosote constituents ( indicators ).
a. Standards for soils treated in the old lagoon and new land
treatment area:
Within two (2) years from its initial seeding the land treatment
process must reduce the concentration of total carcinogenic
indicator chemicals to 100 parts per million ( ppm ) throughout
the volume of the material treated. This level for total
carcinogenic indicator chemicals corresponds, to an approximate
1 x 10-6 soil ingestion risk level. Upon successful completion
of the bioremediation in the land treatment area, the land
treatment area shall be revegetated.
- 33 -
-------
b. Standards for the plant area, the wood storage area and other
site areas:
Upon conpletion of the remedial activities in the plant area,
the wood storage area, and any other site areas, the soils
must contain no more than a 100 ppm total for the carcinogenic
indicator chemicals.
- 34 -
-------
4. RATIONALE FOR SELECTION OF FINAL REMEDY
The removal described in F.I. above, accordina to enaineering
calculations, caused the majority of the contamination-( i.e.,
the contaminated sludges and soils in and around the laaoon )
to be properly disposed of at the CW facility in Eire lie, Alabama,
or in Pinewood, South Carolina. The selected remedy described in
F.'2. above comprises the maior activities necessary to complete
remediation of the site in a technically feasible and cost-effective
manner consistent with CERCLA/SARA the NCP, and annlicable or
relevant and appropriate requirements ( ARAR's ).
With reqard to EPA's decision to enter into an Administrative
Consent Order with AMAX/Brown for the removal described in F.I.
above, Sections 104(a)(2) indicates that any removal action
undertaken by either the President ( EPA ) or by potentially
responsible parties with the President's ( EPA's ) approval
should "contribute to the efficient performance of any long
term remedial action" with respect to the release or threatened
release concerned. Section 122 (e)(6) indicates that once a
remedial investigation and feasibility study has been initiated
"no potentially responsible party may undertake any remedial
action at the facility unless such remedial action has been
authorized by the President" ( e.g., EPA authorizes activity by
means of a Consent Order ).
The removal activities not only eliminated the major source of
contamination and set the stage for the final site remediation
activities, but contributed to the acceleration of the site
along the Superfund enforcement process track. The final remedy
consists of those remaining activities which will bring the site
into compliance with those cleanup standards described in the
Risk/Health Assessment and addenda ( See Attachment B ) and
approved by both EPA and the State of Florida.
The selected remedy is a combination of two general source control
remedial actions as defined in 40 CFR 300.68(e)(2). These two
measures are: 1) removal to the OW facility in Emelle, Alabama,
or in Pinewcod, South Carolina, of the most severely contaminated
soils/sludges; and 2) onsite biodegradation of the less severely
contaminated soils to achieve the cleanup levels indicated in
the Risk/Health Assessment. In case number one, the governing
regulatory factor has been the RCRA Land Disposal Restriction
( LDR ) phase-in. The LDR phase-in schedule has allowed the
removal of creosote contaminated sludges and soils thus eliminating
potential onsite contamination by transporting the hazardous
substances to a secure hazardous waste landfill at a different
location. In case number two, an alternative and innovative
technology, biodegradation, is to be used on less severely
contaminated soils to "polish" those soils to acceptable levels
of cleanliness. EPA's Office of Solid Waste and Emergency
Response ( OSWER ), the Office of Technology Assessment ( OTA ),
and the Office of Research Development ( ORD ) have, in response
to CERCLA/SARA, recognized that our ability to characterize or
assess the extent of contamination,
- 35 -
-------
the chemical and physical character of the contaminants, or the
stresses imposed by the contaminants on complex ecosystems is
1 united, and new, innovative technologies are needed. The Brown
Wood Preserving Site new provides EPA with an excellent opportunity
to utilize and re-assess biodegradation in a controlled situation
in EPA's Region IV. The selected remedy dictates that the
bioreroediation will occur in a lined area which is designed with
interior drainage and spray irrigation systems. The land treatment
area will also have a perimeter security fence and an exterior
drainage system.
- 36 -
-------
III. RESPONSIVENESS SUMMARY
A. BACKGROUND ON COMMUNITY INVOLVEMENT
Historically, community concern regarding the Brown Wood Preserving
Site has been low, according to EPA, the FDER, local officials, and
the local news media. There have been no recorded corplaints from
the local residents. On the contrary, the community seems genuinely
glad that EPA and the FDER have acted rapidly to clean up the site.
B. SUMMARY OF PUBLIC COMMENTS AND AGENCY RESPONSES
There were no comments from the public during or after the public
comment period. An article indicating the tire and place for the
public meeting was placed in the Suwannee Democrat newspaper more
than two weeks prior to the actual meeting date. Few people
attended the public meeting and there were only three or four
general questions which had no bearing on the selected remedy.
Additionally, the administrative record recjuired under Section
113 of CERCLA/SARA was placed in the site repository given below
on September 29, 1987:
Suwannee River Regional Library
ATTN: Ms. Faye Roberts
Reference Librarian
207 Pine St.
Live Oak, Florida 32060
(904) 362-2317
Self-addressed and stamped envelopes were left with the administrative
record. The first draft of the Record-of-Decision and an instruction
sheet for commenters was also deposited with the other records.
However, the EPA Regional Project Manager received no 'mail or
telephone calls from the community regarding either the proposed
remedy or the actual onsite activities.
The Agency has responded to the low level of interest by periodically
speaking with newspaper reporters who are with the Suwannee Democrat.
Newspaper atricles followed these telephone and in-person discussions
with the reporters.
C. EXPLANATIONS OF DIFFERENCES BETVCEN PROPOSED PLAN AND
THE SELECTED REMEDY
The proposed plan in the Feasibility Study is generally the same as
the selected remedy in the Record-of-Decision. There are basically
five differences in the two: (1) the removal of lagoon sludges has
already occurred; (2) the grubbing of the majority of the land
treatment area has already occurred; (3) waters pumped from the
lagoon prior to the excavation of the sludges have already been
treated and properly disposed of; (4) the land treatment area will
be lined; and (5) the cleanup standards with reqard to the original
Risk Assessment were chanaed.
- 37 -
-------
D. COMMUNITY RELATIONS CONDUCTED AT THE SITE PRIOR TO AND DURING
THE PUBLIC COMMENT PERIOD
From 1984 through 1987 EPA, FDER, and AMAX/Brown representatives
maintained contact with Live Oak City officials while undertaking
the Remedial Investigation / Feasibility Study. Numerous newspaper
articles appeared in the local Suwannee Democrat describing the
onsite activities. Several City Council meetings were attended
and the community was made aware of EPA, FDER, and PRP activities.
The administrative record was installed in the site repository in
September, 1987, and through the Suwannee Democrat the community
was made aware of the progress made towards the cleanup. In
November through the middle of December, 1987, the public comment
period occurred. In December the public meeting occurred. There
were no public comments forthcoming from the community. Only
comments of a specific, technical nature were received from the
PRPs' contractor.
- 38 -
-------
ATTACHMENT A
Mernorandum on the Extent of Removal Activity
which occurred December 1987 throuqh February 1988
-------
MEMO TO: Distribution
FROM: J. Ryan
DATE: February 11, 1988
RE: LIVE OAK INTERIM REMOVAL
INTRODUCTION
This memorandum summarizes activities completed at the Live Oak
site as part of interim removal activities during the period
December, 1987 through the present. These activities have
included:
o mobilization,
o removal of lagoon water to allow excavation of sludges,
o excavation, treatment and disposal of sludges and
contaminated soil at the CWM secure landfill, Emelle,
AL, and
o treatment of the lagoon water removed as part of
excavation.
Extensive sampling and analyses of soil and water were completed
during these activities.
MOBILIZATION
Mobilization occurred the first week in December. An office
trailer and decontamination trailer were brought to the site as
veil as the necessary excavation and pumping equipment. A health
and safety plan was reviewed with the site workers and the site
separated into "clean" and "contaminated" zones for truck staging
and decontamination purposes.
diversion ditches were excavated around the perimeter of the
"lagoon to divert run on to the north. The location of these
itches are shown on Figure 1. A sedimentation trap was
.nstalled in the diversion ditches consisting of a series of hay
.bales olaced across the ditch.
-------
REMOVAL OF LAGOON WATER
Standing water was pumped from the lagoon using a vacuum truck.
This water was stored in four former product storage tanks on
site with an estimated capacity of 250,000 gallons. Once all
standing water was removed, a series of trenches were excavated
in the low part of the lagoon to continue dewatering the lagoon
while excavation was ongoing. The vacuum truck continued to
remove the free liquids while the excavation was underway.
EXCAVATION. TREATMENT AND DISPOSAL
The' lagoon sludges were treated with kiln dust in place by nixing
the dust with the sludge using power shovels. When a portion of
the lagoon was stabilized, it was excavated and staged on the
sides of the lagoon for subsequent loading onto the transport
vehicles.
Sampling of the upper four feet of soil was conducted in the
lagoon by excavating test pits with a ^racfchoe and collecting
samples from the pit walls. Only soil waa^s«mpled (no sludge) to
identify the depth of contamination that exceeded 1000 ppm of
total creosote substances (TCS) . Figure 1 shows the location of
the sampling points and the attached tables summarize the
analytical results. These results are prior to sludge
excavation. In general, no contaminated soil was found on the
western end of the lagoon whereas the eastern end had significant
contamination.
Table 1 presents a daily log of materials transported to Emelle.
A total of 15,000 tons of contaminated material was hauled froi.i
the site between December and the end of January. This total
includes over 6000 tons of sludges which exceeded 100,000 ppm
TCS. The remainder was highly contaminated soils greater than
5, 000 ppm TCS.
Low level contaminated sand and some of the underlying clays
also excavated. These soils are currently stockpiled around the
eastern end of the lagoon. It is estimated that these soils
represent approximately 10,000 tons of materials. Table 3
presents data from various samples collected from the lagoon and
the stockpiled soils which represent the average composition of
these stockpiled soils. The concentration of TCS ranges from
less than 1000 ppm to 5000 ppm with an average around 3000 ppm.
Carcinogenic PAH range from 100 to 200 ppm. It is this
stockpiled material which will be treated biologically.
Approximately one foot of native clays were backfilled in th_
lagoon and compacted in place to provide a contoured surface.
The lowest part of the lagoon is now approximately 10 feet lower
than its original surface. Standing water covers the bottom to a
-------
depth of approximately two feet. This water is run-off and has
an oily sheen. Figure 1 shows the current elevations of the
of the lagoon.
TREATMENT 07 WATER
The water which was removed from the lagoon to facilitate
excavation was stored in four tanks in the process area. In
addition, approximately 70,000 gallons of water were pumped out
of the retort pit. Approximately 200,000 gallons of water were
pumped into these tanks. Table 4 presents analyses of the water
prior to treatment.
The initial attempt to treat the water indicated that the water
would clog the fil-ters due to a strong emulsion. Subsequently a
flocculation step was added to break the emulsion. The current
treatment system consists of flocculation, followed by sand
filtration followed by a micron filter followed by carbon
adsorption. The treated effluent is then discharged to a series
of four temporary storage tanks, (20,000 gallons capacity per
tank) for sampling purposes. The stored effluent.. is then spray
irrigated over a three acre area on-site located to the west of
the plant.
The initial 40,000 gallons treated were sampled for volatiles,
semi-volatiles and copper chromium and arsenic. Results of this
sampling are noted below and the complete laboratory results are
presented as Table 5.
Tank «1 Tank 12
2 Butanol 39 ppb ND
Phenanthrene 5 ppb ND
Fluoranthene 5 ppb ND
Pyrene 3 ppb ND
Arsenic <5 ppb <5 ppb
Copper 3 ppb 3 ppb
Chromium 13 ppb 6 ppb
These results were transmitted verbally to Dr. Walker of the FDER
on Monday, February 1, 1988 and Ms. Danner, of U.S. EPA, on
Tuesday February 2. Both individuals gave permission to spray
irrigate the water. Ms. Danner also requested that we continue
to sample each tank for PAH and report the results prior to
spraying. Tanks 1, 3, and 4 were subsequently analyzed for PAH
compounds and these results were verbally transmitted to Dr.
Walker on Monday, February 8. All the PAH compounds were below
detection levels of two parts per billion in Tanks 3 and 4. Tank
1 had detectable quantities of certain PAH compounds.
-------
Fluoranthene was detected at 40 parts per billion and pyrene was
detected at 29 parts per billion. The remaining compounds were
less than 20 parts per billion. Dr. Walker gave verbal approval
to spray irrigate the water on Tuesday. To date, 100,000 gallons
of water have been spray irrigated. Treatment of the water will
continue until the contents of the tanks are emptied. It is then
proposed to proceed with the dismantling of the plant as
described in the attached memorandum .concerning a conceptual plan
for the final remedy.
-------
ATTACHMENT A
-------
TABLE 1
TRANSPORTATION LOG SUMMARY
LIVE OAK, FLORIDA
DATE
12-15
12-16
12-17
12-18
12-19
12-20
1-4
1-5
1-6
1-7 .
1-8
1-9
1-10
1-11
1-12
1-13
1-14
1-15
1-18
1-19
1-20
1-21
1-22
1-23
1-26
TOTAL
NET
TONS
304.11
602.13
473.45
698.73
517.13
322.61
1097.25
1231.27
1298.27
494.38
437.00
324.97
1318.33
1623.51
619.50
527.30
291.33
136.10
167. 18
643.39
430.67
639.79
595.72
238.92
22.51
15057.55
-------
TABLE 2
PAH CONCENTRATIONS IN LAGOON SOIL
PAGE 1
COMPOUND SAMPLE, CONCENTRATIONS IN ug/g (ppm)
NAPHTHALENE
2 -METHYL NAPHTH-
ALENE
ACENAPHTHYLENE
ACENAPHTHENE
FLUORENE
PHENANTHRENE
ANTHRACENE
FLUORANTHENE
PYRENE
BENZO (A) ANTH-
RACENE
CHRYSENE
BENZO (B) FLUOR-
ANTHENE
BENZO (K) FLUOR-
ANTHENE
BENZO (A) PYRENE
INDENO(1,2,3-CD)
PYRENE
DIBENZO( A, H) ANTH-
RACENE
BENZO (GHI)PERYLEN
TOTAL PAH
CARCINOGENIC PAH
2-3A
160
120
0
100
89
180
0
150
110
14
0
3
0
0
0
0
0
926
17
2-3B '
0.96
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1.96
0
2-3C /
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3-5A
20
. 0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
20
0
3-5B,
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3-5C -
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-------
TABLE 2
PAH CONCENTRATIONS IN LAGOON SOIL
PAGE 2
COMPOUND SAMPLE, CONCENTRATIONS IN ug/g (ppn)
NAPHTHALENE
2-METHYL NAPHTH-
ALENE
ACENAPHTHYLENE
ACENAPHTHENE
FLUORENE .
PHENANTHRENE
ANTHRACENE
FLUORANTHENE
PYRENE
BENZO(A) ANTH-
RACENE
CHRYSENE
BENZO (B) FLUOR-
ANTHENE
BENZO (K) FLUOR-
ANTHENE
BENZO (A) PYRENE
INDENO(1,2,3-CD)
PYRENE
DIBENZO( A, H) ANTH-
RACENE
BENZO (GHI)PERYLEN
TOTAL PAH
CARCINOGENIC PAH
4-6A
130
44
0
18
25
36
0
9
9
0
0
0
0
0
0
0
0
271
0
4-6B
130
120
0
140
110
180
0
120
100
13
0
0
0
0
0
0
0
913
13
5-3A
0
0
0
0
0
6.7
0
3
0
0
0
0
0
0
0
0
0
9.7
0
5-3B
0.6
0
0
0
0
6.8
0
3.5
1.5
0
0
0
0
0
0
0
0
12.4
0
6-1A
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6-1B
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-------
TABLE 2
PAH CONCENTRATIONS IN LAGOON SOIL
PAGE 3
COMPOUND SAMPLE, CONCENTRATIONS IN ug/g (ppm)
NAPHTHALENE
2 -METHYL NAPHTH-
ALENE
ACENAPHTHYLENE
ACENAPHTHENE
FLUORENE
PHENANTHRENE
ANTHRACENE
FLUORANTHENE
PYRENE
BENZO(A) ANTH-
RACENE
CHRYSENE
BENZO(B) FLUOR-
ANTHENE
BENZO (K) FLUOR-
ANTHENE
BENZO( A) PYRENE
INDENO(1,2,3-CD)
PYRENE
DIBENZO(A,H}AJJTH-
RACENE
BENZO (GHI)PERYLEN
TOTAL PAH
CARCINOGENIC PAH
6-2A
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6-2B
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6-3A
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
6-3B
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7-1A
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7-1B
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-------
TABLE 2
PAH CONCENTRATIONS IN LAGOON SOIL
PAGE 4
COMPOUND SAMPLE, CONCENTRATIONS IN ug/g (ppm)
7-2A 7-2B
NAPHTHALENE
2-METHYL NAPHTH-
ALENE
ACENAPHTHYLENE
ACENAPHTHENE
FLUORENE
PHENANTHRENE
ANTHRACENE
FLUORANTHENE
PYRENE
BENZO(A) ANTH-
RACENE
CHRYSENE
BENZO(B) FLUOR-
ANTHENE
BENZO (K) FLUOR-
ANTHENE
BENZO(A)PYRZNE
INDENO(1,2,3-CD)
PYRENE
DIBENZO( A, H) ANTH-
RACENE
BENZO ( CHI )PERYLEN
TOTAL PAH
CARCINOGENIC PAH
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7-3A
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7-3B
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
-------
COMPOUND
TABLE 2
PAH CONCENTRATIONS IN LAGOON SOIL
PAGE 5
SAMPLE, CONCENTRATIONS IN ug/g (ppm)
NAPHTHALENE
2 -METHYL NAPHTH-
ALENE
ACENAPHTHYLENE
ACENAPHTHENE
FLUORENE
PHENANTHRENE
ANTHRACENE
FLUORANTHENE
PYRENE
BENZO(A) ANTH-
RACENE
CHRYSENE
BENZO(B) FLUOR-
ANTHENE
BENZO (K) FLUOR-
ANTHENE
BENZO(A)PYRZNE
INDENO(1,2,3-CD)
PYRENE
DIBENZO (A, H) ANTH-
RACENE
BENZO (GHI) PER YLEN
TOTAL PAH
CARCINOGENIC PAH
2-1A
46
23
0
72
17
13
0
4.6
1.9
0
0
0
0
0
0
0
0
177.5
0
2-1B
2700
1800
0
1100
1600
1500
280
940
680
150
150
46
0
75
0
0
0
11021
421
2-2A
1000
470
0
580
.560
1200
31
940
660
150
187
64
0
41
0
0
0
5883
442
2-2B
90
64
0
62
110
64
0
86
61
0
24
0
0
0
0
0
0
561
24
3-1A
860
350
0
330
340
530
22
440
320
55
44
26
0
12
0
0
0
3329
137
3-1B
500
470
0
250
380
, 340
0
320
140
0
90
11
0
0
0
0
0
2501
101
-------
TABLE 2
PAH CONCENTRATIONS IN LAGOON SOIL
PAGE 6
COMPOUND SAMPLE, CONCENTRATIONS IN ug/g (ppm)
NAPHTHALENE
2-METHYL NAPHTH-
ALENE
ACENAPHTHYLENE
ACENAPHTHENE
FLUORENE
PHENANTHRENE
ANTHRACENE
FLUORANTHENE
PYRENE
5-1A
15
0
0
0
5.3
4.6
0
0
0
5-1B
0
0
0
0
0
3.6
0
0
0
BENZO(A) ANTH-
RACENE 0 0
CHRYSENE 0 0
BEN.ZO(B) FLUOR-
ANTHENE 0 0
BENZO (K)FLUOR-
ANTHENE 0 0
BENZO(A)PYRENE 0 0
INDENO(1,2,3-CD)
PYRENE 0 0
DIBENZO(A,H)ANTH-
RACENE 0 0
BENZO(GHI)PERYLEN 0 0
TOTAL PAH 24.9 3.6
CARCINOGENIC PAH 0 6
-------
TABLE 3
CONCENTRATIONS OF PAH COMPOUNDS IN SOILS TO BE LAND TREATED AT LIVE OAK SITE
COMPOUND SAMPLE CONCENTRATIONS IN mg/Kg (ppm)
NAPHTHALENE
2 -METHYL NAPHTHALENE
ACENAPHTHYLENE
ACENAPHTHENE
PLUORENE
PHENANTHRENE
ANTHRACENE,
FLUORANTHENE
PYRENE
BENZO (A) ANTHRACENE
CHRYSENE
BENZO (K) FLUORANTHE
BENZO (B) FLUORANTHE
BENZO (A) PYRENE
DI BENZO (A,H) ANTHRA
INDENO (1,2, -CD) PYR
BENZO (G.H.I) PERYLE
SPOILS
PILE
3-1
1100
410
<3
230
250
1500
<3
330
290
42
63
32
12
18
8
<3
3
SPOILS
PILE
LO-2
1800
650
<3
420
270
170
<3
190
74
23
45
41
<3
28
4
<3
5
SURFACE
SOIL
LO-1
1200
480
<3
280
270
340
44
470
200
39
40
25
7
23
3
<3
2
SPOILS
PILE
1-1
<3
<3
<3
41
43
81
56
420
320
48
96
32
29
35
<3
<3
<3
3-1A
860
350
<2
330
340
530
22
440
320
55
44
<2
26
12
<2
<2
<2
3-1B
500
470
<2
250
380
340
<2
320
140
<2
90
<2
11
<2
<2
<2
<2
AVERAGE
910
393
0
259
259
494
20
362
224
35
63
22
14
19
3
0
2
MAX
1800
650
0
420
380
1500
56
470
320
55
96
41
29
35
8
0
5
TOTAL PAH
4288
3720
3423
1201
3329
2501
3077
4288
-------
TABLE 4
ANALYTICAL RESULTS:
LAGOON (OR POND) WATER BEFORE TREATMENT
COMPOUND
NAPHTHALENE
2 -METHYL NAPHTHALENE
ACENAPHTHYLENE
ACENAPHTHENE
r LUORENE
rriENANTHRENE
ANTHRACENE
FLUORANTHENE
PYRENE
BENZO ( A) ANTHRACENE
CHRYSENE
BENZO (K) FLUORANTHENE
BENZO (B) FLUORANTHENE
BENZO (A) PYRENE
DIBENZO (A, H) ANTHRACENE
INDENO (1,2, -CD) PYRENE
BENZO ( G , H , I ) PERYLENE
PENTACHLOROPHENOL
SURROGATES
% RECOV. FLUOROBIPHENYL
% RECOV. TRIBROMOPHENOL
% RECOV. TERPHENYL-D14
SAMPLE CONCENTRATIONS IN ug/1 (ppb)
INFLUENT
11
20500
2700
<5
5100
4600
10600
<5
3800
2100
320
370
56
210
400
<5
<5
<5
2100
34
120
55
IBFLUENT
#2
560
180
<5
920
1200
9300
<5
4800
4000
400
720
40
360
480
<5
<5
<5
<1000
52
110
98
-------
-------
EXTRA 3
6-3
X 6-2
N~XTRA 2 *
V
SITE PLAN
100
100 200 300 400
SCALE :EET
-------
ATTACHMENT A
-------
Lciucks
Testing Laboratories, Inc.
o-C S .i:h H.'"\-T5i Seattle U>*hinw''on 98!CE (CO*<767-?C*C
Certifies
and Technical Sernccs
CLIENT: Remediation Technology, Inc.
19219 West Valley Hwy., Suite M103
Kent, WA 98032
ATTN: John Ryan
REPORT ON: WATER
SAMPLE
IDENTIFICATION:
Submitted 01/29/88 and Identified as shown:
1) Water II 01/28/88 16:15
2) Water #2 01/28/88 16:30
LABORATORY NO. 8028
DATE: Feb. 2, 19P
Proj. No. C86-006-910
TESTS PERFORMED
AND RESULTS:
parts per million (mq/L)
Arsenic
Copper
Chromium
<0.005
0.003
0.013
<0.005
0.003
0.006
Samples were analyzed 1n accordance with Test Methods for Evaluating Solid
Waste. (SW-846). U.S.E.P.A., 1982. Method 8270 (semi-volatile extractables)
Extractables (bv GC/MS)
Phenol
*An1l1ne
B1s(2-Chloroethyl)Ether
2-Chlorophenol
1,3-Dlchlorobenzene
parts per billion (uq/L)
Lab
Blank
To* t«oex t
. o< ti,
UM en m* «•«• o" ">
co"**"! »ce»O'» "0
« yooo i*m tne «ccai»ng to IM 'von e< i«* MM* i-xi o<
-------
Lauchj
Testing Laboratories, Inc.
Certificate
QJICJ
Chen i <;r? Mcrobdof? and Technca! Ser?>ces
Remediation Technology, Inc.
PAGE NO. 2
LABORATORY NO. 8028
parts per billion (uq/L)
1,4-Dichlorobenzene <2.
*Benzyl Alcohol <2.
1,2-Dichlorobenzene <2.
*2-Methylphenol <2.
B1s(2-Chloro1sopropyl)Ether <2.
*4-Methylphenol <2.
N-N1troso-D1-n-Propylam1ne <2.
Hexachloroethane <4.
Nitrobenzene <2.
Isophorone <2.
2-NHrophenol <2.
2,4-01methylphenol <2.
*Benzo1c Acid <50.
B1s(2-Chloroethoxy)Methane <2.
2,4-D1chlorophenol <4.
1,2,4-Trlchlorobenzene <2.
Naphthalene <4.
*4-Chloroan1l1ne <2.
Hexachlorobutadlene <2.
4-Chloro-3-Methylphenol <4.
*2-Methylnaphthalene . <2.
Hexachlorocyclopentadlene <4.
2,4,6-Trlchlorophenol <4.
*2,4,5-Tr1ch1orophenol <4.
2-Chloronaphthalenc <2.
*2-N1troan1l1ne <4.
Dimethyl Phthalate <2.
Acenaphthylene <2.
*3-N1troan1l1ne <10.
Acenaphthene <2.
2.4-01n1trophenol <20.
Lab
Blank
<50.
<20.
in" i C0""*oa" •"
-------
Laucfej
Tc "ting Laboratories, Inc.
9-sO
\iTjfhmcion 98ICS
Certifica^
ChemiMry Mcrobdu^y and Technical Service?
Remediation Technology, Inc.
PAGE NO. 3
LABORATORY NO. 8028
4-Nitrophenol
*D1benzofuran
2,4-D1n1trotoluene
2,6-01n1troto1uene
Dlethyl Phthalate
4-Chlorophenyl-Phenylether
Fluorene
*4-N1troan111ne
4,6-01nitro-2-Methylphenol
N-N1trosod1pheny1amine
l,2-D1phenylhydraz1ne
4-8romophenyl-Phenylether
Hexachlorobenzene
Pentachlorophenol
Phenanthrene
Anthracene
Di-n-Butyl Phthalate
Fluoranthene •
Pyrene
Benz1d1ne
Butyl benzylphthalate
3,3'-01chlorobenz1d1ne
Benzo(a)Anthracene
B1s(2-Ethylhexy1)phthalate
Chrysene
01-n-Octyl Phthalate
Benzo(b)Fluoranthene
Benzo(k)Fluoranthene
Benzo(a)Pyrene
Indeno(l,2,3-cd)Pyrene
Dlbenzo(a,h)Anthracene
Benzo(g,h,1)Perylene
parts per bHHpn (uq/L)
1 2
<20. <20.
<20.
<20.
5.
3.
<50.
<2.
<20.
<20.
<20.
<50.
<2.
<20.
Lab
Blank
<20.
<20.
<20.
<50.
<2.
<20.
t 0* &^T P*C0*^* 9 I
id !•«* «nd «ccqiO*^ 10 *• 'w*M 0* t^« I
-------
Lducks
Testing Laboratories, Inc.
Certificate
Seattle
Mrr McmboiccF and Technical Services
Remediation Technology, Inc.
PAGE NO. 5
LABORATORY NO. 8028
parts per bUHon (uq/L)
Lab
Blank
*4-Methyl-2-pentanone
*2-Hexanone
1,1,2.2-Tetrachloroethane
Tetrachloroethene
Toluene
Chlorobenzene
trans-l,3-Dlchloropropene
Ethyl benzene
c1s-l,3-D"ichloropropene
*Styrene
*Total Xylenes
Key
* - additional compounds from the EPA's Hazardous Substances List
< - "less than"
Respectfully submitted,
Laucks Testing Laboratories, Inc.
'I • !
'/•'•/ •'.-:„•••• ' '
I
J. M. Owens
OMOremt
UM e< m* •%•>«> o<
*cco>»ng ID m* »«*« o> <*• "M* *"d
-------
— V • • ~^^^" V^HI^M W^B^V ^^^^^^"
Testing Laboratories, Inc.
Certificate
ChemiMry Mcrobctos? and Technical Sertxcs
APPENDIX
Surrogate Recovery Quality Control Report
Attached are surrogate (chemically similar) compounds utilized 1n the analysis
of organic compounds. The surrogates are added to every sample prior to extraction
and analysis to monitor for matrix effects, purging efficiency, and sample.
processing errors. The control limits represent the 95X confidence Interval
established 1n our laboratory through repetitive analysis of these sample types.
T KC*B>t
r O- *"•
u »iccn'
«n« o< to*nc«
-------
CLIENT ,-'
.CONTACT " /... !
PHOENIX ANALYTICAL LA&QRATOftlES, INC.
CHAIN-OF-CUSTQDY
TVP£.SIZE
CLIEN
PROJECT NO.
scredaneiM fount L
FtELlNOUISHED 6Y:
FRtNT NAnE
SIGNATURE
DAY
TIME
RECEIVED BY:
PRINT
DAY
TIME
RELINQUISHED BYi
PRINT NAME
SIGNATURE
DAY
TIME
RECEIVED BY:
PRINT NAME
SIGNATURE
DAY
TIME
RELINQUISHED BYi
PRINT NAME
SIGNATURE
DAY
TIME
RECEIVED BYi
PRINT NAME
SIGNATURE
DAY
TIME
REMARKS
INITIALS,
-------
ATTAOT1FNT *
Risk Assessment and addenda
-------
Document No. P-D897
Risk Assessment to Accompany
the Feasibility Study of
the Live Oak Superfund Site,
Live Oak, Florida
Prepared for:
Amax, Inc.
Golden, CO
The James Graham
Brown Foundation
Louisville, KY
August 1987
-------
TABLE OF CONTENTS
Page
EXECUTIVE SUMMARY ES-1
1. INTRODUCTION 1-1
2. BASELINE ASSESSMENT 2-1
2.1 Selection of Indicator Chemicals 2-2
2.2 Exposure Assessment 2-8
2.3 Toxicity Assessment 2-24
2.4 Risk Characterization 2-30
3. DEVELOPMENT OP PERFORMANCE GOALS 3-1
4. SHORT-TERM RISKS OP REMEDIAL ALTERNATIVES 4-1
4.1 Dermal Toxicity 4-1
4.2 Transportation Risks 4-1
4.3 Air Emissions 4-2
REFERENCES
APPENDIX A GROUND WATER CONTAMINANT MODELING
APPENDIX B DETERMINING SURFACE SOIL CONCENTRATIONS
7779D PD-897
-------
EXECUTIVE SUMMARY
INTRODUCTION
The following report is a health baseline risk assessment
of the Live Oak Superfund Site in Live Oak, Suwannee County
Florida. The cepoct follows the guidelines outlined in the
Superfund Public Health Assessment Manual (EPA 540/1-86/060)
and is intended to accompany the Feasibility Study Report for
the site developed by Remediation Technologies Incorporated. A
summary of the history of the site and its use as a wood
preserving facility is presented in the Feasibility Study
Report as well as the Remedial Investigation Report, prepared
by Fishbeck. Thompson. Carr and Huber and Environmental
Engineering and Management. Limited.
SELECTION OF INDICATOR CHEMICALS
Sampling of media for chemical analysis has been carried
out' by a large number of agencies and contractors. The
chronology of sampling, which occurred between 1982 and 1986 is
presented in Table 2-1 of the following report. Most analysis
was done with a focus on constituents typically used at wood
preserving sites. Creosote constituents and a small amount of
pentachlorophenol were found to be present in soil, sediment.
and surface water at the site. Certain low molecular weight
fractions of creosote were also found in ground water. No
metals wece found at the site. Based on these findings.
creosote constituents and pentachlorophenol were selected as
Indicator Chemicals for detailed health risk assessment at the
site. The creosote constituents were evaluated as a whole for
their potential to produce acute, non-carcinogenic effects on
akin. Additionally, six high molecular weight components of
creosote were evaluated for potential to produce a carcinogenic
response (Benzo[a]anthracene. benzo[aJpyrene.
benzo[b]fluoranthene. chrysene. dibenzo[a.h]anthracene. and
ES-1
7857D PD-897
-------
indeno [1.2.3.c.d.]pycene). These compounds were selected
because the U.S. EPA weight-of-evidence system indicates they
re "possible" oc "probable" human carcinogens. Another
creosote constituent, fluoranthene. was selected for evaluation
because it is present in significant amounts at many of the
contaminated areas of the site, and because of .the existence of
an Ambient Water Quality Criterion for this compound.
Pentachlorophenol was evaluated for its potential to produce
acute, non-carcinogenic effects on skin and also for its
potential to produce systemic toxic effects.
EXPOSURE ASSESSMENT
At any site, humans may be exposed to contaminants in air.
water, or solid media such as soil and sediments. They may
ingest, inhale, or absorb a compound. In certain cases, skin
contact without absorption into the system may be considered an
exposure. Consideration of the location of the site, its
accessibility, and regional hydrogeology indicated that the
pertinent routes of potential exposure for the Live Oak site
were ingesti.on of drinking water from the Ploridan aquifer if
it were to become contaminated, and acute dermal contact with
or Lngestion of constituents in surface soils by trespassers or
visitors to the site.
Concentrations of Indicator Chemicals used in the risk
assessment were determined from analytic results or estimated
with the aid of mathematical models. The models used for
determining potential impact on ground water were a leaching
model developed by EPA (the Organic Leaching Model), an
analytical one-dimensional unsaturated zone model, and an
analytical saturated zone model (the Horizontal Plane Source
Model). The models are described in detail in Appendix A of
this report. The PAH inputs for the model were obtained from
data on sediments and soils in the lagoon area, sampled by the
EPA and P.E. LaMoreaux on June 20-24. 1983. Data from the
lagoon were used because this is the area with the largest
ES-2
7857D PD-897
-------
waste load and the most likely place foe ground water
contamination to occur. The exposure point chosen for the
drinking water risk assessment was the nearest downgradient
offsite well (Well Number 15 on Sheet 10 of the Remedial
Investigation Report). For the direct contact and soil
ingestion assessments, on-site surface soil and sediments not
normally covered by water were considered.
METHODS USED TO ASSESS BISK
Risk Assessment was carried out using two methods:
• The analytical and predicted concentrations were
compared to relevant standards and criteria; and
• Potential intake of Indicator Chemicals using
exposure scenarios in which individuals were assumed
to drink water from potentially-affected wells vor
inadvertently ingest surface soils and sediments
while visiting the site (the lat.uer scenario is
assumed to be relevant to children who soil their
hands and put them in their mouth). Potential
intakes of non-carcinogenic Indicator Chemicals (PCP
and fluoranthene) were compared to "acceptable
intakes" developed by the EPA. Potential intakes of
the carcinogenic PAH were used to calculate cancer
risks using "potency factors" developed by the EPA.
For dermal contact, concentrations in the soil were
compared to apparent minimal-effects levels
determined from the literature.
COMPARISON OP INDICATOR CHEMICAL CONCENTRATIONS TO STANDARDS
AND CRITERIA
Comparison of measured or predicted concentrations of
Indicator Chemicals in off-site well water to Ambient Water
Quality Criteria (the only criteria that appear to be relevant
for the Live Oak site) indicate that no impact from
ES-3
7857D PD-897 . -
-------
pentachlorophenol or fluoranthene is expected. The Ambient
Water Quality Criteria foe carcinogenic PAH gives a range of
values for various risks of cancer. The summed concentration
of the carcinogenic PAH predicted to be in the nearest off-site
well is less than the concentration associated with a cancer
risk of 1 chance in 1.000.000.
ESTIMATION OF CAHCINOGENIC RISK
The potency factor for carcinogenic PAH is based on
observations of carcinogenic actions of a single compound.
benzo[a]pyrene. As this is among the most carcinogenic PAH.
uHing a potency factor based on benzo[a]pyrene for all
carcinogenic PAH is conservative. The factor is
0.0115/microgram/kilogram body weight day. This value
indicates that an individual taking in one microgram of
carcinogenic PAH per Kilogram of body weight every day. ^for
life would have a cancer risk of .0115 (i.e. slightly more than
one chance in one hundred) in excess of "background" risk.
Using the factor with the predicted intake of carcinogenic
PAH that could occur as a result of lifetime ingestion of water
in the nearest downgradient well if contamination were to occur
indicates that the upper limit cancer risk from this Activity
is slightly less than 1 chance in 1.000.000. This low risk is
often considered "virtually safe."
Using the factor with predicted intake from a less likely
exposure scenario, Che assumption that children could
infrequently visit the site and ingest soils and sediments
contaminated with PAH. indicates that an upper-bound cancer
cisk of eighty eight chances in one hundred thousand.
COMPARISON OF PREDICTED INTAKES TO ACCEPTABLE INTAKES OF
NON-CARCINOGENS
Acceptable intakes for pentachlorophenol and fluoranthene
are 30 micrograms/kilogram body weight day. and 6.12
micrograms/kilogram body weight day. respectively. These
ES-4
7657D PD-897
-------
values vece developed by the EPA under the assumption that
these substances, as non-carcinogens, have a "threshold" for
their toxic effects. That is. there is a dose below which
virtually no risk of a toxic response exists. The acceptable
intakes are assumed to be below the threshold dose for
pentachlorophenol and fluoranthene. No estimated intake for
pentachlorpphenol or fluoranthene using any exposure scenario/
was in excess of the acceptable intakes developed by the EPA.
A no-effect level for the dermal effect of creosote and
pentachlorophenol was determined by extrapolating from reports
in the literature on the photosensitization effects of creosote
constituents and the dermal irritation produced by
pentachlorophenol. It was determined that concentrations of
creosote or pentachlorophenol in excess of 1000 ppm might
produce acute dermal effects. Inspection of analytical results
indicate that some surface toils and sediments that are not
continuously covered with water have concentrations of creosote
and pentachlorophenol above 1000 ppm. Thus, tome risk of
transient dermal effect may exist for the current condition of
the site.
DEVELOPMENT OF PERFORMANCE GOALS
For source control remedial alternatives, the Supecfund
d\
Public Heal Evaluation Manual suggests that an analysis to
determine "acceptable" levels of residual contamination be
performed. No evaluation was deemed necessary for
pentachlorophenol or fluoranthene because the baseline
assessment indicated no present or future health impacts of the
compound. The low level of potential carcinogenic risk from
drinking water from off-sitt alto precluded the necessity of
performance goals based on this limited health impact. The
most likely health impact for the Live Oak site is the
potential for non-carcinogenic acute dermal effects from
creosote and pentachlorophenol. Th,us. performance goals of no
more than 1000 ppm creosote or pentachlorophenol in soils or
sediments is recommended. If all media containing
ES-5
7857D PD-897 ...
-------
pentachlorophenol or creosote in excess of these concentrations
i
is removed, it has the secondary effect of decreasing the
potential cancer risk produced by inadvertent ingestion of
soils and sediments. Using the exposure scenario developed in
the baseline risk assessment, the Bean concentration of
carcinogenic PAH calculated to remain following removal of
creosote compounds in excess of 1000 ppm. the cancer risk from
ingestion of toils is calculated to be slightly greater than
one chance in 1.000.000.
ES-6
7857D PD-897
-------
1. INTRODUCTION
The following report is provided to support the
•asibility Study of the Brown Wood Preserving Site in Live
Oak. Florida prepared by Remediation Technologies for Amax,
Inc. and The James Graham Brown Foundation. Inc. It is a
baseline health risk evaluation, and documentation of 'the
method used foe developing performance goals based on human
health considerations. The format follows the guidance
provided in the Superfund Public Health Evaluation Manual (EPA.
1986).
The baseline evaluation is a health risk assessment of the
current condition of the site and. as such, represents a
screening of the "no-action alternative." It indicates if a
remedial action is needed and how quickly steps must be taken.
The methodology developed in the baseline assessment also
provides the framework for developing performance goals used
for screening remedial alternatives.
The remedial alternatives (other than "No Action") that
have been selected for screening for the Live Oak site are
source control measures. EPA guidance (EPA. 1986) suggests
that such alternatives must be screened for their ability to
fulfill requirements of the National Contingency Plan and best
engineering judgment. However, as suggested in the Superfund
Public Health Evaluation Manual, health-based criteria can be
useful in deriving acceptable residual levels of constituents
in soils (performance goals). Performance goals will be
calculated to provide public health protection at the exposure
points identified in the baseline assessment. Specifically.
performance goals will be set which would ensure exposure below
the Acceptable Intake foe Chronic exposure (AIC) for
non-carcinogenic toxic constituents and provide for low risk
from carcinogenic substances. These values provide an
objective, health-based criteria for determining the relative
effectiveness of remedial alternatives.
1-1
7709D PD-897
-------
To the extent possible, an analysis of the short term
health risks associated with the remedial actions will be
provided. Under EPA health assessment guidance (EPA. 1986)
this analysis is intended to provide guidance for the health
and safety plan accompanying the chosen action.
1-2
7709D PD-897
-------
2. BASELINE ASSESSMENT
The format of this baseline assessment follows the 4-step
methodology recommended by the EPA (49 FR 46304. November 23,
1984; 50 FR 1170, January 9. 1985. and EPA, 1986). It should
be noted that the terminology used in the Federal Register and
the Superfund Public Health Evaluation Manual (SPHEM) is
different. Although the differences are only semantic, readers
may be familiar with only one terminology. To avoid confusion,
the terms and a short description of the methods involved with
each step are provided here.
Step 1: Selection of Indicator Chemicals in the
Superfund Public Health Evaluation Manual is
comparable to the Hazard Identification in the
Federal Register. The assessment reviews the
contaminants found in various media. Indicator
chemicals are then chosen based on
concentration, distribution, toxicity and
consistency of detection.
Steps 2 and 3: Estimation of Exposure Point Concentrations and
Estimation of Chemical intakes are comparable to
the Exposure Assessment mentioned in the Federal
Register. The section reviews the potential
exposure pathways, compares relevant standards
and criteria to concentrations at exposure
points and calculates expected doses from
plausible exposure scenarios.
Step 4: Toxicity Assessment is comparable to Dose-
Response Assessment. The section presents a
toxicity profile and develops a dose response
relationship for each Indicator Chemical. The
general source foe this information is the
supporting literature for the standards and
criteria for the constituents (even if a
standard or criterion is not pertinent to the
2-1
7711D D897
-------
exposure situation, the toxicity literature is
often useful). Updated information will also be
analyzed.
Step 5: Risk Characterization is titled the same in the
Federal Register and the SPHEM. In this section
expected doses are compared to the dose response
estimation in order to quantitate risk for .the
site specific conditions.
The assessment of conditions at the Live Oak site is based
on field observations as presented in the Remedial
Investigation (RI) performed by Fishbeck. Thompson. Carr. and
Huber. and Environmental Engineering and Management Ltd.
Analytical data on a variety of samples taken by several
organizations (Florida DER [FDER], EPA. P.E. LaMoreaux and
Associates [PELAJ. Law Engineering and Testing Co. [LETCO], as
well as Environmental Engineering and Management Ltd [EEM])
between 1982 and 1987 were assessed. Table 2-1 presents the
chronology of sampling.
2.1 SELECTION OF INDICATOR CHEMICALS
All of the sampling rounds at the Live Oak Site have been
focused on constituents potentially present, given the past use
of the property as a wood preserving operation. Creosote is
the principal preservative reported to have been used at the
site during operations by Brown and Amax. Pentachlorophenol
and arsenic-metal complexs (e.g. Chromated Copper Arsenate.
Chromated Zinc Chloride) are also common wood preserving
compounds. Although not reported to have been used by Brown or
Amax. they have been addressed in sample analysis. The reasons
for selection or rejection of a compound as an Indicator
Chemical is detailed below and the final list of chosen
constituents is given in Table 2-2.
2-2
7711D D897
-------
TABLE 2-1
CHRONOLOGY OF ENVIRONMENTAL SAMPLING
LIVE OAK, FLORIDA
FDER
EPA
EPA
EPA
EPA
EPA/PELA
PELA
LETCO
EEM
Date
Summer/82
2/8/83
2/8/83
2/9/83
• 2/83
6/83
6/24/83
9/83
9/84
Sampled
surface and
ground water
(private wells)
air
ground water
(private wells)
surface water
sediment
(ditch and lagoon)
borings
surface water
borings
surface soil
n
4
7
3
2
3
7
7
For
•purgeable organics'
EEM
EEM
EEM
EEM
EEM
EEM
1/85
-8/85
8/85
9/85
10/86
1/87
(plant area)
borings
(ditch and plant)
surface soil
(plant)
surface soil
(wood storage)
ground water
(monitoring wells)
ground water
(monitoring well)
ground water
(monitoring well and
private well)
10
26
16
5
metals
cfeo. constituents, pep
creo. constituents, pep
N
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents, pep
creo. constituents
•creo. constituents « phenol, dibenzofuran and PAH associated with creosote.
*pcp • pentachlorophenol
7710D P-D897
2-3
-------
TABLE 2-2
INDICATOR CHEMICALS
Compound
Weight of
Evidence***
Benzo (a) Anthracene B2
B«nzo (b) Pluoranthene B2
Benzo (a) Pyrene B2
Chrysene . B2
Dlbenzo (a,h) Anthracene B2
Indeno (l,2,3,cd) pyrene C
Fluoranthene not ranked
Pentachlorophenol D
Solubility(t))
14
0.8
3.8
2
0.5
0.2
260
14000
5.61
6.06
6.06
5.61
6.77
4.90
5.01
(a) EPA (1986). There are only a limited number of chemical compounds that
nave been demonstrated unequivocally to be human carcinogens. However,
experimental and epidemiologic data are available that are suggestive of
the carcinogenic activity of certain compounds. The EPA
"weight-of-evidence" system for ranking from A to D (in decreasing
order) the level of certainty that a compound is a human carcinogen is
explained in the text. There are no A or B-l level carcinogens present
at the Live Oak Site. Certain PAH compounds present at the site have
been rated at B-2 or C-level potential carcinogens. To obtain an
appropriately conservative risk assessment the assumption is made that
compounds are human carcinogens, even if there is limited certainty that
this effect is real. As such, potential carcinogens down to level C
weight of evidence have been selected. (EPA, 1986).
(b) Solubility (in ppb). at 25*C. Data for PAH from Craun and Middleton,
1984, data for pentachlorophenol from EPA, 1980c.
(c) Log octanol/water coefficient. Data for PAH from EPA, 1980b.
Data for pentachlorphenol EPA, 1980c.
2-4
-------
2.1.1 Creosote Constituents
Creosote is defined by its physical properties (selected
specifically to make the material suitable for wood
preservation) rather than its chemical composition. As a
complex mixture, the toxicity of creosote may vary with each
lot. Creosote is a mixture of aromatic compounds, including:
• Light aromatic compounds, including benzene.
naphthalene and substituted aryl structures such as
xylenes and toluene. These constituents are present
at low levels because their boiling points are lower
than the creosote fractionation temperature range.
• Phenol and substituted phenols (e.g. cresols).
• Polynuclear aromatic hydrocarbons (PAH).
Benzene and substituted benzenes have been detected once.
by the FDER in 1982. At that time, they were found to be
present in surface water of a shallow hole dug in the drainage
ditch on site, but not in four ground water wells off site.
The FDER and EPA (in summer of 1962 and February of 1983.
respectively) did not find naphthalene, phenol or PAH in
private ground water wells offsite (no constituents were found
in the private well sampled by EEM in 1987). nor were they
found in ground water monitoring well samples during the 1985
sampling rounds by EEM. No phenol was found by EEM in 1986
sampling of ground water, while very low concentrations of
naphthalene and low molecular PAHs (acenaphthene. dibenzofuran.
fluorene. phenanthrene and anthracene) were reported in two
on-site monitoring wells in 1986 and 1987.
Phenol, naphthalene, and PAH were detected in surface
water on site by FDER and EPA (dates mentioned above), but
analyses by PELA in Juno of 1983 were negative.
Phenol was not detected to a significant extent in soils
analysis during the Remedial Investigation, and was calculated
to be present in aic to the extent of about 0.5 ppb (EPA-Fit
Study No. Z0830202. 1983). a concentration which is four orders
of magnitude below the NIOSH Exposure Limit (NIOSH. 1985).
2-5 ..
7711D D897
-------
Thus, phenol has been eliminated from further consideration in
the baseline risk assessment due to its limited presence at the
site. Soil analyses done during the Remedial Investigation
indicate that naphthalene, methylnaphthalene. and certain PAH
are present in soils in specific areas at the site. However.
naphthalene and methylnaphthalene have faicly low acute
toxicity (lethal dose in humans is 2-15 g. Sandmeyer. 1981) and
their chronic toxicity is inadequately studied. Therefore, the
naphthalenes were not assessed further. PAH with an adequate
toxicology base were selected as Indicator Chemicals.
The toxic effect of primary concern foe PAH is
carcinogenicity. The U.S. EPA has developed a
"weight-of-evidence" system to classify the data that is
suggestive of human carcinogenic activity of individual PAH
compounds. The weight of evidence categories are:
1) A - human carcinogen. Demonstrated human
carcinogenic activity.
2) B-l - probable human carcinogen. Suggested by
limited studies in humans.
3) B-2 - probable human carcinogen. Suggested by
lifetime studies in animals.
4) C - possible human .carcinogen. Suggested by limited
studies in animals.
5) D - no data, or no demonstrated carcinogenic activity.
There are no A oc B-l level carcinogens among the PAH detected
at the Live Oak Sit*. Six PAH compounds found at the site have
EPA eatings of "probable- (B-2) to "possible" (C) human
carcinogens (EPA. 1986) and have been chosen for detailed risk
analysis in the present assessment. They are listed in
Table 2-2.
Additionally, a non-carcinogenic PAH. fluoranthene. has
been selected because there is adequate dose-response data to
assess the toxicity of this compound (EPA. 1980a).
As will be detailed below, an acute dermal toxic effect of
PAH may also be important at the Live Oak Site (An acute effect
2-6
7711D D897 ...
-------
is one which may occur after a single instance of contact with
the chemical, and generally happens within hours to days after
exposure). All PAH will produce sun sensitivity to varying
degrees. It is neither possible or necessary to determine the
dermal toxicity risk of each PAH. Rather a value intended to
protect against photosensitivity will be developed for total
PAH. Even the PAH which have been eliminated from assessment
foe systemic toxicity will be included la this value.
2.1.2 Pentachlorophenol
Pentachlorophenol has been detected in surface water but
not in ground water samples from the on-site or offsite
monitoring wells. The compound appears to be less widely
distributed in soils than creosote constituents. However there
are concentrations in the sediments that could conceivably pose
a risk through direct human contact oc as a source for future
ground water contamination. Thus. Pentachlorophenol will be
included as an Indicator Chemical in the quantitative risk
assessment.
2.1.3 Metal-Arsenic Complexes
There are no reports of the use of metal-arsenic
preservative at the Live Oak Site and analytic data does not
indicate contamination with these compounds. The EPA (in
February of 1983) detected copper in only one off-site well at
30 ppb. a value well below the Ambient Water Quality Criteria
(1 ppm. EPA. 19804) or proposed Recommended Maximum
Contamination Level (1.3 ppm. FR 50 46968. November 13. 1985).
Zinc was found in all veils sampled (at concentrations of 0.04
to 1.3 ppm) but adverse health effects from this compound have
never been identified (FR 50 46981. November 13. 1985). PELA
detected zinc and copper in the low ppb range (levels
consistent with natural occurence. Bond & Straub. 1973) in
lagoon water samples in June of 1983. At this time they found
no arsenic or chromium. The limited distribution of metals and
2-7
7711D D897 - -
-------
aresenic supports the claim that metal-arsenic preservatives
were not used at the Live Oak Site. It is therefore not
necessary to include these compounds among the Indicator
Chemicals.
2.1.4 Summary
Selected Indicator Chemicals ace listed in Table 2-2.
2.2 EXPOSURE ASSESSMENT
The Live Oak site is located west of the City of Live Oak.
Suwannee County, Florida, a town of 6700 people (1980
population. Bureau of the Census. 1983) in the North Central
portion of the state. There are private water supply wells in
all directions from the site. The area surrounding the' site is
rural. A defunct sawmill operation and a construction company
^
are located to the west and east of the site, respectively, but
there are no permanent residents at these locations. The
nearest permanent residents are located in a trailer court
north of the site, and in new houses to the south. The site is
posted with No Trespassing signs and locked cable gates have
been installed across roadways. The property is partially
fenced and. while the presence of unauthorized persons on site
was documented in the Remedial Investigation, this is a rare
occurrence due to the gates.
2.2.1 Exposure Pathways
At any site, humans may potentially be exposed to
contaminants in ale. water, oc solid media (soils, sediments or
sludges); directly, oc via the food chain. The coute of intake
may be by ingestion. inhalation, oc dermal absorption. Dermal
contact even without absorbtion may also be pectinent for the
indicator chemicals selected at the present site.
2-8
7711D D897
-------
Exposure to contaminants in aic is not a likely exposure
^athway for the Live Oak Site. The volatility of the indicator
chemicals is generally low, and the nature of the waste
precludes substantial entrainment. Finally, a majority of the
local population lives at distances from the site that make air
dispersion and dilution significant factors in diminishing
•xposure by this pathway.
Water
A significant potential exposure pathway off site is
ingestioo of water if the Ploridan aquifer were to become
contaminated as a result of leaching from the site. This
pathway will receive substantial analysis in the following
sections.
N
A pathway for direct contact with media onsite. which
would include surface water is developed below, in the section
on solid media.
Solid Media (soils, sediments, sludges)
A second pathway that may be significant for a subset of
the population is the exposure to contaminants by direct
contact with materials (sediments, soils and lagoon water)
while on site. As the only authorized individuals on the
property would be representatives of the PRPs or Agencies who
are aware of appropriate protective measures for the location.
the potentially at-risk population would be limited to
trespassers.
Food Chain
Food chain exposure via plant crops is an unlikely pathway
as surface run off from the site is captured by the lagoon or a
2-9
7711D D897
-------
swampy area to the west, the crops nearby have not required
irrigation (FTCH. 1987. pg. 92). Wildlife species which may
inhabit the site, and maybe hunted include raccoon, oppossum.
and bobwhite. However. PAH and pentachlorophenol do not tend
to accumulate appreciably in terrestrial animals, probably
because the compounds are extensively metabolized and
eliminated (EPA. 1980b). Thus, food chain exposure is
eliminated as a major pathway for exposure.
2.2.2 Exposure Point Concentrations
The analysis of plausible exposure pathways presented
above indicates that humans may potentially be exposed to
Indicator Chemicals by drinking wate'r or making direct contact
with materials on the property. The following section
determines what concentrations of Indicator Chemicals may exist
at these exposure points now or in the future. The data will
be utilized to determine potential human health impacts by
comparing the values to relevant Standards and Criteria (see
Section 2.2.3) and using the data for calculating human intakes
and applying dose-response relations (see Section 2.3). The
exposure point concentrations developed below are presented in
Table 2-3.
Drinking Water Pathway
The EPA reported no detection of "extractable organics" in
private veils during sampling in February of 1983. The new
private well, sampled by EEM in January. 1987 had no detectable
creosote constituents. The Remedial Investigation report
indicated no indicator parameters in ground water samples taken
in 1985 from monitoring wells at detection limits of 20 ppb for
carcinogenic PAH compounds and 110 ppb for pentachlorophenol.
In October 1986. fluoranthene and other low molecular weight
PAH (acenaphthene. dibenzofuran. fluorene, phenanthrene. and
anthracene) were detected in wells MW-4 and MW-8. Resampling
2-10
7711D D897
-------
TABLE 2-3
CONCENTRATION OF INDICATOR CHEMICALS*
AT EXPOSURE POINTS
LIVE OAK, FLORIDA
Drinking Water Pathway
Indicator
Analytical Data (ppb)
Private
at KV-8 Veils
Model (ppb)
atMV-8
1600 ft,
Veil
Carcinogenic PAH
Acenaphthene
Fluoranthene
Pentachlorophenol
Phenanthrene
(20)
45/40
2/2C
(100)'
8/11C
d
d
d
d
d
.006
.0003
.1
31
a. All concentrations given in ppb (raicrograras per liter). Analytic data
for MV-8 is from October 1986 sampling, reported in the Rl. Private
well data is from EPA, 1983 sampling information. Acenaphthene and
phenanthrene concentrations are given in addition to Indicator
Chemicals. Acenaphthene was considered because it has an "organoleptic1
AVQC. Phenanthrene concentrations were used to check the transport
model values versus analytic data.
b. Not detected by REM in latest 1986 sampling round. Value given in
parentheses is detection limit reported by laboratory.
c. Results of sampling In October. 1986 and January, 1987, respectively.
These values were estimated by the laboratory. The normal detection
limits are affected by the matrix, but generally are about 20 ppb.
d. Not detected by EPA. 1983. no detection limits given.
2-11
-------
TABLE 2-3 (Continued)
Direct Contact Pathway
Surface Water* Surface Soils andb
Indicator Ditch Lagoon Accessible Sediments
Total PAH 329 ppb 76 ppb 12,357 ppo
Carcinogenic PAH 14 ppb 14 ppb 992 ppm
Fluoranthene 69 ppb 25 ppb 2,141 ppm
Pentachlorphenol 94 ppb 53 ppb 5,363 ppm
a. Data from sampling round by NUS from ditch and lagoon (1 each)
February 9, 1983, except Pentachlorophenol In lagoon, which Is mean
of sampling results from February 9, 1983 and June 24, 1983.
b. Mean concentration of 37 samples of surface soil and ditch
sediments, as shown In Appendix B.
2-12
-------
in January, 1987 indicated again that low molecular weight PAH
were present in MW-4 and MW-8. although the compounds were
detected in lower concentrations than the previous sampling
round.
In order to determine what future impacts might occur as a
result of constituents at the Live Oak Site, a leachate and
ground-water transport modeling effort was undertaken. The
methods and results are detailed in an appendix to this report.
The PAH concentration inputs for the model were obtained
from data on the sediments and soils in the lagoon area (PELA
and EPA sampling of June 20-24. 1983) because this is the area
of the largest waste load and the most likely place for ground
water contamination to occur. Inadequate data were available
to model possible leaching of pentachlorophenol. so this value
is not reported.
The model utilized for determining the potential magnitude
N
of leaching from the lagoon material (the EPA Organic Leaching
Model, or OLM). proved valid only for soils beneath the sludge
layer in the lagoon. For sludge contaminants, the model
consistently predicted leachate concentrations in excess of the
solubility limits of the carcinogenic PAH. Thus, the
solubility limits were used as a sludge source term for
transport models. Transport in the unsaturated zone was
calculated by a one-dimensional analytical model. This model
provided mass fluxes as a source term to the saturated zone
model (the HPS model), which was developed by ERT. These
models (described in detail in Appendix A) were utilized to
predict potential drinking water impact of the lagoon in its
present condition.
The primary point of impact chosen was the nearest
downgradient well from the source. This it Well Number 15 on
sheet 10 in the Remedial Investigation report, which is
approximately 160C feet from the source. Although an actual
well was chosen, the models make the assumption that the point
is directly down gradient. Thus, the values predicted from the
models will be generally valid for 1600 feet down gradient.
2-13
77IIP PB97
-------
regardless of direction of ground water flow. Additionally.
solutions to the transport models were calculated using MW-8 as
an exposure point. This activity served a dual purpose. As
some materials were detected in MW-8 during the 1986 sampling
round, it was possible to compare model output to actual data
in order to ensure that the model was conservative (the only
Indicator Chemical found in this well was fluoranthene;
transport of an additional low molecular weight PAH.
phenanthrene. was modeled to provide an additional comparison
to analytical data). Additionally. MW-8 is only a few feet
from the lagoon, the area of highest waste load and potential
for leaching to ground water. It therefore may serve as a
point for making conservative, possibly "worst-case",
predictions of potential ground water contamination. Predicted
concentrations are given in Table 2-3.
Direct Contact Pathway
Of concern for direct contact are those soils, sediments
and surface waters to which project- personnel or trespassers
may have access. For the analysis, all surface water, the
ditch sediments, and surface soils were considered to be
accessible. The highest value and the mean concentration of
the indicator chemicals In each of the accessible media is
given in Table 2-3.
2.2.3 Comparison of Exposure Point Concentrations to
Standards and Criteria
Drinking Water Pathway
There are few ccitecia and no standards which are relevant
to and applicable to the exposure pathways at the Live Oak
site. The Ambient Water Quality Criteria (AWQC). which were
intended for use with surface water, may be used to predict
health risk from drinking water. These criteria are intended
2-14
7711D D897
-------
to minimize health risk from ingestion of water as well as
ingestion of aquatic species whose flesh may contain
contaminants partitioned from the water. By eliminating the
allowance for ingestion of aquatic biota, AWQC values may be
applied to aquifer water, where only the drinking exposure
occurs. The adjusted AWQC. taken from the Superfund Public
Health Evaluation Manual (pages 61-64) are given in Table 2-4.
An AWQC for a low molecular weight PAH, acenaphthene. is
also given in Table 2-4. This is the only non-Indicator
Chemical detected in the on-site wells for which an AWQC
exists. The AWQC for acenaphthene is not designed for
protection of health (there is little data on the health
effects of acenaphthene. which is why it was not chosen as an
Indicator Chemical). Bather it is an "organoleptic" limit,
indicating a concentration at which water might be impacted
relative to taste or smell.
The only other relevant value is a proposed Maximum^
Contaminant Level Goal for pentachlorophenol under the Safe
Drinking Water Act. which is also given in Table 2-4.
Chemical analysis of samples from private wells (EPA
sampling of February. 1983) revealed no Indicator Chemicals.
This would indicate that the potential human health impact of
indicator chemicals either on-site or off-site, via the
drinking water pathway is low. However, the EPA data is
difficult to interpret because detection limits were not
provided. Further, the AWQC for carcinogenic PAH is an
extremely small concentration and analytic methods with
detection limits in the range of the criterion are not
practically feasible. Thus, the detection limits for
carcinogenic PAH reported in the Remedial Investigation are
above the AWQC. It is therefore not possible to use the method
of comparison to criteria to determine the extent of public
health impact of carcinogenic PAH using either the EPA or the
Remedial Investigation data. If it is assumed that the EPA
achieved detection limits comparable to those reported for
2-15
7711D D897
-------
TABLE 2-4
STANDARDS AND CRITERIA FOR INDICATOR CHEMICALS
Ambient Water Recommended Maximum
Compound Quality Criteria^) Contaminant Level
Carcinogenic 0.031
PAH
Acenaphthene 20
Fluoranthene 188 ppb
Pentachlorophenol 1010 ppb 200 ppb
a. AWOC has been adjusted for ingestion of water only. All values
given in units of ppb (micrograms per liter).
b. The value is a proposed RMCL (50 FR 47002. November 13. 1985).
\
c. EPA has calculated cancer risks for various concentrations. The value
for a 1 chance in 100.000 risk is given here.
d. The actual Awgc for acenapthene is indicative only of an
"organoleptic" level at which some imapct on water taste or smell
might be affected. No health impact is expected at this level.
2-16
-------
analyses in the Ri for fluoranthene. pentachlorphenol. and
acenaphthene (20 ppb. 100 ppb. and 20 ppb. respectively), no
impact on human health or welfare would be expected, based on
the method of comparison to criteria.
In the Remedial Investigation study, fluoranthene was
estimated to be present in well MW-8 at a concentration of 2
ppb in October. 1986 and 11 ppb in January. 1987; and
pentachlorophenol was not detected (detection limits. 100 ppb
October. 1986). These concentration values are substantially
below the AWQC listed la Table 2-4. Provided that the
monitoring wells provide a generalized picture of ground water
quality at the Brown Wood Treating Site, no impact on public
health is indicated by the method of comparison to criteria for
these Indicator Chemicals. The October. 1986 sampling
indicated 45 ppb acenaphthene is present in MW-8. This does
not indicate any health impact, but does suggest the
possibility that the odor or taste quality of water could be
impacted at this "worst-case" location.
Data from the ground water modeling study predict higher
concentrations of fluoranthene and phenanthrene in MW-8 than
have been detected. The difference in values may be due to
conservatism of the model. Therefore, the predicted
concentrations appear to be conservative values for estimating
the current impact of Indicator Chemicals potentially leaching
to ground water. Model output for total carcinogenic PAH and
fluoranthene at the 1600-foot well are 0.0003 and 0.1 ppb.
respectively (pentachlorophenol values could not be
calculated). Concentrations predicted for on site location.
MW-8 are 0.006 ppb and 3 ppb foe carcinogenic PAH and
fluoranthene. respectively. The carcinogenic PAH values in
each of these locations are below the concentrations set in the
Ambient Water Quality Criteria Documents as values where
carcinogenic risfc would be relatively low. The predicted
fluoranthene concentration at the off-site and on-site well
locations are substantially below the criterion. Thus, by the
method of comparison of modeling results to criteria, impact of
the Indicator Chemicals from the site appears very low.
2-17
7711D D897
-------
Direct Contact Pathway ^
Standards oc Criteria foe allowable concentrations in
soils, sediments, oc lagoon water ace not available. The
significance of contamination in these media have been assessed
using standard risk analysis procedures.
2.2.4 Estimation of Chemical Intakes
Chemical intakes have been calculated with the aid of
exposure scenarios relevant to the pathways identified above.
Plausible mechanisms by which intake may occur has been
outlined and an estimate of the magnitude of the intake has
been calculated from standard- values for human activities
leading to the exposure (e.g. volume of daily fluid intake).
Intake values have been converted to units of micrograms of
Indicator Chemical per kilogram of body weight per day. to make
them compatible with the dose-response relations developed in
the subsequent Section.
Drinking Water Pathway
Most of the population of Live Oak consumes water from
wells in the Floridan aquifer. Although the general flow of
ground water in the area of the site is away from populated
areas, the possibility has been considered that exposure to
constituents could occur if ground water were contaminated by
materials from the Live Oak Site. To calculate the magnitude
of exposure (in micrograms of constituent pec kilogram body
weight pec day) one multiplies the amount of drinking water
consumed daily (assumed to be 2 litecs pec day in adults and 1
litec pec day in childcen. EPA. 1986) by the concentration of
Indicator Chemicals predicted at the exposure point and
corrects for body weight (assumed to be 10 kg for small
children and 70 kg for adults):
2-18
7711D D897
-------
c x we
Water Ingestion. Daily dose (ug/fcg day) - •
o W
where.
GW - Concentration of constituent in water (vg/L)
WC « Water Consumption (1 L/day for children; 2L/day for
adults)
BW . Body Weight (10 kg for children. 70 kg foe adults)
No Indicator Chemicals have been detected in ground water
at the off-site exposure points. For the on-site location,
only fluoranthene was detected in analytical samples. Thus.
only model predictions for carcinogenic PAH (as shown in Table
2-3) were used. In the case of fluoranthene. model predictions
were used for the off-site location, and both analytic results
and 'model predictions were used for on-site assessment. " This
is equivalent to assessing the health risk now and in the
future, given the time projection of the transport model.
Because neither analytic data nor model prediction data were
available for pentachlorophenol. the intake (daily dose)
equation was applied to the detection limits of the RI study.
The values represent a "worst case" for intake because the
actual concentrations are below the detection limit, if they
are present at all. For the off-site location, the intake
prediction is even more conservative, because dilution of
pentachlorophenol would occur between the entry point into the
aquifer at the site and ultimate exposure points.
Both child and adult intake values were calculated for
pentachloeophenol and fluoranthene. because it is presumed that
toxicity could appear aftec a relatively short period of
exposure. In contrast, the value that must be applied to dose
response relations for carcinogenic toxicants is a lifetime
daily dose. Because most of a lifetime is lived as an adult.
only an adult intake was calculated for carcinogenic PAH. The
calculated values are presented in Table 2-5.
2-19
7711D D897 ..
-------
TABLE 2-5
PREDICTED INTAKES OF INDICATOR CHEMICALS
DRINKING WATER PATHWAY
LIVE OAK, FLORIDA
On-Site
-------
TABLE 2-5 (Continued)
Predicted Intakes of Indicator Chemicals
Direct Contact Pathway
Live Oak, Florida
vg/kg day
Carcinogenic PAH 0.0076
Pluoranthene 0.58
Pentachlorophenol 0.23
Intake of materials by children visiting the site once monthly and
consuming 55 mg soil per visit for five years. All values in units of
micrograms constituent per kilogram body weight per day (vg/kg day).
Values are lifetime average daily doses for carcinogenic PAH and
average daily doses for fluoranthene and pentachlorophenol.
2-21
-------
Direct Contact Pathway
Because the site is not completely secured, it is possible
that people might trespass and make direct contact with
contaminated media onsite. There are two aspects to such an
exposure. People might experience a systemic exposure as the
result of inadvertent ingestion of materials clinging to hands
or articles which may be placed in the mouth. Contaminated
soils and ditch and lagoon sediments (those not covered
continually by water) are the media of concern for this
scenario. People night also experience a dermal effect as a
result of contact with contaminated materials. Soils.
sediments and surface water are all of concern for this
scenario.
Ingestion of materials is normally considered to be of
concern for children only. It has recently been suggested that
a child between the ages of two and six years may be assumed to
ingest 55 mg of soil per day while outdoors (Clausius.
et al. 1987). The question for a rural site such as the Live
Oak property is whether children of this age would ever reach
the site and. if so. how often. It is unlikely that frequent
visits by this age group would occur in the present condition
of the site. For the purposes of quantitation. intake
predictions will be made for children visiting the site once
monthly from age two through six years. This seems
conservative, given the distance of the site from residences.
the limited access to the property, and the fact that the site
is not on the route to a school, playground or other
destination attractive to children. Unknown future uses of the
property, unless otherwise restricted, might include more
frequent presence of children. Notice that because a lifetime
average daily dose is required for ascertaining chronic risks
of carcinogenic PAH. an adjustment must be made for the
proportion of a 70 year lifetime in which exposure occurs.
This correction is not applied for fluoranthene or
pentachlorophenol. where toxicity may be expressed after
2-22
7711D D897
-------
shorter periods of dosing. Children in this age group are
assumed to have an average body weight of 17 kg (EPA. 1986)
Expressed mathematically the intake would be:
C x SC
Soil ingestion daily intake (ug/kg day) - rr:— x f x d
ow
where.
C • Constituent concentration in soil (ug/fcg)
SC - Soil consumption (0.000055 kg/day)
BW - Body weight (17 kg)
f • Frequency of exposure (1 day/30 days)
d - Duration of exposure (used for carcinogens only, to
obtain a lifetime average daily intake; 5 years/70
years)
This formula has been applied to the mean soil
concentration data to model the potential impact of an
individual taking a "random walk" around the site and picking
up contaminated materials from several places. The results are
presented in Table 2-5.
A more plausible scenario for health impacts from direct
contact with contaminated material at the Live Oak Site is
toxic effects on the skin. This scenario is more plausible
because it would affect adults as well as children, and
children are less likely to be on site. Further, this is an
acute effect of the Indicator Chemicals. Thus, it is not a
function of frequency of contact. It is not really necessary
to calculate an "intake" for this type of topical effect. In
Section 2.3.4 concentrations of PAH (total) and
pentachlorophenol that might be associated with dermal effects
are developed.
2-23
7711D D897
-------
2.3 TOXICITY ASSESSMENT ;-
2.3.1 Carcinogenic PAH
PAH are formed as a result of elevated temperature
processes such as fires, petroleum-synthetic mechanisms in the
deep subsurface, and anthropogenic activities such as operation
of internal combustion engines and incineration or other
combustion of refuse, forest, and agricultural products. The
largest contribution of PAH to the environment are the man-made
combustion sources mentioned above. PAH are not generally
intentionally synthesized, but are obtained by refining natural
materials for use as fuels, lubricants, preservatives and
starting materials for petrochemical manufacture.
Only certain PAH have been identified as having the
potential to cause cancer. Such carcinogenic PAH generally
tend to be high in molecular weight, have at least 3 aromatic
rings (usually more), have low water solubility, are easily
absorbed by humans, and have very low acute toxicity. PAH have
not been unequivocally demonstrated to be carcinogenic in
humans. However, by extrapolation from health effects in
individuals who smoke, and from animal data on certain PAH
compounds there is reason to believe that some PAH are
carcinogenic in humans. The U.S. EPA "weight-of-evidence"
system to classify carcinogen data has been described
previously (Section 2.1.1). There are no A or B-l level
carcinogens among the PAH detected at the Live Oak Site. Six
PAH compounds found at site have EPA ratings of "probable"
(B-2) to "possible" (C) human carcinogens (EPA. 1986) and have
been chosen for detailed risk analysis in the present
assessment.
In determining one set of criteria, the AWQC, EPA (1980a)
used animal dose-response data for benzofa] pyrene to establish
a criterion for all carcinogenic PAH (summed quantities).
This approach is very conservative, because the carcinogenic
potency of benzo[a]pyrene is probably greater than other PAH.
2-24
7711D D897 ...
-------
The approach is also an oversimplification because the potency
of an individual PAH may change according to the route of
exposure and the presence of other compounds in the exposure
mixture. Applying dose-response data from benzo[a]pyrene to
other PAH is. nonetheless, the only method currently more
available.
Studies on chemical carcinogenesis suggest that, for some
compounds, no threshold foe the effect exists. That is.
certain carcinogens, even in extremely small doses, will pose
some risk of cancer. This assumption is incorporated into the
cancer dose-response assessment for PAH. Neal and Rigdon
(1967) gave mice feed containing between 1 and 250 milligrams
per kilogram (ppm) benzofaJpyrene and found that more treated
animals developed stomach tumors than the control group. The
increased tumor incidence was dose dependent. After adjusting
the doses to correct for presumed differences in mouse versus
\
human metabolism, this data was used by the EPA Carcinogen
Assessment Group in a computer program which calculates the
upper 95 percent confidence interval on the slope of a dose
response line fitted to an equation modeling the assumed no
threshold, multistage mechanism of chemical carcinogenesis.
The value, called a "potency slope." is 0.0115 per microgram
per kilogram body weight per day for ingestion exposures to
benzo[a]pyrene. The potency slope indicates that an individual
consuming 1 microgram benzo[a]pyrene per kilogram body weight.
daily, for life, might have a risk of contracting cancer of
about 1 chance in 100 over that of the non-exposed individual
(note that this if an upper bound on the estimate, the actual
risk is more likely to be lower). Because the dose-response
relation is presumed to be linear, multiplying the predicted
lifetime daily intake of carcinogenic PAH by the potency slope
will give an upper bound estimate of excess cancer risk from
exposure to constituents at the Live Oak Site (by the routes
previously outlined).
2-25
7711D D897
-------
2.3.2 Fluoranthene
Fluoranthene is among the PAH compounds which has been
demonstrated to lack carcinogenic activity both in skin
painting tests and by subcutaneous injection. Fluoranthene has
also been shown to lack mutagenic activity in the Ames test
(mutagenic activity is often related to carcinogenicity).
Thus, risk analysis foe fluoranthene will be conducted on other
potential toxicities.
The acute toxicity of fluoranthene appears to be low
(LDgQ in cats 2g/kg; Smyth, et al. 1962 as quoted in EPA.
19BOa). The AWQC for fluoranthene (EPA. 1980a) was calculated
on the basis of an extremely limited study by Hoffman, et al.
(1978). It is questionable whether the data are truly
sufficient to establish a criterion. The data are applied to
this risk analysis with skepticism. Hoffman, et al. applied 50
microliters of a 1% solution of fluoranthene to the backs of
mice 3 times weekly for one year and saw no mortality in the
animals (for up to 15 months). If one presumes that the entire
amount of fluoranthene was absorbed and that the average weight
of a mouse is 35 g. one may calculate a "no-effect level" of
6.12 mg/kg body weight. Presumably this type of toxic effect
has a threshold. That is. there exists a dose below which no
individual will respond with a toxic effect. It is common
practice to apply a "safety factor" to an experimentally
determined no-effect level to provide reasonable certainty that
a sub-threshold foe toxic effects is obtained. Often the no
effect level is divided by 10 to allow for possible differences
in sensitivity between man and animals and another factor of 10
to correct foe possible differences in sensitivity among
humans. In the case of fluoranthene. an additional 10 fold
safety factor was applied by the EPA due to the small amount of
data available in the study. Thus, the corrected no effect
level, called an "Acceptable Daily Intake" (ADI) for
fluoranthene was determined to be 6'. 12 ug/*g day. EPA
calculations of fluoranthene burdens from other sources
2-26
7711D D897 ...
-------
(.016 mg/man day from diet and .0001 mg/man day from the air)
do not appreciably change this value. The SPHEM gives no
acceptable intake value for fluoranthene. Thus, the ADI
developed in the AWQC will be used for assessing risk from
fluoranthene at the Live Oak site. Applying the threshold
concept, if the exposure is below the ADI. no risk would be
expected. If exposure is above the ADI risk may be present.
This risk is not quantifiable, but it may be qualitatively
stated that the greater the exceedance of the ADI. the more
likely it is that a toxic effect will be seen.
2.3.3 Pentachlorophenol (PCP)
The sole use of PCP is as an antimicrobial preservative of
wood products. PCP is well absorbed by the dermal, inhalation,
and ingestion routes of exposure. Acute toxic episodes in
humans have been reported after dermal and inhalation exposures
(EPA. 1985b). The acute toxic effects of sweating, fever, and
rapid heart rate are probably related to the ability of PCP to
interrupt energy metabolism. Other acute toxic effects of PCP
are related to the irritative properties of the compound and
include reddening and painful sensations of skin immersed in
the compound, irritation of the throat after drinking
contaminated water (12.5 milligrams per liter), and congestion
of eyes and nasal passages. Effects reported in humans with
possible chronic exposure to PCP may include liver, kidney.
bone marrow damage, and infections which may be related to poor
immune function.
To establish a dose-response relationship for the systemic
effects of PCP. the EPA has relied on a study conducted by
Schwetz. et al. (1978). which reports on chronic toxic effects
and the reproductive ability of rats. Schwetz. et al. noted
that in the reproductive study when animals were fed 0. 3 or
30 milligrams per kilogram body weight (ppm) per day of PCP
prior to and during the gestation period, animals in the
highest dose group only had a lower percentage of live births
2-27
7711D D897
-------
than controls. Each dose group contained 10 males and 20
females. The offspring of the high dose mothers were lower in
weight and survived less often than untreated animals. This
would make the lower dose (3 milligrams per kilogram (ppm) per
day) a "no effect" level.
The two year chronic study contained 5 different dose
levels (0. 1, 3. 10. 30 milligrams PCP per Kilogram per day).
Each dose group contained 25 rats of each sex with 2 additional
cats pec group maintained foe tissue specimens used foe
chemical analysis. The no-observable-adverse-effeet-level
(NOAEL) foe PCP. based on chronic toxicity, was 10 milligrams
pec kilogram pec day among males and 3 milligrams per kilogram
per day among females.
The EPA had used the standard practice of applying a 100
fold safety factor to the NOAEL dose to arrive at an acceptable
intake, subchconic (AIS) of 0.03 milligrams per kilogram body
weight pec day. No further uncertainty factors were applied to
the AIS to derive the acceptable daily intake, chronic (AIC):
it is also 0.03 milligrams per kilogram body weight pec day.
Presumably toxic effects such as those observed in the PCP
experiments here have a threshold. That is. there exists a
dose below which no individual will respond with a toxic
effect. The purpose of the safety factor (which is a factor of
10 to allow foe possible diffecences in sensitivity between man
and animals and another factor of 10 to coccect foe possible
diffecences in sensitivity among humans) is to pcovide
ceasonable ceetainty that the AIC will be below the thceshold
foe toxic effects. The AIC decived by the EPA will be used foe
the present assessment of PCP impact at the Site. If predicted
intake is below this value, no cisk of toxic effect is
expected. If intake is greater than the AIC. a toxic effect
may occuc. The lacgec the intake value is ovec the AIC. the
greater likelihood of a toxic effect.
2-28
7711D D897
-------
2.3.4 Noncarcinogenic Dermal Toxicity of Indicator
Chemicals
The values presented above are derived from studies of the
systemic toxicity of Indicator Chemicals. As has been
suggested in previous sections of this report, it is .possible
that the most likely exposure for persons on site at the Live
Oak Site may be brief dermal contact with contaminated soils.
Such an exposure might not result in significant systemic
absorption of compounds or might not be of sufficient frequency
for a chronic systemic effect to occur. However, it could
result in dermal irritation or other effects. Some suggested
concentration values, that may result in these effects, are
derived from the data presented below.
Creosote Constituents
N
A number of papers have noted that various PAH compounds,
when applied to human skin, will produce an enhanced sunburn
reaction on exposure to ultraviolet light. The phototoxicity
effect is usually reversible, but could cause transient
problems for workers or trespassers at the Site. In 1980. an
exchange of correspondence between Urbanek and Walter in the
Journal of the American Academy of Dermatology indicated that
0.25 percent anthracene dissolved in petroleum did not produce
phototoxicity but did cause a stinging sensation and transient
hive-like reaction in a small number of patients while
Sstu.-r»«*
0.1 percent^was without any toxic effect. For the purpose of
the current assessment. 0.1 percent (1.000 ppm) will be
considered a no-effect level for all PAH. This value may be
overly conservative la that the availability of PAH adsorbed to
soil particles may not be as great as that in a homogeneous
solution in medicinal preparations. The value may be a poor
estimate toxicity in that PAH other than anthracene are present
in soil at the Live Oak site. The other PAH have not been
studies, but may be more or less photoreactive than anthracene
2-29
7711D D897
-------
alone. It is notable, however, that a 1 percent (1QVOOO ppm)
coal tar solution was only minimally effective in producing a
phototoxic reaction (Tannenbaum. 1975). This preparation would
be expected to contain a variety of PAH compounds.
Pentachlorophenol
Deichman. «t al (1942) reported on the effects of PCP
administered dermally in animals. la this study, application
of 10 ml of 1 percent PCP in mineral oil (10.000 ppm) for 4
hours was without local dermal effect in 21 days of dosing.
Solutions of 5-10 percent gave positive or negative dermal
results depending on the volume of material applied and the
vehicle in which the PCP was dissolved. Using a 10 fold safety
factor for animal to human extrapolation on the
10.000 milligrams per Kilogram (ppm) no effect level for dermal
irritation would indicate an acceptable concentration fo-r
protection from this effect would be 1.000 milligrams per
kilogram (ppm). This value may be overly conservative in that
the availability of PCP adsorbed to a soil or sludge particle
may not be as great as that in a homogeneous solution in
organic solvents.
2.4 RISK CHARACTERIZATION
Table 2-6 presents numerical estimates of carcinogenic
risk and other toxic effects for the scenarios developed in
this report.
It i« not possible to calculate a drinking water cancer
risk from actual data because of the lack of detection limit
information for the sampling round on private wells. The
modeling effort for the potential future condition of the
aquifer if no action is taken at the site indicates that a
cancer risk of slightly less than 1 chance in 10.000.000 could
be incurred by drinking water from the aquifer if leaching
2-30
7711D D897
-------
TABLE 2-6
HEALTH RISKS FROM INDICATOR CHEMICALS
FOR PRESENT CONDITION OF SITE
LIVE OAK, FLORIDA
Carcinogenic PAH*
Source Risk
Ingestion of Drinking water
On Site 2.3 chances in 1,000,000
Off Site 9.9 chances in 100,000,000
Ingestion of PAH in Surface Soils 8.8 chances in 100,000
a. Calculated for the sum of carcinogenic PAH, assuming all
have equivalent potency as benzo(a)pyrene
(potency slope - .0115/ug/kg day).
2-31
-------
TABLE 2-6 (Continued)
HEALTH RISKS FROM INDICATOR CHEMICALS
FOR PRESENT CONDITION OP SITE
LIVE OAK, FLORIDA
Health Impacts of Fluoranthene
Source Predicted Intake/API (a)
Ingestion of Drinking Vater Child Adult
on site 0.03-0.05 0.01-0.015
off site 0.002 0.0005
Ingestion of Fluoranthene in
Surface Soils 0.09
a. A value of less than one Indicates Intake is below ADI
(6.12 vg/kg day) and virtually no risk is likely.
Values above one indicate possible risk of toxic effect.
2-32
-------
TABLE 2-6 (Continued)
HEALTH RISKS FROM INDICATOR CHEMICALS
FOR PRESENT CONDITION OF SITE
LIVE OAK, FLORIDA
Health Impacts of Pentachlorophenol
Source
Predicted Intake/API (a)
Ingestion of Drinking Water
Child
Adult
on site
<.33
off site
<.33
Inge.stion of Pantachlorophenol
in Surface Soils
0.008
a. A value of less than one indicates intake is below ADI
(30 vg/kg day) and virtually no risk is likely, values
above one indicate possible risk of toxic effect.
2-33
-------
occurs. For the onsite water location, the risk is
approximately 2 chances in 1.000,000. It should be pointed out
that this risk estimate is an upper bound for the following
reasons:
• The potency slope value used to estimate risk is the
95% upper confidence bound on the dose response
relation. Thus, the actual risk is likely to be
lovec.
• The concentrations are predicted for the closest
existing off sit* well and a monitoring well
extremely close to a potential source of
contamination.
Cancer risks from direct contact (ingestion) with
contaminated soils onsite are greater in magnitude than the
drinking water scenario at slightly less than 1 chance in
10.000. but also less likely to occur. The location and
characteristics of the site preclude regular visits by small
children. Although this scenario does not represent a probable
situation for evaluating the current impact of the site, it
should be considered when planning future uses of the property.
No exposure scenario revealed a health impact from
fluoranthene. All predicted intakes were less than the ADI.
indicating virtually no risk of a toxic effect from this
compound.
Ground water analysis indicated pentachlorophenol was
undetectable at a limit of about 100 ppb. Even if
pentachlorophenol were present at the level of detection, no
health risk would be expected as ingestion of this level
pentachlorophenol would provide a dose that is still below the
acceptable daily intake.
A possible risk not quantified in the Tables is the
potential for dermal effects from direct contact with surface
materials on site. As with other scenarios, it is difficult to
quantify the likelihood that an individual would be on site and
2^34
7711D D897 " '
-------
make contact with contaminated soils, sediments or surface
water. However, because dermal toxicity is an acute effect.
requiring only a single visit to the site, one would
intuitively rate this as a more likely scenario than those
which require regular presence on the property. In the
Toxicity Assessment section it was conservatively estimated
that 1,000 ppm of total PAH could cause photosensitization and
1.000 ppm pentachlorophenol might cause dermal irritation.
There are surface sediment and soil locations where these
values are exceeded. However, surface water concentrations are
substantially below this concentration. Thus, dermal effects
from contact with solid media seem plausible, but surface water
does not present this hazard.
2-35
77IIP D897
-------
3. DEVELOPMENT OF PERFORMANCE GOALS ;-
In compliance with the Supecfund Public Health Evaluation
Manual guidance on source control remedial alternatives, an
analysis determining "acceptable" residual levels of chemicals
is presented here.
No evaluation of fluoranthene and pentachlorophenol were
done as the baseline assessment indicated no present or future
health impacts of the compounds. The low level of carcinogenic
risk for the drinking water scenario also precluded the
necessity of performance goals based on this limited health
impact.
It was determined in Chapter 2 that the.most likely health
risk of the Live Oak Site was the potential for non-
carcinogenic, acute dermal effects produced by direct contact
with PAH and PCP in surface soils and sediments. Thus, a
concentration limit should be put on PAH and PCP in soils with
which people might make direct contact. Total PAH should not
exceed 1000 ppm in surface soils in order to avoid possible
phototoxic reactions. Pentachlorophenol should not exceed 1000
ppm to avoid dermal irritation. As detailed in the Feasibility
Study Report, the preferred remedial alternative has been
designed to meet this performance goal. All materials in the .
lagoon, ditch, and soil containing concentrations in excess of
1000 ppm PAH or PCP will be removed.
A calculation may also be made foe the risk reduction
achieved by the preferred alternative foe the soil ingestion
scenario, although it is a less likely exposure, given the
current use of the site. In Section 2.4. the cancer risk
associated with the mean concentration of surface soil at the
site was estimated for a scenario in which children ingested
soil at the site once a month foe 5 years. Because the risk
estimate was celated to soil concentration in a lineae mannec.
the cisk ceduction associated with the decreased average soil
concentration at the site may be easily calculated using the
equation developed on page 2-22 and the potency slope
(0.0115/mg/kg day). If one remediated soil containing greater
3-1
7714D D897 • •
-------
than 1000 ppm creosote substances in order to limit the risk of
transient dermal effects, the average across the site of .
carcinogenic PAH concentrations remaining in the soil would be
approximately 14 ppm. This value was calculated by
substituting the concentration of carcinogenic PAH in the next
strata sampled (if available) for each sampling point where
values of PAH or PCP were in excess of 1000 ppm. The
calculations are presented in Appendix B.
As shown in Table 3-1. 14 ppm of carcinogenic PAH would be
associated with a substantially reduced cancer risk due to
ingestion of soil; 1.2 chances in 1.000.000.
3-2
7714D D897
-------
TABLE 3-1
COMPARISON OF HEALTH RISKS
BEFORE AND AFTER
PREFERRED REMEDIAL ALTERNATIVE
LIVE OAK, FLORIDA
Acute
Direct
Contact
Soil
Ingestion
Drinking
Water,
Off-Site
Drinking
water,
On-Site
PAH (a)
PCP (b)
Before
8.8
0.099
0.23
Before
Fluoranthene (b)
Before
After
possible unlikely possible unlikely included in total PAH
0.12
0.008
NC (c) 0.09
NC (c)
NC (c) <0.1-<0.33 NC (c) 0.002-0.0005 NC (c)
NC (C) <0.1-<0.33 NC (c) 0.03-0.015 NC (c)
(a) PAH risk is calculated as cancer risk per 100,000 chances, except
in the case of acute direct contact, where the likelihood of
dermal photosensitlzation is estimated qualitatively.
(b) PCP and Fluoranthene risk is calculated as the proportion of the
AIC represented by the estimated intake except in the case of
acute direct contact, where the likelihood of dermal irritation
is estimated qualitatively. As noted, fluoranthene Is among the
total PAH assumed to have photosensitizing properties and is
Included under the PAH category.
(c) Where small risk was estimated for the current condition of the
site, risk values for the remediated site were not calculated
(NC).
3-3
-------
4. • SHORT-TERM RISKS OF REMEDIAL ALTERNATIVES
Other than the No Action alternative, the remedial actions
chosen for screening for the Live Oak site are source coni.rol
measures. The alternatives are off-site disposal, off-site
incineration, on-site incineration, solvent washing techniques.
and combinations of these techniques. Short-term health risks
from these activities are discussed in a qualitative manner
•
belov.
4.1 Dermal Toxicity
The possibility that individuals might make contact with
materials at the Site has been covered in the baseline
assessment. It was stated that site access was difficult and
it was unlikely that individuals would be present at the Kite
with any frequency. When remedial measures are implemented.
the number and frequency of personnel on the site will
increase. This may increase the short-term risk due to direct
contact with contaminated materials. However, it should be
recognized that an approved Health and Safety Plan will be
implemented. That plan will provide for minimizing prolonged
contact with media containing more than 1000 ppm
pentachlorophenol or 1000 ppm creosote substances and frequent
cleansing of unprotected skin.
4.2 Transportation Risks
Off-site remedial activities involving transportation
(e.g.. off-site disposal, off-site incineration, transportation
of recovered material) may be associated with vehicular
accidents. If vehicles involved in the remedial action are
subject to the same risks as Florida motorists in general.
statistics would indicate that 3.3 traffic deaths will occur
for every 100.000.000 miles driven (Land, et al.. 1985). The
mileage estimated from remedial actions (maximum, about 350.000
miles) would be associated with a population risk of accidental
4-1
7713D PD-897 "
-------
death of about 1 chance in 100 (n.b this is a population risk
which cannot be compared to the individual risks derived for
carcinogens previously - if a comparison is pertinent, the
transportation risk would be roughly equivalent to the
population cancer risk of the City of Live Oak (6700 residents)
if the average individual risk was 2 in 1.000.000).
A second aspect of vehicle accident risk is the human
health impact of dispersion of contaminated materials. A
contingency plan will address this possibility if off-site
remedial actions are chosen.
4.3 Air Emissions .
If incineration alternatives are chosen, there is a
possibility that emissions from this technology could have a
health impact. Adequate remediation design is required to
minimize this risk.
4-2
7713D PD-897
-------
REFERENCES
Agency for Toxic Substance and Disease Registry (1986) Memo to
Gael Hickam (EPA) from the ATSDR Office of Health
Assessment ce Bayon Bonfouca Supecfund Site, SI-86-035B.
Bond. R.G. and C.P. Straub, 1973. Handbook of Environmental
Control Vol. Ill: Hater Supply and Treatment CRC Press,
Boca Raton, pp. 36.
Bureau of the Census. 1983. County and City Data Book.
Clausius, P.B.. B. Brunkceef. and J.H. van Wijen (1987), A
Method for Estimating Soil Ingestion by Children. Int.
Arch. Occup. Environ. Med. 59: 73-83.
Craun and Middleton. 1984. Handbook on Manufactured Gas Plant
. Sites Prepared for the Utility Solid Waste Activities
Group and The Edison Electric Institute. ERT project
Number PD-215.
Deichman. H.. W. Machle. K.V. Kitzmiller. and G. Thomas. 1942,
Acute and chronic effects of pentachlorophenol and sodium
pentachlorophenate upon experimental animals. Phatmacol.
exp. Ther. 76:104-117.
EPA. 1980a, Ambient Water Quality Criteria for Fluoranthene.
EPA 440/5-80-049.
EPA. 1980b. Ambient Water Quality Criteria for Polynuclear
Aromatic Hydrocarbons. EPA 440/5-80-069.
EPA. 1980c. Ambient Water Quality Criteria for
Pentachlorophenol. EPA 440/5-80-065.
EPA. 1980d. Ambient Water Quality Criteria for Copper.
EPA 440/5-80.
EPA. 1980e. An Exposure and Risk Assessment for
Pentachlorophenol. EPA 440/4-81-021.
EPA. 1984. Health Effects Assessment for Coal Tar. EPA
540/1-84-024.
EPA. 1986. Superfund Public Health Evaluation Manual. EPA
540/1-86/060.
EPA. 1986b. Record of Decision for United Creosote Site. Conroe.
Texas.
EPA. 1985a. Guidance on Feasibility Studies Under CERCLA. EPA
540/G-85-003.
7712D PD-897
-------
REFERENCES (Continued)
EPA. 19855. Draft Drinking Water Criteria Document for
Pentachlorophenol. EPA 600/X-84-177-1.
FTCH. (1987), Report on the Remedial Investigation. Brown Wood
Preserving Site. Live Oak. Florida.
Hoffman et. al.. 1978. Fluoranthenes: Qualitative
determination in cigarette smoke, formation by pyrolisis.
and tumor initiating activity. J. Nat. Cancer Tnst.
49:1165.
Lane. H.U. et. al.. (editors). 1985. The World Almanac and Book
of Facts. Newspaper Enterprise Association. Inc. New
York. pp. 780-781.
NIOSH. 1985. Pocket Guide to Chemical Hazards. U.S. Department
of Health- and Human Services. Publication No. 78-210.
Neal J. and R.H. Rigdon (1967) Gastric tumors in mice fed
benzo(a)pyrene: a quantitative study. Tex. Rep. Biol.
Med. 25:553.
>
Sandmeyer. E.E., 1981. Aromatic Hydrocarbons in Patty's
Industrial Hygeine and Toxicology. 3rd Ed.. Vol. 2A. G.D.
and F.R. Clayton, eds. Wiley Interscience. New York. pp.
3253-3431.
Schwetz. J.F. Quast. P.A. Keeler. C.G. Humiston. and R.J.
Kociba. 1978. Results of 2-year toxicity and reproduction
studies on pentachlorophenol in rats. In
Pentachlorophenol: Chemstry. Pharmacology and
Environmental Toxicology. K.R. Rao* ed. Plenum Press.
N.Y. pp. 301.
Tannenbaum. L. J.A. Parrish. M.A. Pathak. R.R. Anderson, and
T.B. Fitzpatrick. 1975. Tar photoxicity and phototherapy
for psoriasis. Arch,, Dermatology 111:467-470.
Urnabek. R.W.. 1980. Side effects of anthracene. J. An. Acad.
Dermatol. 2:240.
Walter J.F.. 1980. Side effects of anthracene: Reply. J. Am.
Acad. Dermatol. 2:240-241.
77I2D PD-897
-------
APPENDIX A
GROUND WATEB
CONTAMINANT MODELING
-------
APPENDIX: GROUND WATER CONTAMINANT MODELING
A.I Modeling Techniques and Assumptions
A.1.1 Leachate Model
A model was used to attempt to determine concentrations of
contaminants in leachate. The EPA OLM (Organic Leachate Model)
model was developed for delisting evaluations by statistically
fitting actual leachate and waste mass concentrations (EPA
1986). The model is expressed mathematically as:
C. - 0.00211 C0-678 S°-373 (A-l)
L
Where;
CL • contaminant concentration in leachate N
C • contaminant concentration in waste
S • solubility of contaminant
As will be seen in section A-2. the OLM model proved
invalid for some of the contaminants at the Live Oak site, for
which it was replaced by the conservative assumption that these
contaminants were present in leachate at their solubility limit.
A.1.2 Ground-Water Models
A. 1.2.1 Unsaturated Zone
An analytical solution (Battelle. 1986) was used to
calculate the moisture content as a function of steady state
flux in the unsaturated zone between the lagoon and the
limestone aquifer. The unsaturated zone moisture content. 9.
was used in calculations to estimate the retardation of the
contaminants as they passed through the unsaturated zone.
7705D PD-897
-------
The following expression was derived from Darcy's equation
for steady state flow and soil characteristic relationships
described by Campbell (1974).
6 • (vTTT) ®8 (A-2)
Where:
6 • the unsaturated moisture content
q • the steady state flux (cm./min.)
Ksat « saturated hydraulic conductivity (cm./min.)
6s • saturated moisture content
m • l/(2b +3) vhece b is negative one times the slope of
the log-log plot of the matric potential, Vm. versus
6.
\
The equation is based on the following assumptions.
• flow is one-dimensional in the vertical direction
• water flow is steady state
• water table conditions exist at the lower boundary
• the upper boundary condition is constant flux
• soil characteristics (6 versus inn and hydraulic
conductivity versus ^rm) are constant with depth and
• the hydraulic gradient is vertically down and equals
unity (drainage is due strictly to gravity and
capillary forces are dominated by gravitational
forces).
The moisture content in the unsaturated zone. 9. was
used in the calculation of the retardation factor. R. for each
constituent from the following equation.
R . x „, P(Koc)U9
-------
Where:
Koc * constituent-specific organic-carbon partition
coefficient (listed in EPA 19B2)
fpc • fraction of organic-carbon in the soil
p • built density of the soil (gm/cm3)
The retardation factor represents the mobility of the
constituent traveling through the soil in an aqueous solution.
relative to the mobility of water. The retardation factor was
used in estimating the mass flux of each constituent into the
saturated zone.
The mass flux (gm/year) for each constituent was
calculated using the following relationship
... »,,., (Leachate Concentration) x (Infiltration Rate)
Mass Flux - Retardation Factor * (A-4)
A.1.2.2 HPS Model
The HPS model is a three-dimensional, transient.
analytical solution of the convective-diffusion equation for
solute transport in ground water (Galya. 1987). The model
incorporates the following features:
• Advection in one dimension with dispersion in three
dimensions.
• Simulation of a horizontal plane source in order to
correctly model transport from landfills or waste
lagoons.
• The ability to perform time-varying as well as
steady-state predictions.
• Simulation of the source strength as a mass flux
input which can vary with time.
• The effects of retardation and degradation explicitly
included in the formulation.
7705D PD-897
-------
Model Formulation
Transport and first-order degradation of contaminants ir.
an aquifer can be described by the three-dimensional
convective-diffusion equation:
.
R Z R 2 R 2 6
where
C » concentration of .contaminant
u -ground-water velocity in x direction
t » time
DX • dispersion coefficient in x direction
" axu s
ax • dispersivity in x direction
D « dispersion coefficient in y direction
y
» «yu
a * dispersivity in y direction
D « dispersion coefficient in z direction
• a u
z
az m dispersivity in z direction
X • decay oc degradation coefficient
R • retardation factor
. i * pb xd/e
Pb • bulk density
*d « distribution coefficient
9 • porosity
M « Mass source flux .
*
Solutions of Equation A-4 for various boundary and source
conditions can be specified in terms of the appropriate Green's
functions as (Codell and Schreiber. 1977; Yen and Tsai. 1976;
Carslaw and Jaeger. 1959):
c " 6R X0(x't) Yo(y't> V2
-------
where
T « degradation function due to degradation oc
decay of contaminants (includes the effects of
biodegradation. hydrolysis, chemical reaction.
etc.)
* exp(-Xt) for first order degradation (A-6)
XO.YQ.ZO * Green's functions for transport in x.y
and z directions, respectively.
Equation A-5 provides the contaminant concentration at any
point in space for an instantaneous release of a unit mass of
contaminant.
For a horizontal plane source extending from -L/2 to L/2
in the x direction and from -B/2 and B/2 in the y direction:
-
»9 O '
»rf -^-" » 3 (A.7)
/4D t/R
where:
L • width of source in x direction
x » distance in X direction to calculation point
-------
Foe a source located at the top of an aquifer of infinite
thickness
ZQ- exp(-Z2/(4Dzt/R))/(TTD2t/R) (A-9a)
and for a source located at the top of an aquifer of thickness h
<6b)
« m2ir2D_t
Zft . * (1+2 £ exp (- - =-*-) cos (mwz/H))
0 H m-1 H2 R
where
z • distance in z direction to calculation point
Substituting Equations A-6 through A-9 into Equation A- 5
produces solutions for an instantaneous unit release of mass
over an area of length L and width B at the top of an aquifer
of infinite or finite thickness. Applying a convolution
integral with respect to time provides the concentration C due
to a continuous release at a rate M(t) for any time T
following the beginning of a release.
0)T M(T)X0(x.t-T)Y0(y.t-T)Z0(Z.t-T)T(t-T)dT
(A-10)
where
M • mass released per unit time
Equation A-10 is generally applied to predict peak
concentrations due to time-varying source terms. Steady-state
concentrations can be predicted by specifying a constant source
rate M and integrating time from zero to infinity.
7705D PD-897
-------
Dispersion in the HPS model is formulated as a function of
the length scale traveled by the contaminants as follows:
ax « 0.1 X (8
• ay ' VRxy <9
where
R* • Ratio of ax to a
R~, • Ratio of a to a
XZ X 2
and all other parameters are as discussed above.
Equation 6 was suggested by Gelhar and Axness (1981) and
Walton (198S) as a means of estimating the longitudinal
dispersivity. Mill* «t al. (1985) suggest that the
longitudinal dispersivity should reach some maximum value at
large distances from the release point, though they do not
propose any specific value. Gelhar et al (1985) present a
range of 2.1 - 5.0 for R values and a range of 30 - 860 for
R values.
A.2 Model Applications and Results
A.2.1 Leachate Concentration Modeling
The OLM model was applied to predict the concentration of
contaminants leaching out of the lagoon sediment and soil
beneath the sediments. The parameters considered are listed in
Table A-l. Table A-l also provides the sediment and soil
concentrations of each of these compounds, obtained by
averaging measured concentrations taken from the PELA and EPA
sampling effort of June 20-24. 1983 (at stations J-05. K-02.
and K-03 for sediment and stations L-02A. L-02B. L-05A. L-05B.
L-06. L-07. and N-08 for soil); the OLM predictions for sludge
and soil; and the solubility for each compound.
»
7
7705D PD 897
-------
TABLE A-l
OLM MODEL PREDICTIONS AND SOLUBILITY VALUES
FOR CARCINOGENIC PAH COMPOUNDS AND FLUORANTHENK
(ALL VALUES IN PPM)
Compound
Carcinogenic PAH
Benzo[a ] Anthracene/
Benzo[a]Pyrene
Bftnzo(b]Fluoranthene
Chrysene
Dibenzo[a.hJ-
Anchracene
lndeno[l,2,3cd]-
pyrene
Fluoranthene
Sludge
Cone, in
Waste
510
770
14360
560
256
9100
OLM
Prediction
0.025
0.013
0.137
0.0090
0.0038
0.62
Soil
Cone. In
Waste
15.4
10.6
317
13.8
6.7
200
OLM
Prediction
o
0.0023
0.00073
0.010
0.00073
0.00032
0.046
Solubility
0.0089(a
O.OOOS
0.002
O.OOOS
\
0.0002
0.26
a. The mean solubility limit of the 2 compounds were used in the model.
-------
As indicated in Table A-l. the OLM model predicts that the
sludge "leachate concentration for each of the contaminants
considered will be greater than the solubility value.
Therefore, the OLM model was considered to be invalid and for
transport modeling it was assumed that the leachate
concentration from the sediment would be equal to the
solubility values given in Table A-l.
A.2.2 Model Applications to Live Oak site
A.2.2.1 Unsaturated Zone
Equation A-2 was used to calculate the moisture content in
the unsaturated zone between the lagoon and the upper boundary
of the limestone aquifer. In order to apply the equation to
the Live Oak Site the following assumptions were made regarding
the site surface and ground water conditions: s
• The drainage area of the lagoon (A) is 17.3 acres
(FTCH 1987)
• THe lagoon area (a) is 3 acres (FTCH 1987)
• The average annual rainfall (I) at the site is 52
inches/year (FTCH 1987)
• The fraction of the total rainfall runoff (C) in the
drainage area which infiltrates into the soil, before
reaching the lagoon, is 0.45.
• The annual evaporation from the lagoon (E) is 137
acre-in./y«ar (FTCH 1987)
• The steady state flux (q) through the unsaturated
zone equals the infiltration rate of lagoon water
into the soil and is calculated by
- c x I x A - E
'a
7705D PD-897
-------
therefore, q is 89 inches per year for the Live oak
site lagoon.
• The soil of the unsaturated zone can be
conservatively assumed to be a 30-foot layer of sand
*vith an organic carbon content of 0.1 percent (foe »
.001). where
Ksat • 1.056 cm/min (Clapp. et al 1978)
b - 4.05 (Clapp. et al 1978)
6B « 0.4 (Clapp. et al 1978) and
8 3
p - 1.5 gin/cm
Using the above assumptions. Equation A-3 vas used to determine
the retardation factor and Equation A-4 was used to determine
the mass flux rate for each compound. These values are
presented in Table A-2.
A.2.2.2 HPS Model
The HPS model was applied to predict contaminant
concentrations at monitoring well MW-8 and at a drinking water
supply well located 1600 ft (488 meters) from the lagoon.
assumed to be directly downgradient.
Input parameters and assumptions used in the model were:
• the waste area (lagoon) is 70 meters long and 173
meters wide
• ground-water velocity in the x direction is 33.40
meters/year. (Conservative value derived using
Darcy's Law; a permeability of 1000 ft/day and a head
gradient of 0.0003).
• the dispersivity in the x direction (a ) is less
than oc equal to 100 •
• the dispersivity la the y direction is equal to
ax divided by 3 (R - 3)
10
7705D PD-897
-------
TABLE A-2
RETARDATION FACTORS AND MASS FLUXES
FOR CARCINOGENIC PAH COMPOUNDS,
FLUORANTHENE AND PHENANTHRRNE
Constituent;
Benzo(«)pyrene/
benzo(a)an t hracene
Benzo(b)fluoranthene
Chrysene
Dibenzo(a,h)anthracene
Indeno(1.2.3-cd)pyrene
Pluoranthene
Phenanthrene
Retardation
Factor
2.1x10
«
4.1x10'
1.5x!0:
2.8x10'
l.OxlO1
Mass
Flux
(gre/yr)
0.012
0.005
0.037
0.0005
0.0005
25.0
261.0
11
-------
• the dispersivity in the z direction is equal to
a divided by 50 (R » 50)
-" XZ
• Mass flux rates were obtained from the unsaturated
zone calculations (Table A-2).
The predicted concentrations, after 50 years of constant
release from the lagoon, of each constituent at MW-8 and the
receptor veil are listed in Table A-3. along with the maximum
measured concentrations in MW-8. Fluoranthene and phenanthrene
were the only constituents detected in MW-8 at concentrations
above detection limits. The predicted concentrations of
fluoranthene and phenanthrene at MW-8 are 50 and 290 percent
greater, respectively, than the measured concentrations at the
same veil, demonstrating the conservative results of the model.
12
7705D PD-897
-------
TABLE A-3
HPS MODEL PREDICTIONS OF CONCENTRATIONS
OF CARCINOGENIC PAH, FLUORANTHENE AND PHENANTHRENE
AT COMPLIANCE WELL
Constituent
Benzo(a)pyrene/
benzo(a)anthracene
Benzo(b)£luoran t hene
Chrysene
Dlbenzo(a.h)anthracene
Indeno(l,2.3-cd)pyrene
Fluoranthene
Phenanthrene
HPS
Predicted Cone.
at MV-8 (mq/11
.1X10
-5
.6X10
.4X10
.6X10
.6X10
.003
.031
-6
-5
-7
-7
HPS Highest
Predicted Cone. Measured Cone.
at Receptor Veil (mg/1) at MV-8 (mg/1)
.6X10
.2X10
.2X10
.2X10
.2X10
.0001
.001
-7
-7
-6
-8
-8
<.020
<.020
<.020
<.020
<.020
.002
.008
13
-------
REFERENCES
Batelle. 1986. Guidance Criteria for Identifying Areas of
Vulnerable Hydrogeology. Appendix C. Technical Guidance
Manual for Calculating Time of Travel (TOT) in the
Unsaturated Zone. EPA. Office of Solid Waste and Emergency
Response.
.Campbell. G.S.. 1974. "A Simple Method for Determining
Unsaturated Conductivity from Moisture Retention Data".
Soil Science. Vol. 117. pp. 311-314.
Carslav. H.S.. and J.C. Jaeger. 1959. Conduction of Heat in
Solids. Oxford University Press. London.
Clapp. Roger B. and George M. Hornberger. 1978. Empirical
Equations for Some Soil Hydraulic Properties. Water
Resources Research. Vol. 14, No. 4. pp. 601-604.
Codell. R.B. and D.L. Schreiber. 1977. "NRC Models for
Evaluating the Transport of Radionuclides in Groundwater".
Proceedings of Symposium on Management of Low Level
Radioactive Waste. Atlanta. Georgia.
EPA. 1986. Hazardous Waste: Identification and Listing.*
Leachate. Federal Register. Vol. 51 No. 145. Tuesday.
6/29/86. Proposed Rules.
Fishbeck. Thompson. Carr and Huber (FTCH). 1987. Report on the
Remedial Investigation Brown Wood Preserving Site. Live
Oak. Florida.
Galya. D.P.. A Horizontal Plane Source Model for Ground-water
Transport. Groundwater. in press.
Gelhar. L.W.. A. Mantoglou. C. Welty and K.R. Rehfeldt. 1985. A
Review of Field Scale Subsurface Solute Transport
Processes Under Saturated and Unsaturated Conditions.
Electric Power Research Institute. Palo Alto. California.
107 p.
Gelhar. L.W. and C.J. Axness. 1981. Stochastic Analysis of
Macro-Dispersion in Three-Dimensionally Heterogeneous
Aquifers. Report No. H-8. Hydraulic Research Program.
New Mexico Institute of Mining and Technology. Soccorro.
New Mexico. 140 p.
Mills. W.B.. D.B. Porcella. M.J. Ungs. S.A. Gherini. K.V.
Summers. L. Mok. G.L. Rupp. G.L. Bowie, and D.A. Haith.
Water Quality Assessment: A Screening Procedure for Toxic
and Conventional Pollutants.
7708D PD-897
-------
REFERENCES (Continued)
•
Yacon B".. G. Dagan and T. Goldshmid.. 1984, Pollutants in Porous
Media - The Unsatucated Zone Between Soil Surface and
Groundwater. Spcingec - Verlag. Berlin. 296 pages.
Yen. Gour-Tsyh and Yun-Jui Tsai. 1976. "Analytical
Three-Dimensional Transient Modeling of Effluent
Discharges." Hater Resources Research. 12(3):533-546. June
1976.
7V08D PD-897
-------
APPENDIX B
DETERMINING SURFACE SOIL CONCENTRATIONS
-------
TABLE B-l
DATA USED FOR DETERMINING
SURFACE SOIL CONCENTRATIONS
LIVE OAK, FLA.
Source
EPA. Feb. 1983
LETCO, Sept. 1983
(0-1 ft depth)
Remedial
Investigation
Test Holes in
Vicinity of Lagoon
(0-1 ft. Above
water Only)
Remedial
Investigation
Individual soil
Samples in Former
Plant Area
(0-0.5 ft depth)
Total
Carcinogenic Cresote
No. PAH constituents
TAB. 1-4
A-l
A-2
A-3
A-4
A-5
A-6
A-7
5(005)
8(027)
9(034)
10(040)
11(046)
12(051)
13(055)
14(062)
300
302
304
306
308
310
312
314
316
318
17500
15
0
12.7
13.9
8.4
269
29.5
594
0
3800
680
166
11600
900
918
17.88
0.919
15.79
4.82
1.18
0.072
1.51
19.36
18.11
10.7
221300
32.8
0
43.8
44.8
23.7
1005.3
51.3
1615
16300
37990
5345
253.8
146389
16770
2713
32.658
1.328
23.724
8.06
2.268
0.0824
2.934
30.736
33.02
18.09
PCP Pluoranthene
3000 29000
830
6.1
158500
985
9550
169.5
0
0
0
0
0
0
20000
0
10
0
0
0
0
0
0
0
0
0
3.7
9.6
13.7
5.1
150
7.9
490
7700
7300
2300
9.3
26000
3300
770
0.16
0.011
4.7
0.026
0.029
0.095
0
0.35
3.8
0
-------
TABLE B-l (Continued)
Source
No.
Carcinogenic
PAH
Total
Creosote
Constituents
PCP
Fluoranthene
Remedial .
Investigation
Composite Sample
of Soil in Wood
Storage Area
(0-0.5 ft depth)
320
322
400
Area 1
Area 2
Area 3
Area 4
Area 5
Area 6
Area 7
Area 8
MAX
MEAN
SD
SEN
6.77
2.02
0.503
17.7
11.4
0.058
3.41
2.91
54.4
2.5
1.53
17500
995
3375
555
21.465
3.3982
0.706
31.2
20.09
0.1515
6.426
4.22
77.7
4.786
3.38
221300
12168
42539
6993
0
0
0
0
0
0
0
0.5
0
0
s
0
158500
5363
26135
4356
0
0.17
0
5.1
3.1
0.036
1.2
0.58
5.5
0.8
0.73
29000
2141
6415
1069
-------
TABLE B-2
MEAN INDICATOR CHEMICAL CONCENTRATIONS
WHEN TOTAL CREOSOTE SUBSTANCES ABOVE 1000 PPM ARE REMEDIATED
LIVE OAK, FLA.
Source
LETCO, Sept, 1983
(0-1 ft, except as
noted)
RI, Lagoon
(0-1 ft, except as
noted)
RI, Plant Area
(0-0.5 ft depth)
Carcinogenic
No. PAH
A-l
A-2
A-3
A-4
A-5
A-6
A-7
5(007,3-5 ft)
8(029.3-5 ft)
9(036,3-5 ft)
10(042,3-5 ft)
11(046)
12(054,5-7 ft)
13(0588,5-7 ft)
300
302
304
306
308
310
312
314
316
15
0
0
13.9
8.4
24.1
29.5
0
10.8
0
0.15
166
8.63
6.65
17.88
0.919
15.79
4.82
1.18
0.072
1.51
19.36
18.11
Total
Creosote
Substances
32.8
0
0
44.8
23.7
70.6
51.3
0.34
105.4
5.09
13.11
253.8
166.5
157.2
32.658
1.328
23.724
8.06
2.268
0.0824
2.934
30.736
33.02
PCP Fluoranthcne
830
6.1
14.90
985
805
169.5 s
0
0
0
0
0
0
0
10
0
0
0
0
0
0
0
0
3.7
0
13.7
5.1
10.4
0.15
23.0
.76
.34
9.3
23
24
0.16
0.011
4.7
0.026
0.029
0.095
0
0.35
3.8
-------
TABLE B-2 (Continued)
Carcinogenic
Source No. PAH
318
320
322
400
RI, Wood Storage Area Area 1
(0-0.5 ft depth) Area 2
Area 3
Area 4
Area 5
Area 6
Area 7
Area 6
MAX
MEAN
SO
SEN
10.7
6.77
2.02
0.503
17.7
11.4
0.058
3.41
2.91
54.4
2.5
1.53
166
14
29
5
Total
Creosote
Substances
18.09
21.465
3.3982
0.706
31.2
20.09
0.1515
6.426
4.22
77.7
4.786
3.38
254
36
56
10
PCP
0
0
0
0
0
0
0
0
0.5
0
0
0
985
83
252
44
Fluoranthene
0
0
0.17
0
5.1
3.1
0.036
1.2
0.58
5.5
N
0.8
0.73
24
4
7
1
-------
MEMO TO: Distribution
FROM: J. Ryan
DATE: February 17, 1988
RE: ACTION LEVELS FOR THE LIVE OAK SITE
INTRODUCTION
This memorandum discusses action levels for final remediation of
the Live Oak site. The document has been prepared based on
discussions with EPA and FDER personnel and additional work on
the Risk Assessment report for the site. Included in this
evaluation are considerations related to: (a) the existing site
conditions following the interim action and (b) the proposed
biological treatment of the contaminated soils remaining after
the interim removal action.
In preparing the action levels the following items were
evaluated:
o criteria for soil concentrations of carcinogenic PAH,
o background concentrations of carcinogenic PAH,
o the proposed remedy,
o leachate concentrations associated with residuals,
o average site concentrations of carcinogenic PAH,
o the future development of the site,
o appropriate risk levels,
o extent and cost effectiveness of removal, and
o the risk associated with ingestion of surficial
contaminated soils.
•
SOIL CRITERIA
Although no standards have been set for clean up levels
associated with PAH contaminated soil, various guidelines and
site specific actions can be used to provide a perspective on
soil action levels. The most significant study is by the Center
for Disease Control in Atlanta which builds on previous work
-------
completed with dioxins.
The Centers for Disease Control's Agency of Toxic Substances
Disease Registry (ASTDR) has evaluated the carcinogenic status of
polycyclic aromatic hydrocarbons (PAH) relative to 2,3,7,8
tetrachlorodibenzo-p-dioxin, an extremely potent animal
carcinogen (3). ASTDR has suggested 100 ppm for PAH in soil as
a safe level.
As stated by Dr. Stephen Margolis of the ASTDR:
"In a published article (4), the Centers for Disease
Control (CDC) derived an action level at which to limit
human exposure for 2,3,7,8-tetrachloro-dibenzo-p-dioxin
(2,3,7,8-TCDD), contaminated residential soil. This
derived value was based upon extrapolations from animal
toxicity experiments (including carcinogenicity and
reproductive effects) to possible human health effects
in order to estimate a reasonable level of risk for
2,3,7,8-TCDD. A 10~6 excess lifetime risk was used in
the development of this TCDD soil level.
The Environmental Protection Agency's Carcinogen
Assessment group has derived a relative potency index
for more than 50 chemicals (5). The order of
magnitude potency index for 2,3,7,8-TCDD is eight,
while that for benzo(a)pyrene is only .three. Thus,
2,3,7,8-TCDD is considered to be five orders of
magnitude more potent as a carcinogen than
benzo(a)pyrene. Using only this order of magnitude
difference in potency between the two chemicals and the
CDC-derived residential soil action level, gives
100,000 ppb of benzo(a)pyrene equivalent to 1 ppb of
2,3,7,8-TCDD in soil."
This comparison used benzo(a)pyrene as the representative PAH.
It must be recalled that benzo(a)pyrene is the most potent of all
PAH compounds studied. Therefore, this model is considered to
be a conservative model when applied to the other suspectcj
carcinogens.
An additional measure of conservatism is added to the ASTDR model
by overestimating the ingestion of contaminated soil. Again, as
stated by Dr. Margolis,
"In the model used to derive the 2,3,7,8-TCDD soil
value, the assumption concerning the amount of soil
ingested has been shown to be high. A recent
published study by CDC has shown the amount of soil
ingested by children of the soil-eating age ranges from
0.1 to 1 gram per day (S. Binder personnel
communication). Thus, the model estimate for soil
-------
ingestion during the period of minimum hygiene is
excessive by at least an order of magnitude. Since
the other soil ingestion rates in the model are also
estimates, there is a good likelihood that they are
also in error, possibly by more than an order of
magnitude. Thus, the model very likely overestimates
the total lifetime soil ingestion exposure by at least
one order of magnitude."
Because of the conservative nature of the recommended CDC action
levels, USEPA Region VI and Region VII have used the 100 ppm as
action levels for carcinogens at PAH contaminated sites.
BACKGROUND CONCENTRATIONS
The remedial investigation included sampling of background areas
where PAH compounds may have been used or occur naturally. One
area was a swamp and the other was abandoned railroad tracks.
Concentrations of carcinogenic PAH at the abandoned railroad
tracks were in excess of 16 ppm and in the swamp were 0.8 ppm.
PROPOSED REMEDY
The proposed remedy is described in the attached memorandum
entitled "Conceptual Plan for Final Site Remediation". The
remedy involves: (a) plant demolition, (b) removal and off site
disposal of hardened creosote sludges from the plant area, (c)
excavation of contaminated soils from the plant area, (d)
biological treatment of contaminated soils within or adjacent to
the former lagoon and (e) site security and monitoring.
It should be noted that the majority of site contamination has
been removed as part of the interim removal action. Over 15,000
tons of sludges and contaminated soils were removed from the
lagoon. (See memo on Interim Removal Action.) An additional 1500
tons of hardened creosote will be removed from the plant site and
disposed at CWM's facility in Eroelle, Alabama, as part of the
proposed remedy. An estimated 10,000 tons of soil containing
greater than 1000 ppm total creosote substances will be treated
on-site biologically. The treatment goal for these soils is 100
ppm carcinogenic PAH. Based on modeling and experience with
biological treatment of similar materials it is estimated that
the treatment process will take less than two years.
The cost effectiveness of the proposed plan can be contrasted to
the cost effectiveness of removing soils containing less than 100
ppm carcinogenic PAH by comparing the overall mass of TCS removed
(or treated). Approximately 6,000 tons of sludge and 9000 tons
of highly contaminated soil were removed as part of the interim
removal. It is estimated that another 1500 tons of hardened
-------
creosote will be removed from the plant area and disposed of at
Emelle "and 10,000 tons of moderately contaminated soil will be
treated as part o.f the final remedial action. The sludge and
hardened creosote are estimated to have an average concentration
in excess of 100,000 ppm total creosote substances (TCS). The
highly contaminated soil is estimated to be in excess of 10,000
ppm TCS and the moderately contaminated soil is estimated to have
TCS concentrations ranging from 1000 to 5,000 ppm TCS.
Surficial soil which is greater than 100 ppm TCS and less than
1000 ppm TCS was not identified in the R.I. Surficial soil which
exceeds 10 ppm TCS and is less than 100 ppm TCS is estimated at
approximately 7500 tons. Surficial soil which exceeds one ppm
TCS but is less than 1,000 ppm TCS is estimated at 31,750 tons.
Table 1 compares the mass of TCS which will have been removed and
the mass of TCS which will be treated to the mass present in the
surficial soil which is in excess of one ppm in the surficial
soils. The cost of removing or treating these materials is
expressed on the basis of $/pound of TCS. It can be seen that
the incremental cost of removing the surficial soils is two
orders of magnitude greater despite the fact that these materials
represent less than 0.03 percent of the total creosote
substances; N
LEACHATE MIGRATION
This pathway was evaluated in detail in the existing Risk
Assessment report and was concluded to represent an acceptabl
risk under the existing conditions in the lagoon. This pathway
is revisited, however, to evaluate the impact of the treatment
drainage returning to the pond. The RITZE model was used to
evaluate the concentrations of carcinogenic PAH in the soil por~
water which would be recycled through the treatment pond. Tl
modeling results indicate that this water should be below
detection limits of one ppb for individual PAH compounds.
These results are reasonable in light of reported sediment/water
partition coefficients for these compounds. Tables 2A and ?B
.present the range of these partition coefficients. In general,
log Koc range from 5.29 to 7.34. The actual water concentration
will be a function of the soil organic carbon content and th_
concentration of the individual compound. Using an organic
carbon content of 0.5 to one percent and an initial concentration
of 20 ppm, the estimated water concentration would range from 21
to <0.5 ppb.
These results have also been shown in TCLP testing of the
contaminated soil and sludges. These results are shown on Table
3 and are expressed as the total concentration and the TCLP
(water soluble) extract.
-------
Table 1
COST EFFECTIVENESS COMPARISON
MATERIAL
SLUDGE
HICHLT CONTAMINATED SOIL
HARDENED CREOSOTE
MODERATELY CONTAMINATED SOIL
SOIL >1 PPM TCS
TONS
MATERIAL
6000
9000
1SOO
10000
37750
TCS
100000
10000
100000
5000
9
TONS
TCS
600
90
150
50
0.34
t/TOM
MATERIAL
200
200
200
50
50
S/POUND
TCS
1
10
1
5
2775
X TCS
REMOVED
67.39
10.11
16.85
5.62
0.03
TOTAL COST
<«)
1200
1800
300
500
1887.5
X TOTAL
COST
21.1
31.65
5.27
8.79
33.19
TOTAL 64250 890.34 100 5687.5 100
-------
TABLE 2 a
RANGE OF VALUES FOR LOG(KOC) FOR SELECTED PAH COMPOUNDS (a)
COMPOUND
BENZO(A)
CHRYSENE
DENZO(B)
BENZO(K)
BENZO(A)
ANTHRACENE
FLUORANTHENE
FLUORANTHENE
PYRENE
DIBENZOf A, H) ANTHRACENE
INDENO(1
, 2,3-CD) PYRENE
REFERENCE NUMBER
6 7(b) 8 (C) RANGE
lower upper lower upper lower upper
5.
5.
5.
5.
5.
6.
6.27
40
36
63
83 6.65
76 5.75
45
5.
5.
6.
6.
5.
6.49 6.22 5.
7.
29
29
25
52
72
65
34
5.
5.
6.
6.
5.
5.
7.
29
47
25
52
72
65
34
5.
5.
5.
5.
5.
5.
6.
29
29
36
63
72
65
45
6.
5.
6.
6.
6.
6.
7.
27
•i
25
t
65
<.
3"
(a) All units for partition coefficients are ml/g.
(b) Data was derived from Kd data -
Kd = Koc * (organic carbon fraction)
(c) Calculated from Kow data - (9,6)
and equation: (10, 11)
log Koc = log Kow - 0.317
-------
TABLE 2b
RANGE OF LEACHATE WATER CONCENTRATIONS
COMPOUND
DEIIZO (A) ANTHRACENE
CHRYSENE
BENZO(B) FLUORANTHENE
UENZO(K) FLUORANTHENE
BENZO(A)PYRENE
DI BENZO (A , H ) ANTHRACENE
1NDENO (1,2, 3 -CD) PYRENE
Koc (1) 0.
lower upper
194984
194984
229087
426580
524807
446684
2818383
1862087
295121
1778279
3311311
4466836
3090295
21877616
WATER CONCENTRATION (ug/1) (2)
.5% organic carbon 1.0% organic carbc
upper lower upper lower
21
21
17
9
8
9
1
2
14
2
1
1
1
0
10
10
9
5
4
4
1
1
7
1
1
0
1
0
(1) From Table 2a
(2) calculated using a constituent concentration of 20 ppm.
-------
TABLE 3
Concentrations of Indicator Parameters
Analyzed in Soil Samples
taken August 1986 from test pits
within the Lagoon during low water level
(concentration in ppm)
Senple N'umber
Test Pi t Number
Soil Type
Visible Contaminated?
Depth (feet)
PARAMETER :
Naphthalene
Acenaphthylene
Acenaphthene
Fluorene
Pentachlorophenol
2,4 , 6-Trichlorophenol
Phenanthrene
Anthracene
Fluoranthene
Pyrene
Benzo( a) anthracene
Chrysene
Benzo(b) f luoranthene
Benzof k ) f luoranthene
Benzo(a)pyrene
Indeno( 1 . 2 , 3-cd Jpyrene
Di benzol a .h)anthracene
Benzo(g,h, i Iperylene
Total Creosote Substance
Total Pentachlorophenol
Total Non-Carcinogenic PAH
Total Carcinogenic PAH •*
Percent Carcinogens of
Total Creosote
02
1
Sand
No
1-2
--- Total
<0.15
<0. 15
0.33
0.33
<0.15
<0. IS
0.41
<0. 15
0.63
0.45
<0. 15
<0.15
<0. 15
<0. 15
<0. 15
<0. 15
<0. )5
CO. 15
2. 15
<0. 15
2. 15
<0. 15
OX
03
2
Sand
Yes
1-3
Concent
4
<0. 15
23
29
<0. 15
<0. 15
75
9.70
36
24
5.1
4.9
3.7
1 .2
1.5
0.43
0.27
0.38
218. 18
< 0 . 1 5
202.28
15.90
7X
04 02
3 1
Sludge & Sand Sand
Ves No
0-0.5 1-2
23000
230
8600
11000
990
<100
24000
13000
9400
7400
1800
1100
600
400
370
110
<100
100
1011 10
990
97130.00
3980.00
<0.01
<0.01
0.01
0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
(0.01
<0.01
0.02
<0.01
0.02
<0.01
4X OX
03
2
Sand
Yes
1-3
Cor.cen
0.09
<0.01
0.03
0.24
<0.01
(0.01
0.27
0.02
0.03
0.02
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
<0.0]
<0.01
0.70
<0.01
0.70
<0.01
OX
04
3
Sludge & Sand
Ves
0-0.5
t rat ions
11
<0.01
0.67
0.44
<0.01
<0.01
0.47
0.08
0.42
0.09
<0.01
(0.01
<0.01
(0.01
(0.01
(0.01
<0.01
(0.01
13.17
(0.01
13.17
<0 . 01
OX
Notes: Analysis by Rocky Mountain Analytical Laboratory.
Re er to RMAL s 6 S97
N. ). indicates not detected
Carcinogenic PA Is : Ber.rol a ) anthrac"1 .• Een:o ( bl f 1 uor jr. t hene . Cirysene
t r "'"''' a I B y r e TI e r )iber>'"">'»,hl«»ntf, ra_ e, l nH^ no'J ?,3->-'^|>^v'"^
-------
AVERAGE SITE CONCENTRATIONS
The exposure assessments related to soil ingestion consider a
long term site exposure rather than a short term exposure related
to a specific area on the site. Any development of the site for
either industrial or residential purposes would require extensive
earthwork. It is appropriate, therefore, to consider a
composite site concentration of carcinogenic PAH. This composite
is based on an average concentration across the site to a depth
of two feet.
The average concentrations have been calculated based on the
conditions which will exist after the contaminated soil from the
plant area have been removed and assumes the residuals in the
former lagoon have been treated to a level less than 100 ppm
carcinogenic PAH.
Figure 1 shows the areas used in these calculations. The former
wood storage area is approximately 520,000 square feet in size.
The plant area is approximately 240,000 square feet in size and
the former lagoon area (and future treatment area) is
approximately 240,000 square feet in size. The overall site area
is about 2,100,000 square feet in size. s
Average concentrations over the plant area and the wood storage
area were calculated using existing surficial data (0-12 inches)
excluding highly contaminated samples from the plant area which
will be removed as part of the final remedial action.
Concentrations from 12 to 24 inches were estimated using 50
percent of the concentration from six to 12 inches. If the
concentration measured at six to 12 inches was less than 1.3 ppm
then the concentration at 12 to 24 inches was assumed to be BDL.
Figure 1 shows the location of the sampling points and Table 4
presents the measured concentrations and the estimated composite
average.
Concentrations in the proposed treatment area were assumed to be
100 ppm in the upper 12 inches and 10 ppm at 12 to 24 inches
based on modeling results. (See the memorandum "Conceptual Plan
for Final Site Remediation". The composite concentrations for
the three areas is 16 ppm and for the overall site is 7.6 ppm.
for carcinogenic PAH.
SITE DEVELOPMENT
Most of the soil criteria revolve around the issue of the future
use of the site. The PRP's are pursuing institutional controls
which would preclude the use of the site as a residential
development. The action levels, therefore consider both
unrestricted residential development and restricted site access
-------
TABLE 4
CONCENTRATIONS OF
CARCINOGENIC PAII (ppm)
WOOD STORAGE AREA
nvcj
composite
(over 2 feet)
SAMIM.E
HUMUER
] .00
2.00
3.00
1.00
•J.OO
6.00
7.00
11. 00
•J.OO
10.00
11.00
1 .\ . 00
J 3.00
11 .00
15.00
i r, . oo
J V . 00
111.00
ly.oo
20.00
21 .00
2:;. oo
2:1.00
2-1 .00
::•.». oo
;if... oo
2V. 00
20.00
2*>.00
30.00
31.00
37.00
te avrj.
0-6
INCURS
11.10
11.40
11 .40
11.40
2.0
C2.00
02.00
62.00
62.00
17.70
17.70
17.70
17.70
0.06
0.06
0.06
O.OG
12.72
6-12
INCHES
0.23
0.23
0.23
0.23
0.06
0.06
0.06
0.06
0.70
0.70
0.70
0.70
1.30
1.30
1.30
1.30
0. 15
0.15
0.15
0. 15
6.40
6.40
6.40
6.40
7.10
7.10
7.10
7. 10
DDL
DDL
UOL
DDL
1.99
4.10
1 2 - 2 •!
inciiL.';
LJDL
HDL
ItUL
IJUL
noi.
LWL
I)OL
UDL
UUL
HUL
UDL
ni>L
nni.
UDL
lll)I>
DDL
UDL
DDL
IJO/.
UDL
3 .20
I».20
3.20
3.20
3.60
3. CO
3 .GO
:».co
HUL
I.JOL
DDL
IJUL
O.U5
AREA
33.00
34.00
35.00
36.00
37.00
30.00
39.00
40.00
41.00
42.00
43.00
44.00
45.00
17.20
0.92
15.00
4.00
1.20
0.07
1.50
19.40
10.10
10.70
6.70
1.60
0.50
0.73
UDL
6. CO
0.62
0.04
0.05
0.02
0.30
11.40
0.06
0.02
2.10
DDL
DDL
UDL
3.30
HDL
DDL
UDL
DDL
L1UL
5.70
UDL
UDL
1.00
UDL
nvg.
composite avcj.
(over 2 feet)
7.SB
1.C9
2.70
0.77
-------
HALF LIFE DAYS
oooooooooooooooooooo
I I I I I I I I I I I I I I I 1 • I I
CO
10
co
a.
to
CD
x H
5
TJ
30
• •
S-
33 Q
5' CD i
(O ==
NAPHTHALENE
ACENAPHTHENE
ANTHRACENE
•" FLUORENE
PHENANTHRENE
BENZO(A)ANTHRACENE MM
•—• CHRYSENE
•—•• FLUORANTHENE
•—• PYRENE
BENZO(A)PYRENE
vBENZO(B)FLUOftANTHENE
BENZOOOFLUORANTHENE
DBENZOCAHJANTHRACENE
ENZO(GHI)PERYLENE
i|NDENO(123-CD)PYRENE
-n
rn
x m
30
P
CO
CD
Cfl
33
m
rn
ffi
D
H
-------
(such as an industrial park or preserve land).
The site_ in its current condition is heavily vegetated with less
than five percent of the site area having little or no ground
c-ver. The probability of extensive residential development on
the property is extremely low given the demographics of the area.
However, if development were to occur, extensive earth moving
would be required to provide utilities, roads and landscaping.
Institutional controls that would limit residential development
could include a notice in the property deed or a restrictive
covenant. A proposed restrictive covenant is included as
Attachment A to this nemo. However, Amax and The Brown
Foundation are not in a position at this time to guarantee
restricted use of -the site until the current property
owners/1ienholders give their permission to this arrangement.
Amax and The Brown Foundation are prepared to make the
appropriate contacts to restrict the future use of the site upon
conceptual approval by USEPA and FDER. Various site development
scenarios were considered in developing this memorandum. The:
include:
Scenario 1: Industrial Adult
This scenario considers an adult who inadvertently ingests 5.0 mg
of soil from the site once every 30 days all year for 35 years
during adulthood (ages 25 to 60 years). This adult has an
average weight of 70 kg during the exposure and has a life
expectancy of 70 years.
Scenario 2: Neighborhood Child
This scenario considers a child who inadvertently ingests 50.0 mg
of soil from the site once every 30 days all year for five years
during childhood (ages 6 to 11 years). This child has an
average weight of 30 kg during the exposure period and has a life
expectancy of 70 years.
Scenario 3: Residential Person
This scenario considers a person who lives on the property ar"
has a life expectancy of 70 years. For the five-year perirj
from ages one to six years, this person inadvertently ingest-
100.0 mg of soil from the site every third day and has an average
weight of 15 kg during the period of exposure. For the five-yea..
period from ages six to 11 years, this person inadvertently
ingests 50.0 mg of soil from the site every third day and has an
average weight of 30 kg during the period of exposure. For the
59-year period from ages 11 to 70 years, this person inadvert-
ently ingests 5.0 mg of soil from the site every third day and
has an average weight of 70 kg during the period of exposure.
The weight and soil ingestion factors are based on recently
12
-------
published data (12,13). No specific criteria are available in
terms of the frequency of exposure. Due to the relative
inaccessibility of the site and the improved hygiene of adults,
the frequency of exposure for the neighborhood and industrial
setting is reasonable. This frequency is increased by an order
of magnitude (10 times a month) for the residential setting.
This also is reasonable in light of the conservative lifetime (0-
70 years) which the assessment spans, and considerations related
to:
o Inclement weather
o close supervision of young children, and
o time spent away from the home (at school, work, etc.).
Summary: Limited exposure currently exists due to contact with
exposed surface soil. The probability of residential development
is extremely low in the foreseeable future. Institutional
controls restricting access are feasible.
RISK LEVELS
A major consideration in establishing action levels'* is the
appropriate level of risk. CERCLA guidance recommends risk
levels which range between 10~4 to 10"'. Generally, a level of
10~5 to 10"6 is considered appropriate for protection of human
health. The lower risk (10-6) is appropriate in instances where
the site conditions represent a substantial and immediate threat
to human health and the environment. The 10~5 risk is
appropriate for situations such as Live Oak where the probability
of risk is low and is driven by conservative modeling assumptions
and future land use considerations.
SOIL INGESTION
The Risk Assessment report provided an estimate of risk for the
site based on a number of highly conservative assumptions and
concluded that the baseline risk associated with the current
conditions (assuming waste removal to 1000 ppm of total creosote
constituents) was acceptable. This assessment is revisited to
evaluate risk the associated with both industrial and residential
development. Important assumptions made in this assessment
include:
THE POTENCY OF FACTOR FOR BENZOfAlPYRENE; The "Superfund Public
Health Evaluation Manual", (1) conservatively estimates a cancer
potency factor (CPF) for benzo(a)pyrene of 11.5 [mg/kg/day]'1.
ICF Clements (the same contractor which prepared the above
Guidance Manual) has completed a recent evaluation of existing
data and the application of a biologically based dose response to
13
-------
evaluate the potency of benzo(a)pyrene; (2) This evaluation has
established the potency of benzo(a)pyrene as 5.74 [mg/kg/day]'1.
This assessment is provided as Attachment B to this memorandum.
THE RELATIVE POTENCY OF THE SUSPECTED CARCINOGENS; The existing
site database was reviewed to evaluate the relative ratios of the
carcinogenic PAH. The data base included the soil data on the
plant area and former wood storage area summarized in Table 4.
The data base indicates the average relative proportions of tl._
carcinogenic PAH are:
Relative % of CPF
Comound Total Carcinogens
benzo(a)pyrene 11.54% 5.74
benzo (a) anthracene 8.71% 0.8323
benzo(b) fluoranthene 49.43% 0.8036
chrysene 18.52% 0.0253
dibenzo(a,h) anthracene 3.56% 6.3714
indeno{l,2,3-cd)pyrene 8.23% 1.3317
Shown above is the cancer potency factor calculated by ICF
Clemens for the other carcinogens. The relative distribution of
the carcinogens and their cancer potency factor was N used in
modeling risks associated with soil ingestion. The cumulative
cancer potency factor* based on the relative distribution of the
carcinogens is 1.4713.
MATRIX/AVAILABILITY EFFECTS; The bioaccumulation of the PAH
compounds is strongly related to the soil matrix containing the
compounds. Low levels of compounds sorbed onto soil cannot we
readily accumulated by the body. A matrix factor of 0.3 is used
in this assessment and is consistent with the value used by the
CDC in evaluating action levels for dioxin contaminated soil.
DEGRADATION FACTORS; A half-life of 0.5 years was used to moc1-!
the photochemical and biological degradation of the carcinogens.
Attachment C presents a database which shows these carcinogens
generally have half lives less than 0.5 years. Figure 1 present ~
the maximum range associated with the 95 percent confidei.v
interval for the attached data base. The half year assumption is
the maximum value for any of the carcinogenic compourjs
evaluated.
RELATIVE RISK
The relative risk associated with the residuals remaining on the
plant site after completion of the proposed remedy was calculated
for each of the relevant areas as wall as for the site as a
whole. Table 5 presents the average concentration associated
14
-------
TABLE 5
LIFETIME RISK ASSOCIATED
WITH
VARIOUS AREAS OP THE SITE AFTER
TREATMENT OF CONTAMINATED SOILS TO
100 ppm TOTAL PAH CARCINOGENS
Location
Risk Levels Under Various Scenarios
Residential
(b)
Industrial
Neighborhood
Child
Wood Storage Area
Plant Area
Treatment Area
Site Area (a)
7.7 x 10~8
5.0 x 10"8
1.0 x 10~s
1.4 x 10
-7
8.2 x 10'11
5.4 X lO'11
1.1 X 10~9
1.5 X 10
-10
1.9 X 10~9
1.3 X 10~9
2.6 X 10~8
3.6 X 10~9
a) weighted average of wood storage area, plant area, treatment area
and the overall site.
b) Note: The residential scenario assumes that houses are constructed
on the site immediately after treatment is completed and that a
person spends his entire life on this site (from birth to 70 years).
This scenario is extremely conservative and highly unlikely given
the demographics of the area.
-------
with each area (weight averaged over two feet) and the associated
risk level with various exposure scenarios. The table shows that
the risks associated with the proposed remedy range from 10~7 to
10~10 for the overall site. This risk is based on the completion
of the proposed remedy which includes treating the contaminated
soils down to a level of 100 ppm total carcinogens. Attachment D
provides the equations used in these calculations.
SUMMARY
This. review has evaluated a variety of considerations which
affect soil action levels for the Live Oak site. The action
levels have considered both average site concentrations as well
as maximum concentrations for individual areas.
The proposed maximum concentrations for surfical soils in any
given area is 100 ppm of carcinogenic PAH. This value is based
on work completed by the Center for Disease Control on tt._
relative potency of PAH carcinogens as compared to 2,3,7,8-TCDD.
This value has been used as an action level at other P"l
contaminated sites.
Under the proposed action levels, the maximum concentration
criteria would be applied to the treatment area. The femainder
of surficial soils on the site are well below the proposeJ
maximum criteria. The risk levels on a site average basis
following treatment to 100 ppm would be 10"' for an unrestricted
development scenario and 10~9 for a restricted development
scenario. Therefore, no further action is necessary for site
soils below 100 ppm carcinogenic PAH. It should be noted that
the unrestricted development scenario assumes that houses are
constructed on the site immediately after the completion of the
proposed remedy and that individuals live on the site from birth
to age 70. Given the demographics of the area, this scenario is
highly unlikely. Even so, the associated risk levels are 10~7
which demonstrates the proposed remedy is highly protective of
human health and the environment.
16
-------
REFERENCES
1. U.S. EPA "Superfund Public Health Evaluation Manual",
October, 1986.
2. ICF-Clement Associates "Comparative Potency Approach for
Estimation of the Total Cancer Risk Associated with Exposure
to Mixtures of PAH Compounds". Draft, January, 1987.
3. IARC Monographs on the Evaluation of the Carcinogenic Risk
of Chemicals to Humans, Volume 32, December, 1983.
4. Agency for Toxic Substances and Disease Registry, Memorandum
to Mr. Carl R. Hickham from the ATSDR Office of Health
Assessment, 1986.
5. Health Affects Assessment for Polycyclic Aromatic
Hydrocarbons, September, 1984, EPA/540/1-86-013.
6. Mahmood, R.J. and R.C. Sims. 1986. Mobility of organics in
land treatment systems. Journal of Envisonmental
Engineering, 112:2 236-245.
7. American Petroleum Institute (API). March 1986.
Treatability data in support of a treatment zone model for
petroleum refining land treatment facilities. Washington,
D.C. 343 p.
8. Karickhoff, S.W. 1981. Semi-empirical estimation of
sorption of hydrophobic pollutants on natural sediments and
soils. Chemosphere. 10:8 833-846.
9. Miller, M.M., S.P. Wasik, G. Huang, W. Shiu, D. Mackay.
1985. Relationships between octonol-water partition
coefficient and aqueous solubility. Environmental Science
and Technology. 19:6 522-529.
10. Karickhoff, S.W., D.S. Brown, and T.A. Scott. 1979.
Sorption of hydrophobic pollutants on natural sediments.
Water Research. Vol. 13. 241-248.
11. Hassett, J.J., J.C. Means, W.L. and S.G. Wood. 1980.
Sorption properties of sediments and energy related
pollutants. EPA-600/3-80-041. Environmental Protection
Agency, Washington, D.C.
12. LaGoy, P. 1987. "Threshold Limit Values and Biological
Exposure Indices for 1987/1988" in Risk Analysis, Volume 7,
No.3, pgs. 355-359.
-------
13. Paustenbach, D.J. 1987. "Assessing the Potential
Environmental and Hunan Health Risks of Contaminated Soil"
in Comments on Toxicology, Volume 1, pgs. 185-220.
14. Kimbrough, R.D., H. Falk and P. Stehr. 1984. "Health
Implications of 2,3,7,8 TCDD Contamination of Residential
Soil in Journal of Toxicology and Environmental Health,
Volume 14, pags. 47-93.
-------
COMPARATIVE POTENCY APPROACH
FOR ESTIMATION OF THE TOTAL CANCER RISK ASSOCIATED
WITH EXPOSURES TO MIXTURES OF POLYCYCLIC AROMATIC
HYDROCARBONS IN THE ENVIRONMENT
FINAL REPORT
Prepared by
ICF-Clement Associates
1850 X Street, N.W.
Suite 450
Washington, D.C. 20006
July 1, 1987
-------
,; TABLE OF CONTENTS
Page
EXECUTIVE SUMMARY ii
TECHNICAL SUMMARY AND CONCLUSIONS vii
I. Introduction 1-1
II. Biological Mechanisms of Action of Polycyclic II-l
Aromatic Hydrocarbons
III. Development of a Comparative Potency Approach III-l
for the Assessment of Cancer Risk Associated
with Mixtures of Polycyclic Aromatic Hydrocarbons
IV. The Basis for and Derivation of Relative Potency IV-1
Estimates
V. Estimation of Cancer Risk due to Mixtures of V-l
Polycylic Aromatic Hydrocarbons using
Alternative Methods
VI. Validation of the Comparative Potency Approach •> VI-1
VII. REFERENCES VII-1
APPENDIX A
APPENDIX B
-------
These estimates represent reductions of 50\ for ingestion
exposure and 90% for inhalation exposure as compared to EPA's
estimates. The relative potency estimates for other
carcinogenic PAHs were found on the basis of data from nine
experimental studies to range from 0.0044 to 4.50 as compared to
B[a]P. The estimates based on the most reliable of the studies
were selected and are summarized in Table IV-26; this table is
reproduced below. The estimates were found not to vary to a
great extent among studies for a particular PAH.
The results of this study indicate that to estimate the
cancer risk due to a mixture of PAHs in the environment, the
concentrations of carcinogenic PAHs should be measured and used
in conjunction with relative potency values to estimate the
total risk.
>(
SUMMARY OF RELATIVE POTENCY ESTIMATES
DERIVED FOR PAHs
Benzo[a]pyrene 1.0
Benz[a]anthracene 0.145b
Benzo[b]fluoranthene 0.140a
Benzo[k]fluoranthene 0.066s
Benzo[ghi]perylene 0.0228
Chrysene 0.0044C
Dibenz [-ah] anthracene 1.11°
Indeno[l,2,3-cd]pyrene 0.232a
•Deutsch-Wenzel et al. (1983).
*>Binghara and Falk (1969).
CWynder and Hoffmann (1959).
vi
-------
TECHNICAL SUMMARY AND CONCLUSIONS
«
A relative potency approach for estimation of the
carcinogenic risk of complex mixtures of polycyclic aromatic
hydrocarbons (PAHs) was developed as an alternative to the
current practice of assuming that all carcinogenic PAHs are
equivalent in potency to benzo[a]pyrene, which has little
scientific support. B[a]P has consistently been demonstrated
to be one of the most potent carcinogenic PAHs to which humans
might be expected to be exposed environmentally. As a result,
estimates of cancer risk using a B[a]P one-to-one equivalency
approach will greatly overestimate the carcinogenic potency of
most mixtures of PAHs. Use of a relative potency approach that
takes into account the differing potencies of carcinogenic PAHs
would yield a more realistic estimate of risk, with a sounder
biological basis.
In this report, a new method is developed for estimating
the cancer risk associated with exposure to mixtures of PAHs
that attempts to rectify the problems inherent in earlier
approaches in use by EPA and others. This study focuses on
three critical elements. First, a biological paradigm was
proposed for PAH-induced carcinogenesis. Such a paradigm is
essential to provide a biological rationale for the
dose-response model and to account for any interactive effects
(i.e., synergistic or antagonistic) that might be postulated.
Second, a two-stage mathematical model was postulated that is
consistent with the paradigm. The model parameters were
vii
-------
estimated using rodent tumor response data following exposure
*
to B[a]P. The two-stage model is a special case of the
Moolgavkar and Knudson (1981, 1986) cancer risk model that was
adopted by Thorslund et al. (1987) to account for exposure to
known levels of carcinogenic agents. The model may ajso be
viewed as a special case of the classic Armitage and Doll
(1954) multistage model in its time-independent form, and a
restricted case of the Armitage and Doll (1957) two-stage model
in both its time-independent or -dependent forms. The
advantages of the model are that it is based on a strong
theoretical argument derived from biological principles yet is
simple enough to obtain estimates for its parameters using very
limited data. Using this model, the estimate of cancer potency
for B[a]P was lowered by 50% for ingestion exposure and by 90%
for inhalation exposure as compared to EPA's estimates.
The third critical element in the development of the method
was obtaining estimates of relative potency for carcinogenic
PAHs other than B[a]P using the structural form of the
previously derived model. This aspect depended upon using
bioassay results that were obtained from systems that are not
suitable for direct extrapolation to humans because the routes
of exposure employed are not comparable to those by which
humans are exposed in the environment. When carcinogenic PAHs
and B[a]P have been tested concurrently, however, such
experiments can be used to estimate the relative potencies of
other PAHs compared to B[a]P. These relative potencies have a
specific biological interpretation: they provide an estimate
of the ratio of exposure-induced mutation rates per unit of
viii
-------
exposure. Under the assumptions of the model, these-mutations
i
are thought to transform a normal stem cell into a
preneoplastic cell and a preneoplastic cell into a malignant
cell. Using nine experimental studies, estimates of potency
for carcinogenic PAHs were obtained that ranged from 0.0044 to
4.50 as compared to B[a]P. The estimates were consistent among
studies for a particular PAH. Those that were obtained from
the most reliable studies are summarized in Table IV-26 (and on
page vi).
Once the relative potency estimates were obtained, they
were used in conjunction with the B[a]P dose response model
that was derived and the dose additivity assumption to obtain
estimates of cancer risk associated with any specified exposure
to multiple PAHs. The dose additivity assumption is synonymous
with what is referred to by Finney (1964) as simple similar
action and advocated for use in the EPA (1986b) guidelines for
assessing the cancer risk of mixtures when inadequate data are
available. Simple similar action implies that all carcinogenic
agents in a mixture induce carcinogenesis by similar
mechanisms. Since carcinogenic PAHs appear to be metabolized
to similar reactive derivatives, produce comparable adducts
with DMA and the same types of tumors in experimental animals,
the assumption of a common mechanism of action (i.e., dose
additivity) is plausible.
Under the assumption of dose additivity, the cancer risk
due to exposure to multiple PAHs can be obtained as follows.
The total exposure units equivalent to B[a]P in a mixture to
which an individual is exposed is calculated by taking the sum
ix
-------
of the products of the relative potencies and the exposure
levels for each PAH. These B[a]P-equivalent exposure units are
then substituted into the dose response model for B[a]P to
obtain the cancer risk associated with exposure to the PAH
mixture. This procedure is evaluated in this document using
bioassay results from an experiment conducted by Schmahl et al.
(1977) using two different mixtures of PAHs. The resulting
predictions were encouragingly close to the tumor rates
observed.
For the sake of comparison, more familiar but less
defensible models from a biological theory veiwpoint have also
been used in this report to estimate relative potencies.
Results reasonably comparable to those from the two-stage model
\
were obtained using the one-hit and linearized multistage
models/ suggesting that the results are not highly
model-dependent.
•
This study has demonstrated that estimation of the total
cancer risk due to exposure to a mixture of PAHs should be
based on measurements of the concentrations of its carcinogenic
components and relative estimates of their potency.
-------
I. INTRODUCTION
The majority of occupational and environmental- chemical
exposures are to complex mixtures. Emissions from combustion
sources and leachates from hazardous waste sites are examples of
complex mixtures that contain hundreds of components that have
not been completely characterized. Estimating the potential
human health hazards associated with such mixtures is difficult
because of inadequate chemical and toxicological
characterization, as well as the extent of heterogeneity between
similar mixtures from different sources.
For complex mixtures of polycyclic aromatic hydrocarbons
(PAHs) such as those found at hazardous waste sites, the"
standard approach to estimating the human cancer risk associated
with exposure has been to assume that the carcinogenic PAH
components (excluding heterocyclic compounds such as the
nitroaromatics) are equivalent by weight to benzo[a]pyrene
(B[a]P) in terms of their carcinogenic potency (EPA 1980,
1984). Based on a limited chemical characterization of the
mixture, a number of PAHs are identified and their contributions
to cancer risk are estimated based on their relative
concentrations in the mixture/ their expected contributions to
human exposure, and the statistical upper-bound on the potency
of B[a]P. This procedure has several serious problems.
The assumption that all the carcinogenic PAHs have
potencies equivalent to that of B[a]P is not supported by the
scientific literature. In studies that have been performed to
1-1
-------
evaluate the comparative potencies of PAHs, B[a]P has
consistently proved to be one of the most potent carcinogens
tested. These studies will be described in detail in Section
IV. Studies of the carcinogenic potencies of complex mixtures
of PAHs as compared with B[a]P have shown that most mixtures are
considerably less potent than B[a]P alone. For example, various
kinds of combustion emissions and B[a]P were tested for potency
as tumor initiators on the skin of SENCAR mice (Slaga et al.
1980). Table 1-1 shows that the PAH mixtures were much less
potent as tumor initiators than B[a]P. The authors calculated
relative potency estimates that ranged from 0.007 for coke oven
emissions extract to less than 0.002 for diesel engine exhaust
extract, using papillomas/raouse-mg as the end point. In another
study, the tumorigenicity of an automobile emission condensate
s
(AEC), a diesel emission condensate (DEC), a representative
mixture of carcinogenic PAHs, and B[a]P were tested for
carcinogenicity by chronic application to mouse skin (Misfeld
1980). The results are shown in Table 1-2. Relative potencies
were calculated by the authors to be 0.00011, 0.0053, and 0.36
for DEC, AEC, and the PAH mixture, respectively, as compared to
B[a]P. Such results make the use of the assumption that all
carcinogenic PAHs are as potent as B[a]P to predict the total
carcinogenicity of PAH mixtures difficult to support.
Another problem with the ZPA's use of B[a]P as a surrogate
for the components of PAH mixtures is that it assumes that the
total cancer risk of the mixture can be predicted on the basis
1-2
-------
TABLE 1-1
RELATIVE TUMOR-INITIATING POTENCY OF VARIOUS
EMISSION EXTRACTS AND BENZO[a]PYRENE
Relative Potency13
Based on
Substance8 (papillomas/mouse-mg)
Benzo[a]pyrene 1.0
Roofing-tar emission extract 0.004
Coke-oven emission extract 0.007
Caterpillar diesel exhaust extract C0
Oldsmobile diesel exhaust extract 0.002
Nissan diesel exhaust extract 0.007
Mustang gasoline-engine exhaust extract 0.002
Cigarette-smoke condensate C0
"Material was applied to Sencar mice once as initiator.
Phorbol myristate acetate (2 ug)/ twice a week, was used as
promoter. Emission exposures that were used in the relative
potency calculations were restricted to the linear portion of
the dose-response curve. \
bAs calculated by authors.
GThese entries refer to potencies that were not
significantly different from zero at p-0.05.
Sources: Slaga et al. (1980) and NAS (1983).
1-3
-------
TABLE 1-2
CARCINOGENIC ACTIVITY OF AUTOMOBILE EMISSION CONDENSATE (AEC),
DIESEL EMISSION CONDENSATE (DEC), AND PAHs
ON MOUSE SKIN
Treatment
%Mice with Tumors
Relative
Potency*3
Solvent control
Benzo[a]pyrene:
3.86 ug
7.69 ug
15.4 ug
AEC:a
290 ug
880 ug
2,630 ug
DEC:3
4,300 ug
8,600 ug
17,150 ug
Mixture of PAHs:
3.5 ug
10.5 ug
0
32.8
60.9
89.1
10.3
44.3
83.3
0
2.6
12.7
1.3
38.7
1
0.0053
0.00011
0.36
aObtained with leaded fuel.
bAs calculated by authors.
Sources: Misfeld (1980) and NAS (1983)
1-4
-------
of response-, additivity. The EPA guidelines for assessing the
cancer risk of chemical mixtures in the absence of combined
exposure data, however, advocate the assumption of dose
additivity (1986b). Dose additivity is synonymous with what is
referred to by Finney (1964) as simple similar action. When
agents induce cancer by the same mechanism, they are
characterized by simple similar action and the cancer risk due
to combined exposure can be obtained by estimating the risk due
to the total additive dose of the agents.
To use this method for mixtures of PAHs, estimates of the
relative potencies of the components are required. The relative
potency of the jth carcinogenic PAH compared to the potency of
B[a]P at response level p can be defined as
where x(p) and y^(p) are the number of exposure units of B[a]P
and the jth carcinogenic PAH, respectively, required to produce
a total carcinogenic response rate of p in the test system.
If the mechanism of action of B[a]P is the same as that of
the jth carcinogenic PAH, it follows that Rj(p) • Rj (i.e.,
the relative potency is independent of the response level).
Under the hypothesis of simple similar action, it follows
directly that the joint response to a set of carcinogenic PAHs
is dependent upon the total PAH exposure, T, expressed in units
equivalent to B[a]P. This exposure may be written as
1-5
-------
m
T - I RjYj * x (1-1)
j-l
where
m - the total number of indicator PAHs exclusive of
B[a]P;
yj • the exposure to the jth indicator PAH;
x • exposure to B[a]P; and
RJ - relative potency of the jth indicator PAH
compared to B[a]P.
The probability of a cancer response given T is P(T), where
P(') is the dose-response relationship for B[a]P. For this
approach to be valid, the function P(') must be monotonic in
the experimental range with values of x and y., expressed in
the same units.
In the case of mixtures of carcinogenic PAHs, dose
additivity has been postulated to be a reasonable assumption
because they appear to be metabolized to similar reactive
derivatives that interact with DNA and produce similar tumors
(IARC 1983). Kaden et al. (1979) found a simple additive
relationship of the mutagenicity of soot and added B[a]P.
Schmahl et al. (1977) tested B[a]P and two mixtures of
carcinogenic and noncarcinogenic PAHs when applied to the backs
of mice. The tumor response is shown in Section VI of this
document to be predictable on the basis of additivity of doses
of the carcinogenic mixture components. Assuming simple similar
action (i.e./ dose additivity) for the purpose of assessing the
cancer risk of PAH mixtures seems plausible.
The potential for interactions (synergism or antagonism)
between individual PAHs is another problem that is not addressed
1-6
-------
by the EPA's use of B[a]P as a surrogate for the carcinogenic
components of PAH mixtures. The agents in any complex mixture
can interact to affect carcinogenesis in a variety of ways.
Chemical interactions between agents may create different
carcinogenic agents. An example of this in drinking water is
the interaction of chlorine used as a bactericide with naturally
occurring organic matter to form trihalomethanes (Bellar et al.
1974, Rook 1974). New compounds may form within the body as
well. For example/ nitrosation of certain compounds in fava
beans by endogenous nitrite, when both are present in the
gastric lumina, leads to the formation of a potent,
direct-acting mutagenic nitroso compound (Yang et al. 1984).
Complex mixtures can also act to modify the exposed
\
individual so that the dose at the site of action for one agent
is dependent upon the exposure levels of the other agents in the
mixture. Any event that affects the absorption, distribution,
metabolism, or elimination of a compound will affect the level
of that compound that is available to react with DNA or other
target molecules. For example, cigarette smoke can induce the
levels of enzymes that metabolize PAHs (Conney et al. 1977),
resulting in higher intracellular levels of reactive derivatives
capable of modifying DNA.
All such chemical-biological interactions are the result of
reactions at many cellular sites with multiple molecules of
agents and are probably unimportant at low doses. Mathematical
models of the cancer response that depend upon such interaction
mechanisms would be expected to be nonlinear at low doses. For
1-7
-------
example/ if two chemicals combined to form a carcinogenic agent/
*
the rate of formation would be proportional to the product of
the concentrations of the two chemicals. A linear reduction in
the concentrations of the chemicals would thus result in a
quadratic reduction in the formation of the carcinogenic agent.
The nonlinearity of the typical chemical-biological inter-
action strongly suggests that mechanisms of carcinogenicity that
depend upon such interactions are only marginally important at
exposure levels several orders of magnitude below observed
responses. Interaction terms that dominate a carcinogenic
response at high exposure levels become relatively insignificant
at low exposure levels. Even so, any information about chemical
interactions or exposure modification should be used in the
formulation of a model of the combined effects of agents/ if
available, by estimating exposure at the cellular and molecular
levels.
Interactions may play a large part in carcinogenesis
resulting from experimental exposure to PAHs. Mahlum et al.
(1984) have shown that different temperature range distillates
of coal liquids have different skin tumor-initiating activities
in mice despite the fact that they contain similar levels of
known carcinogenic PAHs. This difference is believed to be due
to the modifying effects of the spectrum of noncarcinogenic PAHs
obtained at different temperatures. Most of the PAH components
of coal liquid fractions obtained at different temperatures will
vary both qualitatively and quantitatively and* as a result,
their abilities to modify carcinogenesis will vary accordingly.
1-8
-------
This report will examine the biological basis for the
mechanisms bf action of carcinogenic PAHs and use what is known
about such mechanisms to develop a mathematical dose-response
model to predict cancer risk resulting from exposure to mixtures
of PAHs in the environment. The model is linear at low doses
because the mechanisms of interaction that affect tumor rates at
the high doses used in the laboratory are hypothesized not to
operate at the low doses to which humans are exposed in the
environment. This method uses a relative potency approach to
account for differing potencies of PAHs as compared to B[a]P/
assumes dose additivity, and has fewer limitations than that
currently employed.
1-9
-------
II. BIOLOGICAL MECHANISMS OF ACTION OF POLYCYCLIC
AROMATIC HYDROCARBONS
This chapter describes the postulated biological mechanisms
of action of PAHs that cause cancer. A series of steps appears
to be involved between exposure to PAHs and tumor development;
these steps are described below. Each step is expected to occur
in a similar manner but to a different extent for different
PAHs, which is why their potencies differ. The form of the
mathematical dose-response model described in the next chapter
depends on the biological mechanisms of action of PAHs.
This section is not intended to be a review of the vast
literature available on PAHs but to provide an underlying
theoretical rationale for the dose-response relationship to be
'used for B[a]P-induced cancer risk. Also discussed is how the
information that is available for B[a]P may be extended to other
PAHs. The organization of this section will follow that
outlined in Figure II-l, which shows the proposed basis for
PAH-induced carcinogenesis.
The paradigm for PAH-induced carcinogenesis can be
summarized as follows. PAHs can be bound to a soil matrix at a
hazardous waste site and upon human exposure, a certain fraction
of the soil-bound PAHs is released and available for absorption
into the body. PAHs are readily distributed throughout the body
and are metabolized enzymatically to a number of water-soluble
derivatives. A small fraction of some PAHs, if it is not
conjugated .or deactivated quickly, can be metabolized to a
II-l
-------
, FIGURE II-l
BIOLOGICAL PARADIGM FOR PAH-INDUCED CARCINOGENESIS
matrix-PAH
I
bioavailable dose
distribution
metabolism
DNA adducts
cell proliferation
repair
mutation
repair
cell proliferation
death
cell proliferation
a
DNA adducts cell proliferation
second mutation
I
death
4-
cancer
II-2
-------
reactive fojrm that may interact with cellular DNA. Once formed,
DNA-PAH adducts may be repaired, or if cell proliferation occurs
prior to repair, a mutation may occur. The mutated cell may die
as a result of the mutation, or may survive and undergo further
proliferation and mutation, which can under certain
circumstances result in the formation of a malignant tumor. The
rest of this section describes each of these steps in greater
detail.
1. PAH Matrix and Bioavailabilfcv
Studies of many complex mixtures have shown that the
toxicity of a mixture as it occurs in the environment seldom
approaches the theoretical toxicity of the sum of the individual
components based on experiments in which pure chemicals are
administered individually (Vostal 1983, Umbreit et al. 1986).
This finding suggests either that some of the components, or
some portions of each component, are not absorbed or that some
antagonism among the components occurs.
Bioavailability can be defined as the fraction of a
compound in a matrix that is released from that matrix, absorbed
by an organism, and, hence, is available to elicit a biological
effect. The release and uptake of a compound from a matrix
constitute facets of the bioavailability of that compound,
although its biological effect is often used as the measure of
its bioavailability. Since risk is considered to be a function
of both exposure and toxicity, bioavailability is an important
consideration in order to determine potential risk from
II-3
-------
components of a complex mixture. The relative proportions of
*
mixture components that reach a tissue are not necessarily the
same as that found in the original mixture, and can therefore
not be predicted solely by chemical analysis of a mixture.
Little work has been done to determine the bioavailability
of PAHs from soil. One study showed that when B[a]P was
adsorbed to carbon particles and instilled into the lungs of
mice, although clearance of the particle-bound B[a]P was slower
than that of unbound B[a]P, less than 15\ of the particle-bound
B[a]P was eluted from the particles and available to react with
respiratory tissue (Creasia et al. 1976).
In another study, Vostal (1983) noted that the in vitro
mutagenic activity of diesel particulates, of which PAHs are a
major component/ was minimal or negative when tested in extracts
obtained using biological fluids; furthermore, chronic
inhalation exposure of animals to high concentrations of diesel
particles did not induce the activity of hydrocarbon-
metabolizing enzymes or adversely affect immune response unless
organic solvent extracts of the particles were administered
intratracheally or parenterally at very high doses. The lack of
biological response to diesel particulate matter in the absence
of radical treatment with solvents to extract PAHs indicates
that environmental exposure to humans could result in a much
smaller toxic effect than would be predicted due to limited
availability. In some cases, however, gradual leaching of PAHs
from retained particles has been shown to lead to an enhanced
tumor response as compared to that obtained with bolus dosing.
II-4
-------
Studies of the bioavailability from soil of 2,3,7,8-TCDD, a
toxic chlorinated polycyclic compound with some physical
characteristics in common with PAHS, have shown that
bioavailability is a function of soil type, varying from 30% to
less than 1% (Umbreit et al. 1986). If PAHs are tightly bound
to their soil matrix, as is suspected to be the case at
hazardous waste sites such as former manufactured gas plants,
the actual exposure to PAHs at the sites will be much less than
predicted from simple exposure models because of restricted
bioavailability. Moreover, differences in physical properties
among the components of PAH mixtures are expected to result in
differential bioavailability, compounds of higher molecular
weight being more strongly bound to soil than those of lower
molecular weight. Differential bioavailability may affe'ct
relative potencies. Bioavailability cannot be incorporated into
the dose response model for PAHs at this time, however, due to
the paucity of data.
2. Distribution
The distribution of many PAHs throughout the body following
absorption appears to occur widely and readily. Wide
distribution of B[a]P to a variety of body tissues can occur
following administration by various routes, and the pattern of
distribution is generally similar for all routes of exposure
(except for high local pulmonary concentration following
intratracheal instillation) (Kotin et al. 1959). IARC (1983)
stated \
II-5
-------
PAH€ are essentially devoid of polar and ionizable
functional groups and would therefore be expected to
dissolve readily in and cross the lipoprotein
membranes of mammalian cells. The demonstrated
tozicity of many PAHs in organs remote from the site
of their administration confirms this expectation.
Furthermore, the fact that isolated cells and tissues
metabolize PAHs by means of intracellular enzymes, and
that some of these metabolites react with
intracellular constituents suggests that uptake across
cellular membranes is an easy process.
..... Levels of PAHs observed in any particular
tissue are dependent on the PAH administered, the
route and vehicle of administration, the
post-administration sampling times and the presence of
inducers of PAH metabolism. Nevertheless., results of
a number of studies (Heidelberger & Weiss, 1951; Kotin
££_&!., 1959; Bock & Dao, 1961; Mitchell, 1982)
indicate that (i) detectable levels of hydrocarbon can
be observed in most internal organs from minutes to
hours after administration; (ii) mammary and other fat
tissues are significant storage depots where
hydrocarbons may accumulate and be slowly released;
and (iii) the gut contains relatively high levels of
hydrocarbons or hydrocarbon metabolites as the result
of hepatobiliary excretion of metabolites or of
swallowing of unmetabolized hydrocarbon following
mucociliary clearance after inhalation exposure.
*
Distribution of PAHs to target tissues thus appears to
occur readily (IARC 1983). For the purposes of relative potency
estimates and risk assessment, the distribution (equilibrium
partitioning) of PAHs will be assumed to resemble that of B[ajr.
3. Metabolism
The metabolism and activation of PAHs have been well
reviewed by several authors (Sims and Grover 1974, Conney 1982,
Cooper et al. 1983). The general scheme for PAH metabolism is
depicted in Figure II-2 for B[a]P. This figure illustrates the
multitude and complexity of the metabolic reactions and
II-6
-------
FIGURE II-2
METABOLIC FATE OF BENZO[A]PYRENE (FROM IARC 1983)
CSH
cenjuqotti
BENZO(«]PTRENE
JO
ASENC OXIDES
4,3-
7.8-
9.10-
DIHYPRQDIQLS
4.3-
7.8-
9.10-
• PHENOL PIOLS
9-OM-4.3-dioi
6-OM-7.8-cJ.ol
M3-)OH-9.lO-dioi
ClucuronidM '
•nd
Mlphat* ttrtri
GSM conjugott*
OIOL CPOXIOCS
TCTMAOLS
9.iO-di«-7,8-«poii«c
Benzo[a]pyrene is metabolized initially by the microsomal
cytochrome P-450 monoxygenase system to several arene oxides
(reaction 1, Fig. 2). Once formed, these arene oxides may
rearrange spontaneously to phenols (reaction 3), undergo
hydration to the corresponding trans-dihydrodiols in a reaction
catalysed by microsomal epoxide hydrolase (reaction 4), or react
covalently with glutathione, either spontaneously or in a
reaction catalyzed by cytosolic glutathione S-transferases
(reaction 5). Phenols may also be formed by the cytochrome
P-450 monooxygenase system by direct oxygen insertion
(reaction 2), although unequivocal proof for this mechanism is
lacking. 6-Hydroxybenzo[a]pyrene is further oxidized either
spontaneously or metabolically to the 1,6-, 3,6-, or
6,12-quinones (reaction 6), and this phenol is also a presumed
intermediate in the oxidation of benzo[a]pyrene to the three
quinones catalysed by prostaglandin endoperoxide synthetase
(Marnett et al. 1977, 1979). Evidence exists for the further
oxidatiye metabolism of two additional phenols;
3-hydroxybenzo[a]pyrene is metabolized to the 3,6-quinone
(reaction 6), and 9-hyrdoxybenzo[a]pyrene is oxidized to the
K-region 4,5-oxide, which is hydrated to the corresponding
4,5-dihydrodiol (reaction 7). The phenols, quinones and
dihydrodiols can all be conjugated to glucuronides and sulphate
esters (reactions 8-10), and the quinones also form glutathione
conjugates (reaction 11).
II-7
-------
; FIGURE II-2 (Continued)
*
METABOLIC FATE OF BEN20[A]PYRENE (FROM IARC 1983)
In addition to being conjugated, the dihydrodiols undergo
further oxidative metabolism. The cytochrome P-450
monooxygenase system metabolizes benzo[a]pyrene 4,5-dihydrodiol
to a number of uncharacterized metabolites, while the
9,10-dihydrodiol is metabolized predominantly to its 1- and/or
3-phenol derivative (reaction 12), with only minor quantities of
a 9,10-diol, 7,8-epoxide being formed (reaction 14). In
contrast to 9,10-dihydrodiol metabolism, the principal route of
oxidative metabolism of benzo[a]pyrene 7,8-dihydrodiol is to a
7,8-diol, 9,10-epoxide (reaction 14), and phenol-diol formation
is a relatively minor pathway. The diol epoxides can be
conjugated with glutathione either spontaneously or by a
glutathione S-transferase-catalysed reaction (reaction 15).
They may also hydrolyze spontaneously to tetraols (reaction 16,
although epoxide hydrolase does not catalyze the hydration).
II-8
-------
transformations that PAHs may undergo. The majority of PAH
»
metabolites are conjugated as water-soluble derivatives, which
facilitates elimination. Some fraction of PAH metabolism may
lead to the formation of reactive diol epozides. The PAH diol
epozide metabolites are thought to be primarily responsible for
the ultimate carcinogenic activity of the biologically active
PAHs following reactions with critical sites on DNA (Sims et al.
1974). The formation of diol epozides from B[a]P is outlined in
Figure II-3. This figure illustrates that four optically active
7,8-diol-9,10-epoxides are formed from B[a]P. These
stereoisomers differ in their rate of formation and in
biological activity; the metabolically predominant stereoisomer
in rats is also that which possesses the greatest carcinogenic
potency (Levin et al. 1982).
Metabolism of PAHs in humans is qualitatively similar to
that in a variety of animal tissues, however, quantitative
•
differences are found both among species and among different
human tissues and may be influenced by dose and the presence of
other PAHs; the susceptibility of a tissue to PAH carcinogenesis
may be partly a result of its ability to activate and detozify
PAHs. The potency of PAHs will clearly be related to the eztent
to which they are metabolized to diol epozides, the rate at
which the diol epozides are themselves metabolized, and the
eztent to which those that are formed are capable of interacting
with DNA at critical sites.
II-9
-------
4. DMA Adduct Formation
i
The term "DNA adduct" refers to a complex between part of a
DNA molecule and a molecule of xenobiotic (or a metabolite).
Covalent binding of reactive metabolites of B[a]P and other
carcinogenic PAHs to DNA can occur and results in the .formation
of DNA adducts. This step appears to be essential for the
s
production of PAH-induced neoplasia (Gelboin 1980, Weinstein et
al. 1978). If the adducts are not repaired, replication of
damaged DNA during cell proliferation may lead to mutation. Tl._
production of neoplasia may be related both to the extent to
which PAHs bind to DNA, and the level of cell proliferation in
the target tissue.
Pereira et al. (1979) applied single doses of B[a]P
^
topically to ICR/Ha mice and found a linear relationship betwet«
the applied dose and the amount of each of the two major
B[a]P-i)NA adducts formed in the epidermis over a dose range of
•
0.01 to 300 mg B[a]P per mouse. The level of adduct formed
reached a maximum at 7 hours and remained constant for the next
49 hours, indicating that repair of such lesions is slow.
Adriaenssens et al. (1983) administered single doses of
B[a]P orally via intubation to rats and examined the formation
of DNA adducts over a dose range of 2 to 1,351 umol/kg (0.048 to
29.7 umol/mouse). The rate of formation of the major diol
epoxide-DNA adduct was linear with respect to dose in the
forestomach and, although the authors stated otherwise, in the
lung as well. The extent of the deviation from linearity in tl._
11-11
-------
lung was slight and consistent with the level of error
associated with this type of measurement.
PAH metabolites have been found to bind to DNA in every
tissue that has been examined, regardless of species, dose, or
route of administration. The physical nature of the adducts
formed/ the levels at which they occur, and the extent to which
they persist unrepaired appear to be similar in various tissues,
i
whether or not those tissues are likely to develop tumors
(Stowers and Anderson 1985). Thus, formation of PAH-DNA
adducts appears to be a necessary, but not sufficient step for
PAH-induced carcinogenesis; another step is also required.
5. Cell Proliferation
\
The rate of cell proliferation can have an important
influence on the rate at which mutation occurs and may account
for the differences in susceptibility among tissues.
Experiments with radiation and chemically-induced mutagenesis
have indicated that at least one round of cell proliferation is
required, before the mutational change becomes permanent and
cannot be repaired (Borek and Sachs 1968, Kalanuga 1974).
Increasing the rate of cell proliferation in a tissue can
increase the susceptibility of that tissue to mutation (Craddock
1971). Cells that are replicating are known to be of greater
susceptibility to mutation than those that are not (Tong et al.
1980). Cells that are not actively proliferating have time to
11-12
-------
repair DNA damage, making the occurrence of a mutation much less
likely.
The role of cell proliferation in PAH-induced
carcinogenesis is two-fold:
(1) Tissues that have higher rates of cell turnover, such
as the skin and lung, appear to be target tissues for tumor
formation, while those with slower rates, such as the liver, ai_
not. For example, B[a]P does not usually cause liver tumors in
rats. Performing partial hepatectomy on rats leads to a high
rate of hepatic cell turnover. If partial hepatectomy is
performed prior to B[a]P administration, tumors will occur in
that organ (Hirakawa et al. 1979).
(2) PAH administration can increase cell proliferation as a
consequence of toxicity. The formation of bulky DNA adducts arj
adducts with other cellular macroroolecules such as RNA and
protein can lead to toxicity and cell death. One response of a
tissue to toxicity is to increase cell proliferation in order to
replace damaged cells. Epidermal hyperplasia has been shown to
occur following topical application of B[a]P and its diol
epoxide to the backs of C57BL/6J mice (Bresnick et al. 1977). A
linear dose-response relationship occurred between the number of
layers of epidermis following a single dose of the carcinogenic
diol epoxide of B[a]P over the dose range of 0.2 to 1.2
umoI/mouse; the increase persisted for 4 days. These
observations indicate that, at least at high doses, B[a]P and
its metabolites may not only form DNA adducts, but may increase
11-13
-------
III. . DEVELOPMENT OF A COMPARATIVE POTENCY APPROACH
4 FOR THE ASSESSMENT OF CANCER RISK
ASSOCIATED WITH MIXTURES OF POLYCYCLIC
AROMATIC HYDROCARBONS
This chapter describes the mathematical models that could
be used to characterize the dose response relationships between
PAHs and cancer using laboratory animals and thereby to predict
the human cancer risk associated with exposure to mixtures of
%
PAHs. The models are derived, put in a biological context, and
their strength and weaknesses noted. A general model is
recommended and made specific for B[a]P using data from
experiments in which mice were fed or hamsters inhaled B[a]P
and their tumor responses were determined. The assumptions
that are required to convert a dose response model based on
\
animal experiments into one that predicts lifetime human cancer
risk, if levels of human exposure are known, are described and
used to obtain a human dose response model for B[a]P. Finally,
•
a method is derived that uses this model in conjunction with
relative potency estimates for other PAHs to estimate the human
cancer risk associated with exposure to mixtures of PAHs.
1. Classic Multistage Model
Armitage and Doll (1954) derived a mathematical model for
carcinogenesis that assumes a complex, time-dependent process
involving a specific ordering of at least two qualitatively
different cell stages. Under the assumption that the
transition rate (r^) between the i-1 and ith cell stages is
constant over time (i • l,2,...k) and that a total of k stages
III-l
-------
must occur to produce a neoplastic cell, the probability that 'a
tumor will occur by age t in the absence of competing risk is
P(t> - l-e
In this formulation, it is assumed that each of the N
susceptible cells in the target organ is independently
undergoing transition to its next stage. It is also assumed
that one or more of the transition rates is a linear function
of the exposure level x. Using the model parameters, this is
equivalent to specifying
ri • ai * bix i • 1/2,..., k,
>
where a^ > 0 and b^ i 0. The latter condition excludes the
possibility that the carcinogen may be inhibitory (i.e., reduce
the transition rate) at any stage. The number of stages, m,
»
that have transition rates that are exposure-dependent defines
the degree of the polynomial relationship. The exponent of the
multistage model in this form is a constant times the product
of m linear equations with non-negative coefficients. In the
resulting polynomial, each of the coefficients associated with
terms of order m or less exist and are subject to certain
nonlinear constraints. For example, if the coefficient
corresponding to .the jth power is denoted as C^ and two
stages are exposure-dependent (i.e., m - 2), it follows that
Cj > 0 for j » 0,1,2 and C^ i 2(C0C2)1/2 (see
Kalbfleish et al. 1983).
IXI-2
-------
It is jioteworthy that this form of the multistage model
has seldom been used in risk assessments but is usually
referred to as the main rationale for low-dose linearity. Its
advantages are that it stems from a reasonable biological
theory of carcinogenesis and yields a positive, nonzero maximum
likelihood estimate of the linear term. Its disadvantages are
that the nonlinear constraints on the C^'s often create
estimation problems and that dose-response relationships with
high upward curvature could not be fitted with this model.
2. Linearized Multistage Model
To avoid the problems posed by the classic multistage
model, Guess and Crump (1977) "generalized" the multistage
N
model by giving it a less restrictive polynomial form:
P (x,t) - l-<
m 4u
- [ I c ,x3 ] tk
where Cj JL 0 for j » 0,1,...,m. When the model is fitted to
data from an experiment with (n+1) groups of animals (usually
one control group and n dose levels), the constraint m ± n is
applied.
As a first step when using the model, the maximum
likelihood estimates for the C's are obtained. Next, a unique
polynomial is derived that gives the maximum risk at a
specified level of exposure, subject to a goodness-of-fit
constraint. The linear tern in this polynomial times an
III-3
-------
model is depicted in Figure III-l. According to the model, the
i
population of cells at risk is proliferating cells, often
referred to as stem cells. Stem cells are those from which
most other cells in an organ arise. A stem cell can
differentiate and leave the pool of proliferating cells,
becoming no longer susceptible to heritable alterations of its
DNA. A normal, susceptible stem cell may do one of several
things. It may divide into two daughter stem cells, terminally
differentiate, die, or undergo mutation at a critical site that
results in formation of a preneoplastic or intermediate cell.
A preneoplastic cell may be defined as a cell that has
undergone one of the changes necessary to become a cancer cell
but is not yet cancerous. The preneoplastic cell may, in turn,
divide into two daughter preneoplastic cells, differentiate,
die, or undergo further mutation at another critical site to
produce a cancerous cell. The cancerous cell will, after a
sufficient length of time, divide into enough cells that it
becomes a detectable tumor. All of these processes can be
described mathematically by postulating specific
exposure-dependent rates for the cell changes. Moolgavkar and
Knudson (1981) showed that to a close approximation, the
age-specific tumor rate at time t for their two-stage model may
be expressed as:
t
I(t) - MOMJ. CQ(V) [exp (B-D)(t-v)] dv (III-3)
III-5
-------
FIGURE III-l
GENERAL BIOLOGICAL STRUCTURE UNDERLYING EXPOSURE- AND TIME-DEPENDENT MODEL
BASED ON THAT DESCRIBED BY MOOLGAVKAR, VENZON, AND KNUDSON
I
ot
Dead Stem Cell
DQ(xtt)
Dead First-
Staqe Cell
MQ(x,t)
Bn(x,t)
•BTs
CQ is a normal, susceptible atea cell.
C. is a transformed, first-stage cell, which can proliferate into a prenalignant clone,
C- is a cancerous cell that vill eventually develop into a detectable tiwor.
Dn(x,t), B0(x,t), and MQ(x,t) are the exposure- and ti«e-dependent death, birth, and
transition or Mutation rates for the noraal stem cell.
D.(x,t), B.Uft), and M,(xvt) are the exposure-'and 11ae-dependent death, birth, and
1 transition or Mutation rates for the first-stage cell.
x is the exposure level, which is as aimed to be constant over tiae.
t is the age of the subject.
-------
where
«
I(t) • age-specific cancer incidence at age t;
MO - transition rate from stem to preneoplastic
cell;
MI » transition rate from preneoplastic to
cancerous cell;
CQ(V) « number of susceptible stem cells per
individual target organ at age v;
B - birth rate or rate of cell proliferation of
preneoplastic cells; and
D • death rate of preneoplastic cells.
Essentially, the equation describes the progression from a
normal stem cell to a cancerous cell under the assumptions of
the two-stage model. This model is a combination of
deterministic and stochastic components. The numbers of
\
preneoplastic and fully malignant cells at any time are assume-9
to be random variables that are dependent upon these event
rates, while the number of normal cells at risk of
•
transformation at time v, denoted by CQ(v), is assumed to be
deterministic and known.
Using this carcinogenesis model, Moolgavkar and Xnudson
(1981) successfully described or explained such phenomena as
• genetic predisposition to cancer;
• patterns in childhood cancer rates;
• hormonally influenced changes in breast cancer rates;
• the results of initiation-promotion experiments for
multiple agents;
• changes in respiratory cancer rates associated with
variable smoking patterns; and
111-7
-------
the fact that for many human carcinomas, the
age-specific incidence rates increase roughly with the
fourth to the seventh power of age.
The Moolgavkar-Knudson model is thus more general in
applicability than the Armitage-Doll model, which was developed
specifically for carcinomas in epithelial cells and predicts an
k 1
age-specific incidence approximately proportional to t ,
where k is the assumed number of stages.
The biological processes described by equation III-3 can
be affected by the level of exposure to carcinogenic agents,
and are more likely to occur as the length of exposure
increases; as a result, they are exposure and time dependent.
Thorslund et al. (1987) have described an exposure- and
time-dependent version of the model. In the context of
biological mechanisms of carcinogenesis, the model can be used
to predict the risk of agents that exert their effects in a
number of different ways. Mutation-inducing initiating agents
would affect the transition rates between cell stages (MQ or
M.), while promoting agents may increase the proliferation
rate of preneoplastic cells (increasing B without affecting
D). Cocarcinogens may increase the proliferation rate of
normal stem cells (CQ), thereby increasing the size of the
target for initiating agents. Inhibitors could remove cells
from the populations of susceptible stem or preneoplastic cells
through toxicity (e.g., increasing D without affecting B) or by
inducing differentiation. Within the context of the model, the
following observations can be explained:
III-8
-------
• tumor rates at very high experimental doses are often
lower than at lower doses, even after adjusting for
differential mortality, as a result of high-dose
- toxicity to stem and preneoplastic cells;
• exposure to multiple carcinogens can give
antagonistic, additive, multiplicative, or
super-multiplicative tumor rate responses by affecting
different model parameters;
• such a response to combined agents appears to be
dose-dependent because some responses such as
increased stem cell proliferation may both occur and
have an impact on tumor rates only at high doses; and
• some preneoplastic, benign cell masses that may be
precursors of tumors will regress and disappear upon
cessation of exposure as a result of a decline in the
birth.rate and no change in the death rate of such
cells.
The multistage model provides no satisfactory explanation for
any of these observations.
4. Restricted Dose Response Model for BFalP
B[a]P is a carcinogenic polycyclic aromatic hydrocarbon
that occurs ubiquitously in air and soil as a product of
combustion. B[a]P can be metabolically activated to form
derivatives that can react with DMA (Sims and Grover 1974,
Gelboin 1980, Weinstein et al. 1984). If B[a]P-DNA adducts
occur at critical sites that control the regulation of a cell's
growth or differentiation, mutations may occur that can lead to
cell transformation and ultimately, cancer. Because the
experimental evidence (Lee and O'Neil 1971) strongly suggests
that B[a]P is a genotoxic agent acting at at least two sites c**
DNA, the two-stage model is a useful approach for estimating
the cancer risk associated with exposure to B[a]P. Without
intermediary experimental information about cell stages and
III-9
-------
differences; in exposure over time, however, it is not possible
i
to estimate the individual background and exposure-induced
mutation rates for preneoplastic and transformed cells using
tumor rate data from a standard animal carcinogenesis bioassay,
nor can the relative transition rates that correspond to each
stage be identified. What can be estimated from bioassay data
are two exposure-induced relative transition rates and a
background transition rate. In the absence of information to
the contrary, the assumption was made that the transition rates
from normal to preneoplastic and preneoplastic to cancerous
cells are equally likely, which could occur if the mutations
that caused the transformations resulted from the same type of
DNA adduct. If these transition rates are linear functions of
\
dose (which is likely at low doses) and the growth rate of
preneoplastic cells is independent of exposure level, the
probability that a tumor will develop by time t as a result of
exposure to a level of genotoxic agent x can be expressed as:
P(x,t) - 1 - exp - (M(l+Sx)2 [exp (Gt) - 1 - Gt]/G2} , (III-4)
where equation III-3 was integrated to obtain the cumulative hazard
function and
M • background tumor rate parameter;
S - fractional increase in the transition rate between
cell stages per unit dose, assumed to be the same for
each stage;
G - the difference between the birth and death rates of
preneoplastic cells, B-D, and is the
exposure-independent growth rate of these cells; and
t - the time (or age) at which risk is evaluated.
111-10
-------
The level of x at the target tissue is assumed to be directly
related to the.administered dose, based on observations of
experiments discussed in Section II in which the rate of
formation of the major B[a]P diol epoxide-DNA adduct was found
to be linearly related with respect to dose in the forestomach,
lung, and skin (Pereira et al. 1979, Adrienssens et al. 1983).
The specific dose-response model derived for B[a]P
(equation III-4) is restricted form of the model developed by
Armitage and Doll (1957), Moolgavkar and Knudson (1981), and
Thorslund et al. (1987). It is restricted by assuming that G
is independent of x and that the two transition rates are
linear functions of x with the same coefficient. The
consequence of these assumptions is that at constant t the
s
dose-response function has only two parameters:
P(x) - 1 - exp-A(l + Sx)2 (ZZZ-5)
where A • M [exp(Gt)-l-Gt]/G and t • age at last observation.
This model has a number of advantages. Among the most
important are the following:
• At low doses, it converges to a linear, no-threshold
form with slope parameter equal to 2AS.
• The model can be fitted to data from an experiment in
which only one exposure group exhibited a positive
response.
• The model has only two parameters that have to be
estimated, so that X2 goodness-of-fit tests can be
run on the data from the standard bioassay with one
control and two exposure groups.
• A stable point estimate of risk can be obtained
directly.
III-ll
-------
• The mathematical form of the model follows directly
from what is currently the most widely accepted
hypothesis of the mechanisms of cancer induction.
• The model is consistent with the mechanisms of
tumorigenesis for B[a]P and other PAHs postulated in
Section II.
It should be noted that equation (III-5) is also a
restricted form of the Armitage-Doll (1954) model, with two
stages affected equally by the carcinogen. Under these
restrictions, equations (III-l) and (III-4) differ only in
their dependence on t. It is not possible to distinguish
between the two models on this basis, however, unless an
integer power of t is inconsistent with the data. In such a
case, the Armitage-Doll model would be rejected.
The assumption that the parameters A and S in equation
s
(III-5) are greater than zero avoids the disadvantages of the
linearized multistage model that are listed in Section III.2;
hence, equation (III-4) or (III-5) can be used to derive stable
•
point estimates of low-dose risk. Such estimates are derived
for B[a]P in the succeeding sections of this chapter, and for
other PAHs in Section IV. However, because equation (III-2)
includes equation (III-5) as a special case, the Guess-Crump
algorithm provides reasonable estimates of an upper bound on
low-dose risk for these chemicals. Such estimates are derived
in Section VI and are compared with point estimates derived in
Sections III and IV, with upper bounds derived from the
multistage and point estimates from the one-hit models.
111-12
-------
5. Dose-Response Pglafcionshin for Oral Exposure to BTalP
The carcinogenicity of B[a]P via ingestion was
investigated in a series of experiments conducted by Neal and .
Rigdon (1967) and Rigdon and Neal (1966, 1969). The most
pronounced dose-response relationship obtained under reasonably
consistent conditions was observed in the Neal and Rigdon
(1967) study. The results of this experiment are reproduced in
Table XIX-1. The end point in this study was either a
papilloma or a sguamous cell carcinoma in the forestomach
(sguamous portion of the stomach), although squamous cell
carcinomas were far more common.
A number of experimental factors that are atypical of the
standard bioassay protocol were employed in the Neal and, Rigdon
(1967) study. These included variable age of first exposure, a
duration of exposure that only lasted about one-seventh of a
lifetime, and an observation period that was less than
one-fifth of a lifetime.
Syracuse Research Corporation (SRC), under contract to the
U.S. Environmental Protection Agency (EPA), had only limited
resources to devote to developing a unit risk estimate for
B[a]P. They therefore decided to use the Neal and Rigdon
(1967) study but restrict the data set to that subgroup of the
experimental population that had been exposed continuously for
the length of the experiment (EPA 1984). In fitting the
linearized multistage mod^l, no adjustments were made for the
variable length of the follow-up period or the age attained at
the end of the observation period. Using the standard EPA
111-13
-------
TABLE III-l
FORESTOMACH TUMORS IN MICE FED BENZO[a]PYRENE
Age First
Exposed
(days)
._
30
30
116
33-67
33-101
31-71
17-22
20-24
18-20
49
56
49
62
49
91
74
48
98-180
Milligrams
of B[a]P per
gram of food
0.0
0.001
0.01
0.02
0.03
0.04
0.045
0.05
0.10
0.25
0.25
0.25
0.25
0.25
0.25
0.25
0.10
0.10
5.0
Number
of days
fed B[a]P
._
110
110
110
110
110
110
107-197
98-122
70-165
1
2
4
5
7
30
7
30
1
Age killed
(in days)
300
140
140
226
143-177
143-211
143-183
124-219
118-146
88-185
155
162
155
168
155
198
182
156
209-294
Number with
Forestomach
Tumors
Number of Mice
0/289
0/25
0/24
1/23
0/37
1/40
4/40
24/34
19/23
66/73
0/10
1/9
1/10
4/9
3/10
26/26
0/10
12/18
17/33
SOURCE: Adapted from Neal and Rigdon (1967)
111-14
-------
approach discussed in Anderson et al. (1983), SRC discarded all
data on animals exposed to 50 ppm B[a]P or more because of
"lack of fit" and obtained a 95% upper-bound estimate on the
chemical's potency to humans of 11.53 (mg/kg/day) . This
number represents an upper bound on the slope of the cancer
dose-response curve for this chemical.
If more detailed data were available on the age at first
exposure and the length of exposure for individual animals,
rather than for exposure groups as a whole, a more precise,
time-adjusted analysis would be possible. However, because
such detailed information is lacking, the approach taken here
is to restrict the analysis to exposure groups that are
comparable with respect to age at first exposure and number of
days exposed. The resulting homogeneous experimental .
subpopulation selected from the entire experimental population
(see Table III-l) is presented in Table XXX-2. This is the
same subgroup used by Chu and Chen (1984).
The two-stage model with identical linear dose-dependent
transition rates (equation III-5) was fitted to the data in
Table III-2. The joint maximum likelihood estimates of the
parameters when there are no animals with tumors in the control
group can yield estimates where A - 0. One way around this
problem is to obtain a positive estimate for A using an
alternative method. Based on a priori biological assumptions,
we know that A>0. This follows from the fact that
^Methods could be developed to use more fully the types
of data that are available, including the information presented
in the other Rigdon and Neal (1966, 1969) papers. However,
such an effort is beyond the scope of the current project.
111-15
-------
TABLE II1-2
DATA USED TO ESTIMATE THE DOSE-RESPONSE RELATIONSHIP
BETWEEN INGESTED B[a]P AND FORESTOMACH TUMORS
Number of
Mice with
Forestoraach Expected Number
Dose Dose (x) Tumors of Tumors Pre-
(ppm in diet) (mg/kg/day) Number Exposed dieted by Model
0
1
10
30
40
45
0.00
0.13
1.30
3.90
5.20
5.85
0/289
0/25
0/24
0/37
1/40
4/40
0.50
0.05
0.20
1.32
2.25
2.72
NOTE: This table contains data only on those groups that are
comparable with respect to age when first exposed and
number of days exposed. The first three columns are from
Neal and Rigdon (1967). The fourth column is based on
the low-dose linear two-stage, dose-dependent,
time-independent, identical transition rate model:
P(x) - 1 - exp [-0.0017256(1 + 0.922x)2]
111-16
-------
if one assumes that B[a]P exerts its effects via a direct
genotozic mechanism of action, the background tumor rate is
roportional to the square of the mutation rate caused by
factors other than B[a]P at the critical gene locus. Since a
variety of other factors such as background levels of-PAHs,
viruses, and solar irradiation have the potential to induce
such mutations, a nonzero background tumor rate estimate is
required for biological consistency. Unfortunately the
"maximum likelihood" estimate of the background tumor rate is
zero (i.e., 0/289 • 0). As a result, an alternative estimation
procedure will be used to estimate the background tumor rate.
When the observed background tumor rate is zero, the multistage
procedure often gives a zero estimate of the true background
\
rate. Such an estimate is illogical from a biological
standpoint, implying that exposure to PAHs is a necessary
condition for the occurence of the tumor in question. The
•
two-stage model is applied in this report so as to make a zero
background rate estimate impossible. The approach taken to
estimate the background rate is to use a Bayesian estimator
with an invariant prior. This approach has been suggested by
Jeffreys (1961, pp. 117-125) and Gart (1966), among others, in
comparable situations where it is known a priori that a zero
estimate is unlikely for a binomial response.
Using this approach, the background tumor rate is
estimated to be
P(0) - 0.5/(n + 1) - 0.5/290 - 0.0017241.
111-17
-------
Equating this value to the parametric form of the two-stage,
dose-dependent, identical transition rate model yields the
following relationship:
P(0) - 0.0017241 « 1 - exp [-A(l + S'O)2] - 1 - exp (-A),
which may be solved directly for A. The solution is
A • 0.0017256, which is put into the model equation to obtain
the relationship:
P(x) - 1 - exp [-0.0017256(1 + Sx)2].
\
The remaining unknown parameter, S, is then estimated by
fitting the above equation in the data in Table III-2. An
approximate maximum likelihood estimate of 0.922 is obtained.
The goodness of fit of the model is also shown in Table III-2.
The expected values are too small to conduct a formal X
goodness-of-fit test. However, it is apparent that the model
does not underestimate the tumor rate or overestimate the
curvature in the experimental range. The data are thus
consistent with the restricted two-stage model represented by
equation III-5.
The low-dose linear term from this model may be expressed
in terms of the parameter values as 2AS, which must be greater
than zero because the background rate A is always greater than
zero and S is greater than zero for all positive dose-response
111-18
-------
relationships. The low-dose linear term in animal exposure
units (mg/kg/day) for a 140-day (30*110) experiment is thus
2AS - q, - (2)(0.001723)(0.922) - 0.00318 (mg/kg/day)"1.
To translate this into a human potency value, two adjustments
are necessary. First, the risk must be adjusted to account for
constant exposure throughout one's entire lifetime. It is
assumed that 2 mice years equal 70 human years (i.e., a
lifetime) and (2 mice years)(365 days/year) - 730 days. EPA's
standard procedures for less than full lifetime follow-up yield
an adjustment factor of (730/140) « 141.8 (Anderson et al.
1983). Second, it is standard practice to calculate •>
species-equivalent exposure units on a mg/surface area basis.
To convert the estimate to these units, it is multiplied by the
additional adjustment factor of (70/0.034)1/3 - 12.72, where
70 and 0.034 kg are the weights of an average human and CFW
mouse, respectively. After making these two adjustments, the
human potency factor is estimated to be
- (0.00318)(141.8)(12.72) - 5.74 (mg/kg/day)"1
for human exposure. This value can be compared to the 95%
upper-bound estimate of 11.53 (mg/kg/day)"1 obtained by EPA's
contractor, which is not an actual estimate; it is only the
upper bound on the actual values and is approximately two times
larger than the linear point estimate obtained here.
111-19
-------
6. Dose-Response Relationship for Inhalafcion Exposure to BFalP
The preferred route of exposure for chronic studies
conducted to assess the risk of cancer posed by airborne B[a]P
is inhalation. However, such studies have difficult
methodological problems associated with them. As a result/
much of the work on this chemical has been done using
intratracheal instillation of B[a]P alone or in combination
with other agents as a surrogate route of exposure for
\
inhalation. Although much valuable qualitative information has
been obtained by this approach, its use in quantitative risk
assessments has not met with much success.
Most carcinogenesis bioassays in which B[a]P has been
administered via inhalation exposure have yielded negative
results. There are several exceptions; most notably, a
positive response was obtained in rats using a combination of
B[a]P and the atmospheric irritant sulfur dioxide; sulfur
dioxide by itself was not carcinogenic (Laskin et al. 1970).
The tumors induced in this experiment were squamous cell
carcinomas; a proportion of the bronchogenic carcinomas found
in humans are of this type.
A well-conducted study by Thyssen et al. (1981) provides
the most clear-cut evidence of the dose-response relationship
between inhaled B[a]P and tumorigenesis. In this experiment,
Syrian golden hamsters were exposed throughout their lives to
B[a]P in a sodium chloride aerosol for 4.5 hours per day, 7
days per week for 10 weeks and then 3 hours per day
thereafter. Respiratory tract tumors were induced in the nasal
IIX-20
-------
TABLE II1-3
DATA USED TO ESTIMATE DOSE-RESPONSE RELATIONSHIP BETWEEN
INHALED B[a]P AND RESPIRATORY TRACT TUMORS
Exposure Average
Rate (z) Survival (t)
(rag B[a]P/m3)
Observed
Effective
Number
(in weeks)
Number of Respiratory
Tract Tumors
Exposed
Predicte
0
2.2
9.5
46.5
96.4
95.2
96.4
59.5
27
27
26
25
0.73
1.88
9.06
12.59
0
0
9
13
NOTE: Based on low-dose linear two-stage, dose- and
time-dependent, identical transition rate model:
P(x,t) • 1-exp [-0.000115(1+0.312x)2][exp(0.057t)-l-0.057t]
SOURCE: Thyssen et al. 1981
111-22
-------
TABLE III-3
DATA USED TO ESTIMATE DOSE-RESPONSE RELATIONSHIP BETWEEN
INHALED B[a]P AND RESPIRATORY TRACT TUMORS
Exposure Average
Rate (x) Survival (t)
(mg B[a]P/m3)
Observed
Effective
Number
(in weeks)
Number of Respiratory
Tract Tumors
Exposed
Predicted
0
2.2
9.5
46.5
96.4
95.2
96.4
59.5
27
27
26
25
0.73
1.88
9.06
12.59
0
0
9
13
NOTE: Based on low-dose linear two-stage, dose- and
time-dependent/ identical transition rate model:
P(x,t) - 1-exp [-0.000115(1+0.312x)2][exp(0.057t)-l-0.057t]
N
SOURCE: Thyssen et al. 1981
IXI-22
-------
To extrapolate this value to human exposure for 24 hours
per day, the average length of exposure per day over the
approximate 2-year experimental period is calculated. The
weighted mean of the two exposure periods is
[(10)(4.5)+(92)(3)]/102 - 3.147 hr/day. Thus, exposure 24
hours per day has a potency value of
- (0.0170X24/3.147) - 0.1295 (mg/m3)"1,
This calculation is based on the assumption that inhalation
exposures are equivalent across species. To convert this
exposure level to (mg/kg/day)"1 to be able to compare it to
exposure via ingestion, the absorption rates for inhalation arj
ingestion are assumed to be the same and the average person is
assumed to weigh 70 kg and inhale 20 m air per day. Under
these assumptions, exposure to 1 mg/m for 24 hours results
in an exposure level of [(1)(20)]/70 - 0.2857 mg/kg/day. Thus,
on a mg/kg/day basis the estimated slope is
q1 - 0.1295/0.2857 - 0.4533 (mg/kg/day)~1,
This value is considerably smaller than the EPA's estimate of
6.11 (mg/kg/day) , which is based upon a species adjustment
for metabolism that corrects for the differences in metabolic
rates in an unclear manner.
The data of Thyssen et al. (1981) can also be fitted to
the Armitage-Do11 model (equation III-l), with the assumption
111-23
-------
that two stages are equally affected by the carcinogen out of a
total of 3 or 4 stages. Thus, both models can be fitted to the
data; the conceptual difference between the two models is that
the Moolgavkar-Knudson model is limited to two stages with
well-defined biological transitions between them, whereas the
Armitage-Doll model requires one or two additional stages and
does not specify the nature of the transitions.
7. Estimation of the Joint Carcinogenic Response to Total
PAH Exposure
In this section, a method for obtaining relative
carcinogenic potency estimates for PAHs is derived. This
method is dependent upon the structural form of the
dose-response model developed for B[a]P. Derivation of .
relative potency estimates requires that point estimates of
risk be obtained because upper bounds on risk cannot be
compared. For these reasons, point estimates of risk were
derived by fitting equation III-5 to dose-response data for
various PAHs. The point estimates are compared to the
corresponding estimates for B[a]P to obtain relative potency
estimates. These relative potency estimates serve as the
parameters Rj in equation 1-1 (presented in Section I).
The two-stage model was shown earlier in this section to
be consistent with the dose-response relationship between B[a]P
and mouse forestomach tumors following oral administration and
hamster lung tumors following inhalation exposure; Lee and
O'Neil (1971) have demonstrated a similar relationship for
mouse skin tumors following topical application. Given this
111-24
-------
demonstrated experimental consistency and the underlying
theoretical rationale for the mechanism of action of the PAHs
(see Section II), it is reasonable to assume that the same
functional relationship exists for all animal test models. The
parameter estimates would undoubtedly be species- and
tissue-dependent, but the underlying structural relationships
would very likely be the same.
Using the low-dose linear two-stage identical transition
rate model developed for B[a]P, the dose-response relationship
for a specified test system can be expressed as
P(x) - l-exp[-A(l+Sx)2]
for B[a]P and
- l-exp-A(l
for the j'th carcinogenic PAH.
It is possible to obtain 100% efficient estimates of the
RJ values by joint maximum likelihood estimation using data
from bioassays of individual PAHs conducted simultaneously.
However such estimation procedures are time-consuming to
develop and lie beyond the scope of the present project.
Fortunately, a simple approximation for the estimates of the
Rj values can Le obtained. To do so, the value of A is
estimated from the control data as previously discussed. It
can be shown that the final potency estimates are not very
111-25
-------
sensitive to this estimate. Given this estimated value of A,
the term SRj can be estimated by an approximate
goodness-of-fit procedure for each PAH for which data are
available. The parameter S is estimated in a comparable manner
using B[a]P dose-response data from the same bioassay. The
final relative potency estimates R., which according to our
general theoretical approach are independent of animal test
system used, are obtained from the ratio of the estimates for
SRj and S, determined from the jth PAH and B[a]P,
respectively.
For many test systems/ the highest dose may be the least
relevant for determining potency due to acute toxicity,
activation of cell proliferation mechanisms, and/or saturation
\
of metabolism or local penetration. When the highest dose
yields a level that is inconsistent with the assumed
dose-response model, it will not be used in estimating the
parameters. In other situations, only a single dose level for
B[a]P may be available. In this case, if B[a]P gives a high
response, a more stable estimate may be obtained by simply
using the PAH exposure level that gives the closest response
rate to that of B[aJP since such a procedure is independent of
any assumed parametric dose-response relationship. Even when a
control value and a single response for B[a]P and the jth
carcinogenic PAH are the only values available, it is possible
to obtain an estimate of the relative potency. To demonstrate
how such estimates can be obtained, consider the following
limited response data:
111-26
-------
OBSERVED RATES
Agent Exposure Response
Control 0 r0/n0
B[a]P x r/n
jth PAH yj rj/nj
Estimates for the relative potencies can be obtained by
equating the observed rate with the function response. This
results in the equations
ro/no " 1-exp-A
r/n - 1-exp-A(1+Sx)2
r./n^ » l-exp-A(l+SRjy.)
which can be solved algebraically by hierarchical substitution
of the parameter estimates for A and S into the equation
defining the relative potencies. This approach yields the
algebraic solution:
J
[ln(l-r/n)/ln(l-r0/n0)]l/2 _
In the next section, the relative potency values (R..)
are estimated for selected PAHs that are likely to be
encountered in the environment.
111-27
-------
IV. THE BASIS FOR AND DERIVATION OF COMPARATIVE
POTENCY ESTIMATES
This chapter describes nine experiments in which B[a]P and
other PAHs were tested concomitantly for carcinogenesis using
several animal species and different methods of administration/
as well as two experiments in which the levels of these
chemicals that interact with cellular DNA were determined.
Unlike the experiments using B[a]P described in Section III,
most of these studies used methods of administration that cannot
be directly compared to those by which humans would be expected
to be exposed. As a result/ dose reponse relationships for
these PAHs cannot be described mathematically in a manner that
is useful for the prediction of human risk unless they are
expressed relative to B[a]P. Cancer potencies relative to B[a]P
are derived for each PAH from each experiment in the following
section. The studies that are considered the best in terms of
performance and relevance to humans are then selected; the
relative potencies for the PAHs from these studies are
summarized at the end.
A. CARCINOGENESIS
It is difficult to compare directly the results of
carcinogenesis bioassays because of inter-laboratory
differences, varying susceptibilities and metabolic capabilities
between species used, and differences in techniques and exposure
routes. As a result, the only studies that are used to make
relative potency estimates meet the criteria that B[a]P was
IV-1
-------
tested in the same bioassay system as the other PAHs, in the
same laboratory, and at the same time. In this section,
individual studies that meet these criteria are summarized and
evaluated regarding quality/ and relative potency estimates are
derived for each of the PAHs for which adequate data are
*
available. In some instances, the studies were conducted with
several PAHs in addition to those commonly encountered at
hazardous waste sites. In these studies/ the results for the
i
other PAHs are included for comparison. For each study
discussed/ a table is included that summarizes all of the
results obtained; another table presents the data actually usrj
to derive the relative potency estimates/ for ease of
verification. The order in which the studies are discussed
corresponds to the reverse of the chronological order iri which
they were performed/ with the most recent study presented
first. Relative potency estimates are derived by the method
described in Section III.
1. Deutsch-Wenzel et al. (1983)
In this study, eight environmentally ubiquitous PAHs were
tested for carcinogenicity by direct implantation into the lungs
of female Osborne-Mendel rats. Test compounds were prepared in
solution with trioctanoin and molten beeswax. Upon injection
into the left lobe of the lung, the mixture congealed into a
pellet/ from which the test compound diffused over time into the
surrounding lung tissue. Epidermoid carcinomas and pleomorphiw
sarcomas were observed and dose-response relationships were
IV-2
-------
obtained. The treatment method involved some trauma to the
animals. Relative potency estimates would not be affected,
however, since this factor was constant for all treatment
groups. Tumor induction times were difficult to observe, so
survival time and the number of tumor-bearing animals were the
responses evaluated. Table IV-1 shows the results for the PAHs
tested.
The transition rate parameter R.S was estimated for each
of the carcinogenic PAHs from the data in Table IV-1. The
measure of response used for modeling was the number of animals
bearing epidermoid carcinomas of the lung at the completion of
the study. (Pleiomorphic sarcomas were excluded in the interest
of consistency and because their etiology probably differs from
that of carcinomas.) The numerical form of the best-fitting
model and the goodness of fit are shown in Table IV-2. No
tumors were observed in the control rats/ so the parameter A was
obtained by Bayesian estimation, as discussed in Section III
(i.e., A - -ln(l-.5/36) » 0.014). The highest exposure groups
were omitted for B[a]P and IND, since the response was
significantly less than that predicted by the model. This
discrepancy is .most likely due to toxicity and reduced survival
in those high exposure groups. A more complex analysis could
use median survival times as an additional variable; this would
most likely reduce the relative potency estimates somewhat and
increase precision.
Note that the low-dose linear two-stage model yields
potency estimates ranging from slightly more than one to three
IV-3
-------
TABLE IV-1
CARCINOGENICITY AND DOSE-RESPONSE RELATIONSHIPS OF EIGHT
FREQUENTLY OCCURRING ENVIRONMENTAL PAHs IN RAT LUNGS
Compound3
B[b]F
B[b]F
B[b]F
B[e]P
B[e]P
B[e]P
B[j]F
B[j]F
B[j]F
B[k]F
B[k]F
B[k]F
IND
IND
IND
ANT
ANT
B[ghi]P
B[ghi]P
B[ghi]P
B[a]P
B[a]P
B[a]P
BW-TC
Untreated
control
Dose
(mg)
0.1
0.3
1.0
0.2
1.0
5.0
0.2
1.0
5.0
0.16
0.83
4.15
0.16
0.83
4.15
0.16
0.83
0.16
0.83
4.15
0.1
0.3
1.0
—
«•—
Number
of
Animals
35
35
35
35
30
35
35
35
35
35
31
27
35
35
35
35
35
35
35
34
35
35
35
35
35
Median Sur-
vival Time/
Weeks (95%
Confidence
Interval)
110(95-118)
113(100-121)
112(98-126)
117(105-123)
111(91-126)
104(92-113)
110(91-118)
117(103-125)
89(75-105)
114(101-122)
95(81-104)
98(87-108)
116(95-123)
109(98-116)
92(82-108)
102(91-116)
88(72-103)
109(101-121)
114(103-123)
106(93-117)
111(95-120)
77(68-99)
54(46-64)
104(91-121)
118(104-133)
Number of
Animals
Bearing
Epidermoid
Carcinomas/
Number of
Animals
Bearing
Pleomorphic
Sarcomas
0/1
1/2
9/4
0/0
0/1
1/0
1/0
3/0
18/0
0/0
3/0
12/0
3/1
8/0
21/0
1/0
19/0
0/0
1/0
4/0
4/6
21/2
33/0
0/0
.
0/0
Tumor
Inciden
(%)
2.9
8.6
37.1
0.0
3.3
2.9
2.9
8.6
51.4
0.0
9.7
44.4
11.4
22.9
60.0
2.9
54.3
0.0
2.9
11.8
28.6
65.7
94.3
0.0
0.0
Abbreviations: B[b]F, benzo[b]fluoranthene; B[e]P,
benzo[e]pyrene; B[j]F, benzo[j]fluoranthene; B[k]F,
benzo[k]fluoranthene; IND, indeno(l,2,3-cd)pyrene; ANT,
anthanthrene; B[ghi]P, benzo[ghi]perylene; B[a]P,
benzo[a]pyrene; BW-TC, beeswax and trioctanoin (vehicle)
SOURCE: Deutsch-Wenzel et al. (1983).
IV-4
-------
<
Ul
3LE IV-2
DATA USED AND ESTIMATES OP RELATIVE POTENCY OBTAINED PROM
RAT LUNGS FOLLOWING INTRAPULMONARY INJECTION
Compound9
Exposure
Levels
in mg
(*)
B[a]P:
0.1
0.3
1.0
B[b]F:
0.1
0.3
1.0
B[k]F:
0.16
0.83
4.15
IND:
0.16
0.83
4.15
B[ghi]P:
0.16
0.83
4.15
Control
0.00
No. Rats
Exposed
35
35
35
35
35
35
35
31
27
35
35
35
35
35
34
35
Epidermoid
Observed
4
21
(b)
0
1
9
0
3
12
3
8
(a) '
0
1
4
0
Carcinomas
Predicted
4.79
19.82
(b)
0.83
1.73
7.40
0.75
2.09
13.79
1.66
11.41
(a)
0.57
0.95
4.07
^••™
Estimated
Transition
Rate
Parameter
(SRj)
22.28
3.12
1.48
5.17
0.486
-—
Estimated
Relative
Potency
(Rj)
1.
0.140
0.066
0.232
0.022
— —
Deutsch-Wenzel
Relative Pptency
Estimates Using
Probit Model
1.
0.11
0.03
0.08
0.01
— —
aAbbreviations: B[a]P, benzotajpyrene; B[b)P, benzo[b)fluoranthene; B[k]F,
benzo[k]£luoranthene; IND, indeno(l,2,3-cd)pyrene; B[ghi]P, benzo[ghi]perylene.
bOmitted because of lack of fit.
Model: P(x) « l-exp-A(l+SRj)2, A = -ln(l-.5/36) = 0.014.
Source: Deutsch-Wenzel et al. (1983).
-------
times greater than, although within the confidence limits of,
those obtained by the authors of the study based on the probit
model.
2. LaVoie et al. (1982)
The tumor-initiating activities of the benzofluoranthenes
were tested on Swiss albino CD-I female mouse skin and compared
with that of B[a]P. Compounds were applied in 10 doses (1 every
other day) to the backs of mice and followed 10 days later with
applications of the tumor promoter phorbol myristate acetate
(TPA) 3 times a week for 20 weeks. Table IV-3 shows the total
initiating dose of each of the compounds tested and numbers of
tumors and tumor-bearing animals obtained. Tumors were
\
principally squamous cell papillomas, with some keratoacanthomas
observed as well. Benzo[b]fluoranthene was approximately ten
times as potent as a tumor initiator than benzo[j]fluoranthene,
•
which was only somewhat more potent than benzo[k]fluoranthene.
B[a]P was more potent than all three.
Two factors make the data obtained in this experiment
difficult to interpret. The application of the promotor TPA
adds an additional element of complexity to the evaluation. Th,
possibility of a qualitatively different interaction between irA
and the various PAHs exists. Furthermore* this experimental
protocol may allow only the first and not the second mutation
that is hypothesized to occur during PAH-induced carcinogenesis
and would therefore not fit a two-stage model. These potential
problems are compounded by having only a single level of B[a]P
IV-6
-------
TABLE IV-3
TUMOR INITIATING ACTIVITY OF BENZOFLUORANTHENES
Number of Skin
Total Initi- Tumor-Bearing
ating Dose Animals/Number Skin Tumors
Compound3 (yg) of Animals per Animal
B[b]F
B[b]F
B[b]F
B[j]F
B[j]F
B[j]F
B[k]F
B[k]F
B[k]F
B[a]P
Acetone
(Vehicle)
100
30
10
1,000
100
30
1,000
100
30
30
••
16/20
12/20
9/20
19/20
11/20
6/20
15/20
5/20
1/20
17/20
0/20
7.1
2.3
0.9
7.2
1.9
0.6
2.8
0.4
0.1
4.9
0
*\
Abbreviations: B[b]F, benzo[b]fluoranthene; B[j]F,
benzo[j]fluoranthene; B[k]F, benzo[k]fluoranthene; B[a]P,
benzo[a]pyrene.
Source: LaVoie et al. (1982).
IV-7
-------
available upon which to base the potency estimate. To minimize
the latter problem, the data used in the evaluation will be
restricted to the single exposure level of each of the PAHs that
gives a tumor response closest to the 85% B[a]P tumor response
rate. This procedure yields results that are less
model-dependent. The relative potencies calculated on the basis
of these restricted data are displayed in Table IV-4.
3. Habs et al. fl980>
Six environmentally relevant PAHs were tested in order to
determine their dose-response relationships as carcinogens when
topically applied to the backs of female NMRI mice throughout
their lifetimes. Table IV-5 lists the compounds tested* doses
used/ frequency of application, and results obtained. Clear-cut
dose-response relationships were seen for B[a]P and benzo[b]-
fluoranthene. BenzoCjlfluoranthene proved to be weakly
carcinogenic, as did cyclopentadieno(c,d)pyrene and coronene.
Benzo[k]fluoranthene and indeno(l,2,3-cd)pyrene were inactive as
carcinogens in this test system.
The relative transition rate using the low-dose linear
two-stage model was calculated for each of the six test agents,
as shown in Table IV-6. The estimates obtained for B[j]F, CP,
B[k]F, and IND are based upon very weak or perhaps nonexistent
dose-response relationships. Even so, the response data
observed are consistent with the underlying response model.
Estimates obtained using these data would, if erroneous, tend to
overestimate the relative transition rates and, as a result, the
IV-8
-------
TABLE IV-4
DATA USED AND ESTIMATES OF RELATIVE POTENCY OBTAINED FROM
MOUSE SKIN INITIATION-PROMOTION EXPERIMENT
Compound8
B[a]P
B[b]F
B[j]F
B[k]F
Total
Dose
(M)
30
100
1,000
1,000
Number of Skin
Tumor-Bearing
Animals/
Number Exposed
17/20
16/20
19/20
15/20
Estimated
Transition
Rate
Parameter
(SRj)
0.263
0.0719
0.0102
0.0066
Estimated
Relative
Potency
(Rj)
1
0.273
0.039
0.025
Abbreviations: B[a]P, benzo[a]pyrene; B[b]F,
benzo[b]fluoranthene; B[j]F, benzo[j]fluoranthene; B[k]F,
benzo[k]fluoranthene.
Model: P(x) - l-exp-A(l •»• SRj)2, A - -ln[l-.5/21]
Source: LaVoie et al. (1982).
0.014.
IV-9
-------
TABLE IV-5
PAH-INDUCED SKIN TUMOR RATES IN NICE
<
Compound8
B[a)P
B[b]F
B[j)F
BtkjF
IND
CP
COR
^m ^^
—
Solvent
Acetone
Acetone
Acetone
Acetone
Acetone
Acetone
DMSO
Acetone
DMSO
Frequency of
Application
2z/week
2z/week
2z/week
2z/week
2z/week
2z/week
4z/week
2z/week
4z/week
Dose
(tig/animal)
1.7
2.8
4.6
3.4
5.6
9.2
3.4
5.6
9.2
3.4
5.6
9.2
3.4
5.6
9.2
1.7
6.8
27.2
5.0
15.0
—
~~
Number of Tumor-
Bearing Animals/
Total Number
of Animals
8/34
24/35
22/36
2/38
5/34
20/37
1/38
1/35
2/38
1/39
0/38
0/38
1/36
0/37
0/37
0/34
0/35
3/38
1/39
2/40
0/35
0/36
Tumor
Incidence
<*)
23.5
68.6
61.1
5.3
14.7
54.1
2.6
2.9
5.3
2.6
0
0
2.8
0
0
0
0
7.9
2.6
5.0
0
0
Abbreviations: B[a]P, benzofalpyrene; B[b]F, benzo[b]fluoranthene; B[j]F,
benzo[j]fluoranthene; B[k]F, benzo[k]£luoranthene; IND, indeno[l,2,3-cd]pyrene;
CP, cyclopentadieno(cd)pyrene; COR, coronene; DMSO, dimethylsulfoxide.
Source: Adapted fro abs et al. (1980).
-------
F.E iv-6
DATA USED AND ESTIMATES OP RELATIVE POTENCY OBTAINED PROM
MICE SKIN PAINTING STUDY
Compound8
Dose
ug/animal
(x)
B[a]P:
1.7
2.8
4.6
B[b]F:
3.4
5.6
9.2
B[j]F:
3.4
5.6
9.2
CP:
1.7
6.8
27.2
B[k]P:
3.4
5.6
9.2
IND:
3.4
5.6
9.2
No. of Tumor-Bearing Estimated
Animals Transition Estimated Habs et al. (1980)
Rate Relative* R«»1;tt-ivi» Pnt-onrv
No. Mice
Exposed
34
35
— —
38
34
__
38
35
38
34
35
38
39
38
38
36
37
37
Observed
8
24
(b)
2
5
(b)
1
1
2
0
0
3
1
0
0
1
0
0
Parameter Potency Estimates Using
Predicted (SRj) (Rj) Probit Model
3.92 1. 1.
10.36
20.60
(b)
0.656 0.167 0.27
2.38
4.30
(b)
0.241 0.061
0.77
1.18
2.36
0.091 0.023 0.05
0.28
0.56
2.74
0.078 0.020
0.43
0.59
0.86
0.081 0.021
0.36
0.48
0.69
aAbbreviations: B[a)P, benzo[a]pyrene; B[b]F, benzotbjf luoranthene;
benzo[j]f luoranthene; CP, cyclopentadieno(cd)pyrene; B[k]F, benzo[k]f luoranthene; IND,
indeno[l,2,3-cd]pyrene.
bOmitted because of lack of fit.
Model: P(x) - l-exp-A(l+SRj)2, A = -m(l-.5/81)
Source: Habs et al. (1980).
0.006192.
-------
relative potencies because the actual tumor response may be much
lower or non-existent. Note that the relative potency estimates
obtained using this model are about a factor of two smaller than
those obtained by the author based upon the log-probit model.
This difference is primarily due to the elimination of the
high-dose information for B[a]P that gave a response that was
lower than expected for no directly discernible reason.
4. Pfeiffer (19771
As part of an experiment designed to evaluate the
interactions between carcinogenic and noncarcinogenic PAHS,
single doses of B[a]P or dibenz[ah]anthracene dissolved in
tricaprylin were injected subcutaneously between the shoulder
\
blades of female NMRI mice. Animals were palpated weekly for
tumor development. Table IV-7 shows the incidence of sarcomas
obtained and the doses of PAHs used. The control tumor
incidence was not reported, although a group of
"non-carcinogenic" PAHs was tested and will be included here as
a pseudo-control group. Furthermore/ the numbers of effective
or surviving animals were not reported* so the tumor incidences
may not reflect potential toxicity at the higher doses.
The dose-response curves for B[a]P and D[ah]A increase at
rates that are less than linear, presumably due either to a
reduced rate of absorption or saturation of metabolic activating
systems at high doses. This effect is more pronounced for
D[ah]A than B[a]P, so that relative potencies for D[ah]A
estimated at high doses are lower than those estimated at lower
IV-12
-------
TABLE IV-7
TUMOR INCIDENCE FOLLOWING SUBCUTANEOUS INJECTION OF
BEN20[A]PYRENE OR DIBENZtAH]ANTHRACENE
Compound3
B[a]P
D[ah]A
Dose
3.12
6.25
12.5
25.0
50.0
100.0
2.35
4.7
9.3
18.7
37.5
75.0
Number of Tumor -
Bearing Animals
Number of Animals
9/100
35/100
51/100
57/100
77/100
83/100
37/100
39/100
44/100
56/100
65/100
69/100
Tumor
Incidence
(%)
9
.35
51
57
77
83
37
39
44
56
65
69
Pseudo-
control"
273.975
549
1,098.37
2,196.75
4,394
8,787
6/100
8/100
6/100
4/100
13/100
5/100
6
8
6
4
13
5
aB[a]P, benzola]pyrene; D[ah]A, dibenztah]anthracene.
b!0 "noncarcinogenic" PAHs. The "noncarcinogenic" PAH
group included chrysene, which is generally considered to be
weakly carcinogenic; at the dose levels used, however, chrysene
alone cannot account for the tumor response observed.
NOTE: Control animals were included in the experiment;
however, whether they were vehicle or untreated controls was not
specified, and their tumor incidence was not reported.
Source: Pfeiffer (1977).
IV-13
-------
doses. The lack of a true control population and the presence
of a tumor response from the population receiving a group of
hypothesized noncarcinogenic PAHs also make fitting the data to
the low-dose linear two-stage model difficult to interpret. As
a result, the relative potency was determined by relating
results at comparable observed response levels and using the
noncarcinogenic PAHs as the pseudo-vehicle control. The data
used for these comparisons are shown in Table IV-8. Two
estimates of the relative potency were obtained; however, the
estimate corresponding to the lower exposure levels (i.e., 2.81)
is probably more reasonable to use for low-dose extrapolations.
5. Binoham and Falk (1969)
\
Graded concentrations of B[a]P or benz[a]anthracene (B[a]A)
were applied topically to the backs of C3H/He mice 3 times a
week for 50 weeks and local tumors were quantitated. B[a]P was
dissolved in decalin, benz[a]anthracene was dissolved in toluei..
and 50 rag by weight of each solution was administered at each
dosing; however, the method of application was not specified.
Table IV-9 shows the results of this experiment. Mo solvent
controls were included. From these results, benz[a]anthracene
appears to be less potent than B[a]P, although the use of
different solvents for each agent introduces a confounding
factor that cannot be directly evaluated.
An extremely shallow dose-response relationship for B[alA
was obtained over a two order-of-raagnitude dose range. The most
likely explanation for this response is that the administered
IV-14
-------
TABLE IV-8
DATA USED AND ESTIMATES OF RELATIVE POTENCY OBTAINED FROM
SUBCUTANEOUS INJECTION STUDY WITH FEMALE MICE
Compound3
Low dose:
B[a]P
D[ah]A
High dose:
B[a]P
D[ah]A
Pseudo-
control
Dose
(w9)
(x)
6.25
2.35
25.00
18.70
0.00
Number
Exposed
•
100
100
100
100
600
Number
Mice with
Tumors
35
37
57
56
42
Estimated
Transition
Rate
Parameter
(SRj)
0.2297
0.6480
0.0964
0.1264
— —
Estimated
Relative
Potency
(Rj)
1
2.82
1
1.31
— —
Abbreviations: B[a]P, benzo[a]pyrene; D[ah]A/
dibenz [ah] anthracene.
Model: P(x) • l-exp-A(l+SRj)2, A « -ln[l-42/600]
Source: Pfeiffer et al. (1977).
0.0726.
IV-15
-------
TABLE IV-9
TUMOR INCIDENCE FOLLOWING DERMAL EXPOSURE OF MICE TO
BENZO[A]PYRENE AND BENZ[A]ANTHRACENE
Number of Animals
Bearing Malignant
Dose Tumors /Number of
Compound3
B[a]P
B[a]A
(% Concentration)
0.02
0.002
0.0002
0.00002
1.0
0.2
0.02
0.002
Animals
5/12
0/20
0/21
0/18
5/29
3/32
1/18
0/32
Tumor
Incidenc-
<%)
42
0
0
0
17
9
6
0
Abbreviations:
benz[ a] anthracene.
B[a]P, benzo[a]pyrene; B[a]A,
Source: Bingham and Falk (1969).
IV-16
-------
dose and the biologically active dose at the target site
(reactive form of the agent at the site of action) are not
'inearly related because of saturation of metabolic activating
systems or permeability. In this case, the responses obtained
»
at the lowest exposure levels would be the most directly
comparable between agents. The data used to obtain the relative
potency estimate for B[a]A are the single levels of agent giving
the lowest response for both B[a]P and B[a]A. The data used and
the relative potency estimate derived are shown in Table IV-10.
Incorporation of the other exposure levels of B[a]A into the
relative potency estimate would decrease the estimate for that
agent.
>
6. Van Dunren et al. f!966)
As part of an experiment to evaluate the tumorigenic
potential of tobacco leaf and smoke constituents, several PAHs
were tested for tumor initiating activity. The compounds were
applied as single doses in 0.1 ml of solvent to the backs of
female ICR/Ha Swiss mice. This treatment was followed by
applications of 25 pg croton resin in acetone as a promoter
three times a week for 63 weeks (or until death). The compounds
tested and the experimental results obtained are presented in
Table IV-11. B[a]P is clearly more potent than any of the other
PAHs tested.
The data used to obtain potency estimates for CH and B[a]F
are shown in Table IV-12. The response employed for the
estimates was number of malignant tumor-bearing animals by the
IV-17
-------
TABLE IV-10
DATA USED AND ESTIMATE OF RELATIVE POTENCY OBTAINED FOR
- BENZ(A)ANTHRACENE FROM MOUSE SKIN PAINTING STUDY
Compound3
Dose
(\ concen-
tration
<*>
B[a]P:
.02
B[a]A:
.02
.20
1.0
Control0:
0
No. Mice
Exposed
(Survived
50 weeks)
12
18
32
29
20
No. Mice
with
Malignant
Tumor
5
1
3
5
0
Estimated
Transition Rate
Parameter
(SRj)
187.0
27.16
5.13
1.81
— —
Estimated
Relati\ _
Potency
(Rj)
1.0
0.145b
0.027
0.0097
— —
a Abbreviations: B[a]P, benzo[a]pyrene; B[a]A,
benz [ a ] anthracene .
lvalue used for relative potency .estimate.
cAssumed value.
Model: P(x) « l-exp-A(l + SRj)2, A - -ln(l-0.5/21)
Source: Binghara and Falk (1969).
0.024.
IV-18
-------
TABLE IV-11
PAH-INDUCED SKIN TUMOR INITIATION IN MICE USING
CROTON RESIN AS A PROMOTER
Papillomas
Carcinomas
Number of
Tumor-Bearing
Animals/
Compound8
B[a]P
CH
B[b]P
B[mno]P
Control (pro-
motion with
acetone only)
Control
(untreated)
Dose
150 pg
1 mg
1 mg
1 mg
—
- —
Solvent
0.1 ml acetone
0.4 ml acetone
0.1 ml acetone
0.1 ml acetone
—
^^
Number of
Animals
20/20
16/20
18/20
4/20
5/20
0/100
Tumor
Incidence
(%)
100
80
90
20
25
0
Number of
Tumor-Bearing
Animals/
Number of
Animals
6/20
2/20
5/20
0/20
1/20
0/20
Tumor
Incidence
<*)
30
10
25
0
5
0
Abbreviations: B[a]P, benzo[a]pyrene; CH, chrysene; B[b]F, benzo[blfluoranthene;
B[mno]F, benzofranojfluoranthene.
Source: Van Duuren et al. (1966).
-------
TABLE IV-12
DATA USED AND ESTIMATES OP RELATIVE POTENCY OBTAINED FROM
MOUSE SKIN INITIATION PROMOTION EXPERIMENT
Compound3
BCaJP
CH
B[b]F
Promoter
control
Doseb
.
1.
0.040
0.125
— —
Abbreviations: B[a]P, benzol a] pyrene; CH,chrysene; B[b]F,
benzo[b]f luoranthene.
administration was followed by 25 pg croton resin
applied three times weekly for 63 weeks. N
Model: P(x) - l-exp-A(l+SRj)2, A -
Source: Van Duuren et al. (1966).
-ln(1-1/20) - 0.0513.
IV-20
-------
63rd week of the test. The necessity of using a promoter to
obtain a carcinogenic response in the test system is a factor
that would tend to reduce the credibility of the potency
estimates if no other information was available, although since
internal comparisons are available for the determination of
consistency/ using the initation-promotion system can supply
valuable additional information. Use of the two-stage model to
evaluate an initiation/promotion experiment is questionable,
however. The potency estimate for CH in this system is highly
variable/ since the response level is not clearly different from
the croton resin control.
7. Hoffmann and Wvnder (1966)
Two different types of experiment were performed:
(1) 'Complete" carcinogenesis experiments/ in which PAHs
were topically administered for 1 year and
(2) Initiation-promotion experiments/ in which PAHs were
administered for only 3 weeks/ followed by application
of a tumor-promoting agent for the duration of the
experiment.
For the "complete" carcinogenicity evaluations/ female
Ha/ICR/rail Swiss albino mice received three weekly topical
applications of B[a]P, benzolghi]perylene, or
indeno(l/2/3-cd)pyrene for one year. The number of papillomas
observed tt each dose for each group is shown in Table IV-13.
B[a]P and benzp[ghi]perylene were dissolved in dioxane, and
indeno(l/2/3-cd)pyrene was dissolved in acetone. No tumors were
observed in the dioxane vehicle controls; no acetone control was
included. The results
IV-21
-------
TABLE IV-13
CARCINOGENIC ACTIVITY OF SEVERAL PAHS ON MOUSE SKIN
Compound3
B[a]P
IND
B[ghi]P
Dose
(% Concen-
tration)
0.05
0.1
0.01
0.05
0.1
0.5
0.05
0.1
Number of Tumor -
Bearing Animals/
Number of Animals
17/20
19/20
0/20
0/20
6/20
7/20
1/20
0/20
Tumor
Incidence
(%)
85
95
0
0
30
35
5
0
Abbreviations: B[a]P, benzola]pyrene; IND, indeno(l,2,3-
cd)pyrene; B[ghi]P, benzo[ghi]perylene. N
Source: Hoffmann and Wynder (1966).
IV-22
-------
show that B[a]P is much more potent as a carcinogen than either
of the other PAHs tested in this model.
For the initiation-promotion experiments, ten doses of the
same PAHs tested above were applied in dioxane 2 days apart to
the backs of mice for a total dose of 0.25 mg per mouse and
followed by 2.5% croton oil in acetone. The frequency and
duration of croton oil administration were unclear. The results
of these experiments are found in Table IV-14. Again, B[a]P was
shown to be much more potent than either of the other PAHs
tested.
For each of the compounds used in the first experiment, the
lowest dose that gave a positive response was employed in the
relative potency estimation. Again, since B[a]P gave a Crouch
higher response than the other PAHs and the slope appears to be
very low, the potency estimate, if in error, is most likely
overestimated. These estimates are shown in Table IV-15.
The initiation-promotion experiment yielded results that
would tend to overestimate risk for two reasons. First, the
responses for IND and B(ghi]P may be spurious, since they are
not statistically elevated over the control response; second,
the response for B[a]P is in a high range in which a lower
percent of the applied dose may be converted to a biologically
active target dose, as compared with lower applied dose levels.
The relative potency estimates obtained from this experiment are
displayed in Table IV-16, although they are probably less
reliable than those obtained in the 'complete* carcinogenses
IV-23
-------
TABLE IV-14
INITIATION-PROMOTION ACTIVITY OF SEVERAL PAHS ON MOUSE SKIN
Compound3
B[a]P
IND
B[ghi]P
Control6
Total
Doseb
(mg)
0.25
0.25
0.25
0
Number of Tumor -
Bearing Animals/
Number of Animals
24/30
5/30
2/27
2/30
Tumor
Incidence
(%>
80
17
7
7
Abbreviations: B[a]P, benzo[a]pyrene; IND, indeno(1,2,3-
cd)pyrene; B[ghi]P, benzo[ghi]perylene.
^Administration of PAHs was followed by 2.5% croton oil as
a promoter.
cCroton oil was administered in the absence of PAH
treatment. N
Source: Hoffmann and Wynder (1966).
IV-24
-------
TABLE IV-15
DATA USED AND ESTIMATES OF RELATIVE POTENCY OBTAINED FROM
MOUSE SKIN COMPLETE CARCINOGENESIS EXPERIMENT
Compound9
B[a]P
IND
B[ghi]Pb
Control
Dose
(\ Concen-
tration)
(x)
0.05
0.10
0.05
0
No. Of
Animals
Exposed
20
20
20
20
NO. Of
Tumor -
Bearing
Animals
17
6
1
0
Estimated
Transition
Rate
Parameter
(SRj)
157.8
28.55
9.238
— *
Estimated
Relative
Potency
(Rj>
1.0
0.181
0.0585
•* —
Abbreviations: B[a]P, benzo[a]pyrene; IND, indeno(l,2,3-
cd)pyrene; B[ghi]P, benzolghi]perylene.
^Response was not dose related.
Model: P(z) - l-exp-A(l+SRj)2, A • -ln(l-.5/21) - .0241.
Source: Hoffman and Wynder (1966).
IV-25
-------
TABLE IV-16
DATA USED AND ESTIMATES OF RELATIVE POTENCY OBTAINED FROM
MOUSE SKIN INITIATION-PROMOTION EXPERIMENT
Compound3
B[a]P
INDC
B[ghi]pc
Control
Doseb
-------
experiment because of the inconsistency of the initiation-
promotion protocol and the two-stage model.
8. Wynder and Hoffmann (1959)
As part of a study of the carcinogenicity of tobacco and
its constituents, several PAHs were tested as complete
carcinogens on the skin of mice. Female Swiss mice received
various concentrations of the test substances dissolved in
acetone three times a week applied to their backs with a brush
throughout their lifetimes. Table IV-17 shows the compounds
tested, concentrations used, and the results of these
experiments. No solvent control group was reported; however,
since no papillomas or carcinomas were obtained for several of
the PAHs tested, a solvent control group would most likely have
been negative as well. The results indicate that B[a]P and
dibenz[ah]anthracene are somewhat similar in potency in this
model and chrysene is much less potent.
In this experiment, the survival rate was much less for
B[a]P-treated animals than for those receiving the other PAHs.
To adjust for the difference in survival rates, the relative
potency estimates will be based on comparisons of tumor rates
observed at 14 months using number of animals bearing carcinomas
as the response. The data used and the relative potency
estimates generated for D[ah]A and CH are shown in Table IV-18.
Estimates obtained from this experiment are considered to be of
high quality because of two factors: First, the response spans
the full range from zero to almost 100% for both
IV-27
-------
TABLE IV-18
DATA USED AND RELATIVE POTENCIES OBTAINED FROM MOUSE SKIN
CARCINOGENESIS EXPERIMENT
Compound3
Dose
(\ con-
centra-
tion) (X)
B[a]P:
.001
.005
.01
D[ah]A:
.001
.01
.1
CH:
1.
Control:
0
No.
Animals
Exposed
29
30
20
20
20
(b)
20
20
No. Animals
Carcinomas
Months
with
by 14
Observed Predicted
0
19
19
2
18
(bj
8
0
2.23
14.01
17.42
1.69
18.31
(bj
8.00
0.48
Estimated
Transition
Rate
Parameter
(SRj)
822.0
912.0
3.60
~—
Estimated
Relative
Potency
(Rj)
1.0
1.11
, 0.0044
~—
Abbreviations: B[a]P, benzo[a]pyrene; D[ah]A,
dibenz[ah]anthracene; CH, chrysene.
bOmitted because of lack of fit.
Model: P(x) - l-exp-A(l+SRj)2, A - -ln(l-.5/21)
Source: Wynder and Hoffman (1959).
0.0241,
IV-29
-------
B[a]P and D[ah]A; second, the model is consistent with the data
for the full range of B[a]P exposures.
9. Bryan and Shimkin (1943)
A study was conducted to establish dose-response curves for
several PAHs following subcutaneous injection into the right
axilla of male C3H mice. The solvent used was tricaprylin,
although no solvent control group was reported. Animals were
palpated frequently following a single injection of the test
compound in order to detect tumors; 99\ of those detected proved
to be spindle-cell sarcomas (the remaining 1% were carcinomas or
mixed). Table IV-19 shows the results obtained for B[a]P and
dibenz[ah]anthracene. Dibenz[ah]anthracene appears to be
somewhat more potent than B[a]P in this system.
Although this classic experiment with 12 exposure levels
per compound gave a full range of response for each compound,
several factors make it difficult to interpret. An exposure
level of 0.0039 mg appears to have been included for both B[a]P
and D[ah]A; however, no results were reported for this dose, and
no explanation for the omission was included. Vehicle control
response levels were also not reported. This problem is
exacerbated by the abnormality of a positive response at the
lowest exposure level for B[a]P (0.00195 mg) in the absence of a
response at exposure levels A, 8, or 16 times higher. For the
analysis conducted here, the data will be restricted to the
monotonically increasing portion of the curve, which is similar
to the approach taken by the authors in their fitting of the
IV-30
-------
TABLE IV-19
TUMOR INCIDENCE OBSERVED IN MICE FOLLOWING SUBCUTANEOUS
INJECTION OF BENZO[A]PYRENE OR DIBENZ[AH]ANTHRACENE
Compound3
D[ah]A
B[a]P
Dose
(mg)
8.0°
4.0
2.0
1.0
0.5
0.25
0.125
0.062
0.031
0.0156
0.0078
0.00195
8.0
4.0
2.0
1.0
0.5
0.25
0.125
0.062
0.031
0.0156
0.0078
0.00195
NuT.be r of Tumor -
Bearing Animals/
Number of Animals
16/21
17/20
19/19
22/22
20/21
19/21
21/23
20/20
16/21
6/19
6/40
2/79
20/21
16/19
19/19
18/20
19/19
14/21
15/19
4/20
0/16
0/19
0/40
2/81
Tumor
Incidence
(*>
76
85
100
100
95
90
91
100
76
32
15
3
95
84
100
90
100
67
79
20
0
0
0
2
aD[ah]A, dibenz[ah]anthracene; B[a]P, benzo[a]pyrene.
^Volume of solvent was Ot5 ml for the 8.0 mg dose of D[ah]A
and 0.25 ml for all other doses.
Source: Bryan and Shimkin (1943).
IV-31
-------
probit model. The only difference is that the authors of the
study included the 0.125 mg level for D[ah]A in their analysis,
ich would have the effect of causing a slight decrease in the
relative potency estimates and yielding a worse fit. In
addition, for the purpose of the present analysis, the
assumption will be made that the lowest dose of B[a]P (0.00195
mg) actually represents a control level. This assumption is
consistent with the ol^erved response data. The dose-response
data used in the potency determination are shown in Table
IV-20. Under the conditions of this experiment; the potency of
D[ah]A is estimated to be 4.5 times that of B[a]P.
B. SURROGATE APPROACHES: DNA ADDUCT FORMATION
Interaction between reactive PAH metabolites and DNA has
been established as a necessary (albeit not sufficient) event
for PAH-induced tumorigenesis (Gelboin 1980, Weinstein et al.
1978). The formation of B[a]P diol-epoxide DNA adducts has been
shown to be linearly related to administered dose for several
organs in rats and mice (Pereira et al. 1979, Adrienssens et al.
1983), and related to a quad.ratic increase in tumor formation
(Lee and O'Neill 1971). It is reasonable to suggest that in tl._
absence of quantitative tumor dose-response information for
carcinogenic PAHs that can be readily compared to B[a]P, the
formation of DNA adducts could be used as a surrogate basis for
establishing relative potency estimates.
It is recognized that the DNA adduct level in a tissue is
not directly related to carcinogenic response. Other factors
IV-32
-------
TABLE IV-20
DATA USED TO OBTAIN A RELATIVE POTENCY ESTIMATE FOR
DIBENZ[AH]ANTHRACENE FROM A
SUBCUTANEOUS INJECTION EXPERIMENT
Tumors
Compound3
Dose
(mg)
B[a]P:
0.125
0.062
0.031
D[ah]A:
0.062
0.031
0.0156
0.0078
Pseudo-
controlb
0.00195
No. of
Mice
Exposed
19
20
16
20
21
19
40
•
•
160
Observed
15
4
0
20
. 16
6
6
4
Expected
12.15
5.73
2.05.
19.71
14.84
6.33
5.85
4.0
Estimated
Transition Estimated
Rate Relative
Parameter Potency
(SRj) (Rj)
42.8 1.0
192.4 4.50
\
_— __
Abbreviations: B[a]P, benzo[a]pyrene; D[ah]A,
dibenztah]anthracene.
^Combined response at a dose level of 0.00195 mg of B[a]P
or D[ah]A.
Model: P(x) - l-exp-A(l+SRj)2, A - -ln(l-4/160) - 0.0253.
Source: Bryan and Shimkin (1943).
IV-33
-------
such as the DNA repair and cell turnover rates in the target
tissue can have a greater effect than the actual number of
adducts on the ultimate probability of tumor formation.
However, under the assumptions of the current model of PAH tumor
induction, the transition rate is linearly related to the DNA
adduct formation rate. As a result, the relative potency
measured as the ratio of DNA adducts per unit administered dose
for two PAHs is assumed to be independent of tissue type, route
of exposure, and test system as long as the ratios of the levels
of distribution, metabolism, and elimination for B[a]P and the
PAH in question remain unchanged. Furthermore, DNA adduct
formation may reflect cancer risk from PAHs at low,
environmental levels of exposure more accurately than
dose-response curves based on tumor rates produced at high
levels of exposure in the laboratory, when additional biological
mechanisms may play a role in enhancing or suppressing tumor
rates.
This section describes the studies in which the levels of
DNA adduct formation were determined for several PAHs, including
B[a]P. The studies are described in reverse chronological
order, and relative potencies are determined for the PAHs tested.
1. Reddv et al. (1984)
A total of 28 compounds were tested for ability to bind to
DNA using a very sensitive technique (detection limit^rof
a
1 aromatic DNA adduct in 10 normal adducts) for the detection
of DNA adducts employing P-postlabelling following
IV-34
-------
administration to rats and mice. The PAHs tested were applied
to the dorsal skin of female BALB/c mice as four doses of 1.2
umole each in acetone at 0, 6, 30, and 54 hours. Animals were
sacrificed 24 hours after the last treatment. DNA was isolated
from treated skin, digested, enzymatically labelled with P
(using T4 polynucleotide kinase and [y- P]ATP), resolved
using thin-layer chromatography and detected using
autoradiography and scintillation counting. Relative adduct
labelling was determined by comparing the radioactivity of the
adducts to that of normal nucleotides. Absolute levels of
adduct formation were not given, and results were expressed as
the ranges depicted in Table IV-21; as a result, point estimates
of relative potency are not obtainable. The only quantitative
conclusion that can be drawn from this study is that B[a]P is
the same potency or up to three orders of magnitude more potent
than the other PAHs tested. The authors noted, however, that
there was a good correlation between adduct level and
carcinogenic potency on mouse skin, and that the noncarcinogens
failed to form detectable adducts.
2. Phillips et al. (19791
The levels of binding of seven radioactively-labelled PAHs
to mouse skin DNA following topical application were
determined. One umole of each PAH in acetone (see Table IV-22
for compounds tested) was applied to the backs of male C57BL
mice and the animals were sacrificed at the time of maximal DNA
binding (from 19 to 72 hours following administration, depending
IV-35
-------
TABLE IV-21
RELATIVE DNA BINDING OF PAHs TO MOUSE SKIN
Relative
Compound3 Level of Binding*3
B[a]P +++
B[a]A ++
D[ah]A ++
B[ghi]P *+
CH ++
AN ND
PY ND
Abbreviations: B[a]P, benzo[a]pyrene; B[a]A,
benz[a]anthracene; D[ah]A, dibenz[ah]anthracene; B[ghi]P,
benzo[ghi]perylene; CH, chrysene; AN, anthracene; PY, pyrene
^Relative binding levels: +++, 1 adduct in 104 - 5zl05
nucleotides; -n-, 1 adduct in 5x10^ - 107 nucleotides.
ND - not detected; detection level - 1 adduct in 10s
nucleotides.
Source: Reddy et al. (1984).
IV-36
-------
TABLE IV-22
EXTENT OF PAH BINDING TO DNA
Estimated Relative Potency
pmoles DNA-PAH/ma DNA
Compound3
B[a]A
D[ac]A
D[ah]A
7 -MB A
3-MC
B[aJP
7,12-DMBA
Extent of Reaction
with DNA
(pmole/mg DNA)
2
10
15
25
25
27
43
pmoles DNA-B[a]P/mg DNA
0.07
0.37
0.56
0.93
0.93
1.0
1.6
Abbreviations: B[a]A, benz[a]anthracene; D[ac]A,
dibenz[ac]anthracene; D[ah]A, dibenz[ah]anthracene; 7-MBA,
7-methylbenz[a]anthracene; 3-MC, 3-methylcholanthrene; B[a]P,
benzo[a]pyrene; 7,12-DMBA, 7,12-dimethylbenz[a]anthracene.
Source: Phillips et al. (1979).
IV-37
-------
on the compound, which for the purpose of the model is assumed
to be proportional to steady-state binding levels under
conditions of continuous exposure). DNA was isolated,
hydrolyzed enzymatically and chromatographed; levels of PAH
binding were determined from the amount of radioactivity eluting
with the fractions containing DNA-bound PAHs. The extent of DNA
binding for the PAHs tested is shown in Table IV-22.
7,12-Dimethylbenz[a]anthracene, considered to be the most
carcinogenic PAH, was found to bind to DNA to the greatest
extent; the weakest carcinogens, benz[a]anthracene and
dibenz[a,c]anthracene, were bound the least. The extent of
binding of the intermediate compounds also appeared to correlate
well with what are thought to be their carcinogenic potencies.
B[a]P is approximately two orders of magnitude and twice as
potent, respectively, as the other PAHs tested,
benz[a]anthracene and dibenz[ah]anthracene.
;
3. Grover and Sims (1968)
The extent to which PAHs react with DNA was determined
using an in vitro system comprised of a microsomal enzyme
activation system/ salmon sperm DNA, and 0.5 ug of
radioactively labeled PAHs. Following incubation for one hour,
the amount of PAH bound to DNA was determined using
scintillation counting. The PAHs tested and results obtained
are shown in Table IV-23. B[a]P was again shown to be the most
active PAH; it was at least twice as potent as the other
indicator PAHs.
IV-38
-------
TABLE IV-23
BINDING OF PAHs TO DNA IN VITRO
Estimated Relative Potency
umoles DNA-PAH/mole DNA
Compound3
B[a]P
D[ah]A
B[a]A
PY
PH
Extent of Reaction
with DNA
(umoles/mole DNA)
1.41
0.44
0.70
0.31
0.05
umoles DNA-B[a]P/mole DNA
(Rj)
1.0
0.31
0.50
0.22
0.04
Abbreviations: B[a]P, benzo[a]pyrene; D[ah]A,
dibenz[ah]anthracene; B[a]A, benz[a]anthracene; PY, pyrene; PH,
phenanthrene.
Source: Grover and Sims (1968).
IV-39
-------
Establishing a credible biological basis for use of an in
vitro system on which to base relative potency estimates is
difficult, since many of the factors affecting adduct formation
in vivo, such as DNA repair, are absent; however, the
information derived from this system may be used in support of
other data.
C. SELECTION OF RELATIVE POTENCY ESTIMATES
In the previous sections, relative potency estimates were
derived for carcinogenic PAHs. These estimates were based on
the results obtained in ten separate studies using five
different experimental systems. The potency estimates and the
systems from which they were derived are summarized in Table
IV-24. Note that up to four independent estimates of relative
potency were obtained for the PAHs.
A composite estimate of the relative potency is required
for each PAH. Major qualitative differences exist between the
studies from which the estimates are derived concerning their
precision and applicability to continuous human ingestion or
inhalation exposure. As a result, no attempt will be made to
obtain a numerical estimate based upon a weighted average of tl._
studies. Instead, the most relevant study will be selected
based upon the following objective criteria and any particular
unique condition that is noted. The selection is provisional; a
more appropriate method of selection might entail a consensus of
opinion from a panel of recognized experts employing a formal
system of weighting such as that suggested by DuMouchel and
Harris (1983).
IV-40
-------
TABLE IV-24
SUMHA8Y OF HELAT1VE POTEMCY ESTIKATES
f(X IMOICATOB. PAMt
TEST SYSTEM
Subcutaneous Intrapulmonary Initiation-
Nous* Skin Injection Adminittration Promotion DMA Adduct
Carctnogeneals Into Mice to Rat»2 on Moust Skin Formation'
•emo CaJ pyrene
lenz Ca] anthracene
lenzoCbl f luoranthen*
BenzoCklf luoranthene
•enzo tgh < ] pery I ene
Chrysene
0 < ben i [ah] anthracene
I ndeno 1 1 , 2 , 3 - cdl pyrene
1.0 1.0
O.H5°
0.167**
0.020b
0.0585e
o.ooud
1.11- I.K«.4.SO'
0.021b,0.181e
1.0 1.0 1.0
0.07
O.UO 0.273h.0.125'
0.066 0.025h
0.022 O/OUT*
O.CKO1
O.S6
0.232 0.163C
Reference*: (a) llnohan and Falk 1969
(b) Habe et al. 1980
(c) HoffMm and Wynder 1966
(d) Wynder and HoffMnn 1959
(e> Pfelffer 1977
(f) Iryan and Shlakin 1943
(0) Oeutach-Wentel et al. 1983
(h) LaVote tt al. 1980
(I) Van Ouuren et al. 1966
(j) Phillip* et al. 1979
VI-41
-------
Some of the criteria considered for making the .selection
are:
(1) The relevance of the route of exposure employed in the
bioassay to anticipated human exposure.
(2) The duration of exposure (longer exposures are
preferred).
(3) The sample sizes used.
(4) The inclusion of a vehicle control group.
(5) A dose-response relationship obtained that is
consistent with the underlying theoretical model, which
is linear-quadratic in form.
(6) The extent to which the observed responses cover the
entire response range.
(7) The absence of additional complicating variables such
as promoters.
The extent to which each of the studies fulfills these
N
criteria, in a relative sense, is indicated in Table IV-25. The
scale used is as follows: «•++, above average; ++, average; +,
below average.
Based on this information and some factors specific to
individual studies, the relative potency estimates that are
considered to have the most reliable basis were selected and are
shown in Table IV-26.
The basis for each of the selections in Table IV-26 is as
follows:
BenzTalanthracene. Only one study was performed from which
a relative potency could be derived.
BenzoTblfluoranthene and BenzoTklfluoranthene. Both the
mouse skin carcinogenesis (Habs et al. 1980) and intrapulmonary
administration (Deutsch-Wenzel et al. 1983) studies were found
IV-42
-------
TABLE IV-25
EXTENT TO WHICH STUDIES USED TO DERIVE RELATIVE POTENCY
ESTIMATES FULFILL APPLICABILITY CRITERIA
CRITERIA1
Reference
Teat tys t «i
Routi of
Exposure
Contlnu-
Exposure
Sample
Sizt
Fits Dose-
Absence Response
Quality of of Com- Full Model for
Controls ptletting Range of 2 or More
Used Factors Responses Doses
Oeutsch- Intrapulaonery
Wtnzel administration
•t al. 1963
LaVoie Initiation-
et al. 1980 promotion
on skin
Mate tt al.
1980
Skin ear-
cinogenesis
Pfeiffer
1977
Subcut
Injection
Knehan and
Falk 1969
Skin car-
einogwwsU
Van Duuran
tt al. 1966
Initiation-
promotion en
skin
Hoffman and
Wynder 1966
Skin car-
cinogenasit
or
Initiation-
promotion
on (kin
nd Skin car-
Moffnn 1959 elnP9tnM<»
•rytn and Subcutamoua
Shinkln 1943 Injection
"Criteria art diacuastd sort fully In the ttxt.
IV-43
-------
TABLE IV-26
SUMMARY OF RELATIVE POTENCY ESTIMATES
DERIVED FOR PAHS
Benzo[a]pyrene 1.0
Benz[a]anthracene 0.145b
Benzotb]fluoranthene 0.1403
Benzo[k]fluoranthene 0.066a
Benzo[ghi]perylene 0.0223
Chrysene 0.0044C
Dibenz[ah]anthracene l.llc
Indeno[1,2,3-cd]pyrene 0.232a
aDeutsch-Wenzel et al. (1983)
bBingham and Falk (1969).
cWynder and Hoffmann (1959).
IV-44
-------
to be of comparable quality; the estimates based on the latter
study were chosen because of the relevance of this route of
exposure to that of the B[a]P study on which the dose-response
model is based. This choice results in an approximately 15%
lower estimate for B[b]F, but in a 3-fold greater estimate for
B[k]F.
BenzoTahi1oervlene. IARC (1983) considers this compound to
be a noncarcinogen. However, evidence from several studies
indicates that it has weak carcinogenic potential. Therefore,
for the purpose of conservatism, B(ghi]P will be considered to
be carcinogenic and a relative potency estimate derived. The
intrapulmonary administration study (Deutsch-Wenzel et al. 1983)
\
proved to fulfill many more of the adequacy criteria than did
the skin carcinogenesis study (Hoffman and Wynder 1966),
therefore, the former was used as the basis for estimation.
Chrysene. The skin carcinogenesis study (Wynder and
Hoffman 1959) was considered to be more appropriate than the
initiation-promotion study (Van Duuren et al. 1966) because of
the additional variables introduced in the initiation-promotion
study. Furthermore, the tumor response observed in the latter
study was not statistically significant.
Dibenzrahlanthracene. On the basis of the extent to which
each of the studies of this compound fulfill the criteria for
study adequacy, the estimate derived from the skin
carcinogenesis study (Wynder and Hoffman 1959) will be used.
The relative potency estimate based on DNA adduct formation
supports the choice of the lower estimate. Estimates of two to
IV-45
-------
four times higher were obtained from studies that employed
subcutaneous injection as a method of administration; however
this,system is less relevant to human exposure.
Indeno Fl.2,3-cdlpyrene. As was true for the
benzofluoranthenes, both the Habs et al. (1980) skin
carcinogenesis study and the Deutsch-Wenzel et al. (1983)
intrapulmonary administration study fulfill most of the adequacy
criteria. The estimate based on the latter study was chosen
because of its relevance to the route of exposure used in the
study of B[a]P from which the dose-response model was derived.
Four independent potency estimates were made and the largest
chosen. Consistency between estimates was apparent; the three
largest estimates varied by no more than a factor of one-third.
IV-46
-------
V. ESTIMATION OF CANCER RISK DUE TO MIXTURES OF POLYCYCLIC
AROMATIC HYDROCARBONS USING ALTERNATIVE METHODS
A number of mathematical models are available that can be
used to describe cancer dose-response relationships for PAHs
instead of the biologically-based two-stage model described in
r
Section III. These models have several limitations; however,
it is useful to see how relative potencies can change depending
on the model used and the assumptions made. For the purpose of
numerical comparison, this section applies two alternative
models (one-hit and multistage) to the tumor data available for
PAHs and compares the results to the relative potencies derived
using the approach described in this report.
N
1. One-Hit Model
The one-hit model assumes that the dose-response
relationship between PAH exposure and tumor response is
one-stage (instead of two-stage or multistage) in form. Under
this assumption, only one change is required to turn a normal
cell into a tumor cell and the probability of a tumor response
may be expressed as
P(x) - 1 - exp -(BQ*B1x), (V-l)
where x is the administered dose, BQ is the background tumor
rate/ and B. is the low-dose linear term (i.e., slope of the
V-l
-------
line described by the model at low doses). The one-hit model
is a special case of the Armitage-Doll and Guess-Crump models
in which time is fixed and only one stage is assumed to be
affected.
Given a single bioassay in which a separate tumor dose
response relationship is obtained for m different PAHs
including B[a]P, the low dose linear term for each PAH (i.e.,
Bli where i-l,2,...,m) can be estimated. The ratio of the
linear terms for a specified PAH to the linear term for B[a]P
is an estimate of the relative potency of that PAH compared to
B[a]P.
The one-hit approach has several advantages. Only one
data point is required to obtain a relative potency estimate.
The estimates are very stable (i.e., they do not change much as
a result of small alterations in the tumor data). The
estimates are dose-independent, so they may be used in all dose
ranges. The main disadvantages are that many of the PAH
dose-response data observed are nonlinear so that systematic
biases.are introduced by assuming linearity. Also, in the
simple form presented in equation V-l, the model can be applied
only to studies in which the dose z and time t are constant.
2. Linearized Multistage Model
Point estimates instead of statistical upperbounds for
multistage model parameters can be obtained and used to predict
tumor rates at various PAH exposure levels. One measure of
potency of a particular PAH is the reciprocal of its exposure
level that results in a specified tumor response rate. For
V-2
-------
example, the reciprocal of the exposure level that induced a
10% tumor response based on point estimates from the multistage
model is used by EPA as a measure of potency for reportable
quantities of potential carcinogens (see Cogliano 1986). The
ratio of two potencies so calculated might be used to obtain a
relative potency estimate. Relative potencies calculated in
this manner may vary at different response levels and as a
result, are unreliable because they are highly dependent on the
tumor response level chosen. For response levels that are much
smaller than those in the observable range (i.e., increases in
tumor rates expected at environmental levels of exposure), a
very minor shift in the tumor response data in an experiment
can be shown to lead to orders of magnitude changes in the
relative potency estimates at environmental levels using this
approach. As a result, this approach has little to recommend
it for assessing low dose potencies over a wide range of
potential environmental exposure levels.
Using the 95\ upper bound linear term obtained from the
multistage model to derive relative potencies instead of point
estimates has a superficial level of appeal. This is the
measure of potency roost often used by EPA in various regulatory
settings (see EPA 1986c). The ratios of upper bound estimates
are not good measures of relative potency, however, since they
are sometimes more dependent upon sample size considerations
than on tumor response levels. Since unit risks (i.e., 95\
upper bound linear terms) are often used to represent the
potencies of carcinogens at environmental levels of exposure,
V-3
-------
using them to estimate relative potencies is an informative
exercise nonetheless.
3. Application of Models
Using the tumor data from Sections III and IV, the
parameters in the linearized multistage and one-hit models were
estimated and a Chi-square goodness-of-fit statistic obtained.
The parameter estimates and X values are shown in Appendix
A. In cases where the models did not fit the data, as measured
by the approximate X goodness-of-fit test, the tumor rate
obtained at the highest dose was omitted and the models were
fitted to the reduced data set. Comparisons of the observed
and predicted numbers of tumors for each agent in each s
experiment are shown in Appendix B.
Comparison of the results in Appendices A and B with the
corresponding results for the two-stage model presented in
Sections III and IV reveals a number of similarities and
differences. The one-hit model often yields a poor fit to the
data, reflecting the fact that dose-response relationships
showed upward curvature in most studies. The multistage model
provided adequate fits to the data in all but four cases. In
all four of these cases, the two-stage model also fitted poorly
to the full data sets and calculations of relative potency wei_
based on selected subsets of the data (Tables IV-4, IV-8,
IV-14, and IV-20); in at least of two of these cases, the
multistage model gave adequate fits to similar subsets. In a
number of cases, the multistage model yielded a
V-4
-------
linear-quadratic dose-response relationship similar to that
assumed in the two-stage model. In other cases, however, the
multistage model included only a linear term or the last term
in the nth degree polynomial. In several cases, the multistage
model yielded slightly better fits to the data than the
two-stage model (e.g.. Tables B-l, B-3, B-7, and B-9).
However/ these "improvements" in fit resulted from the larger
numbers of parameters in the multistage model with resulting
variability in the number of terms (i.e., in the shape of the
dose-response curve). With only two exceptions, the two-stage
model yielded reasonable fits to the data after elimination of
one or more high-dose terms. The exceptions are D[ah]A in
Table IV-7 and both D[ah]A and B[a]P, shown in Table IV-19; in
these cases a multistage model could not fit the data either.
Thus, the available dose-response data are not sufficient to
determine which of these two models provides the better
description of the data. The two-stage model is preferred for
parameter estimation, for two reasons: (a) it has the same
form in all cases, whereas the multistage model invokes
different numbers of stages to fit different data sets; and (b)
it is based on a specific model of the two affected
transitions, whereas the multistage model leaves both the
nature and the number of the affected transitions unspecified.
Two parameters were used to obtain relative potency
estimates: the maximum likelihood estimate of the linear term
in the one-hit model and the 95% upper bound linear term in the
multistage model, which are also shown in Appendix A. The
v-5
-------
ratio of each of these parameters to that derived for B[a]P
"ields the relative potency estimates for the one-hit and
multistage models, respectively.
Tables V-l and V-2 summarize the relative potency
estimates derived for each PAH from the bioassay data in
Sections III and IV using the one-hit and multistage models.
To evaluate these estimates and their consistency as compared
to those obtained using the two-stage model, the following
analyses were conducted.
For each PAH with three or more estimates of relative
potency, the variance of the log1Q relative potency was
calculated in order to evaluate the consistency of the
\
estimates. The log transformation was used because the
relative error is of primary importance, not the absolute
variability. The results of this analysis are displayed in
Table V-3. The only statistically significant difference
(using a standard two sided "F" test at the p - 0.05 level) in
the estimates is for the comparison of the consistency of the
two-stage model versus the one-hit model for the relative
potencies of B[b]F. The statistical power of the "F" test is
poor due to small sample sizes, however, so there may be other
differences in consistency that remain undetected.
Obtaining average relative potencies is an approach that
appeals to some investigators instead of attempting to select
the best study on which to base the relative potency
estimates. To illustrate this approach, the geometric means of
the potencies obtained for each PAH using the three different
V-6
-------
TABLE VI
ESTIMATES Of RELATIVE POTENCY (ONE-HIT MCDEL)
ENpcrlaantat Stud/
• (•IP BtblF I(J1F
• (UF IND
• UM1P OlahlA IU1A CH
BdnolF
Dautsdi-Uenzal at mi. (1965) (IV-2)b 1* 0.0929
laVola at at. (1962) (IV-J)
Mafoa tt al. (1980) (IV-6)
Pfelffar (1977) (IV-7)
• In^iM and Falk (1969) (IV-9)
Van Duurcn tt al. (1966) (IV-11)
Hoftaan ft Wynder (1966) (IV-1J)
ioffavi ft Wynder (1966) (IV-U)
Wyndtr ft Noffawt (1959) (IV-18)
•ryan ft Shlakin (1943) (IV-19)
O.OS36 0.146 0.0123
1 0.619* O.U30* 0.0258
1* 0.080* 0.021 0.008 0.0177 0.018
1* 1.17*
<0.001
0.0606s 0.00980
0.0736 0.00517
0.0101
1.046*
0.456*
<0.001
<0.001
0.00270
•Ulttiout hl^i doM gr.
**n* tabta ruber* In Uiidi th« bloasuy dita and relative potency ettlaatet derived using the
two-stag* aodel for cadi bloattay are shown In parentheses.
Abbreviations: IJalP. beraolalpyrene; ilbjf. bemolblfluoranthene; i[JJF. beniojjlfluoranthene; CP. eyclopentadlenofc.dlpyrene; BlklF.
beniolklfluoranthene; l». Indenolc.dlpyrene; Bl^iJlP. benzolghilperylene; OlahJA. dibeni(ah)anthracene; BU1A.
bentUlanthracene; CN, chrysene; Btmojf. benzolanolfluoranthene.
-------
TABLE V 2
ESTIMATES OF KLATIVE POTENCY (MATISTAGE MODEL)
Experimental Study
• (•IP ilbJF IIJ1F 0>
ItkJF I MO
DCahlA IU)A CN
• IwwlF
Deutsch-Ueniel at al. (1963)
(IV-2)C 1*
LaVole at al. (1982) (IV-3) 1
Naba at •(. (I960) (IV-6) 1*
Pfclffer (1977) (IV-7) l"
iM and Falk (1969) (IV-9) 1
Duur«n «t »l. (1966) (IV-11) 1
MoffMn 1 Uyncter (1966) (IV 13) 1
NoffMn & Uyndtr (1966) (IV-U) 1
Uyndsr t Noffm (1959) (IV-IB) 1
•ryan t Shlrtln (1943) (IV-19) 1
0.105 0.085 0.246 0.025*
0.562* 0.134* 0.0238
0.201* 0.0648 0.0268 0.0279 0.0302
<0.001
0.0292* 0.0250
0.131 0.0665
1.17*
0.0137
4.05*
0.488b
<0.001 <0.001
0.0132
•without high «*>•• gr.
^)ld not •Uxtantlally change after dropping high dose groqp.
eTht table nurtwra In Oildi the blouuy data and relative potency estimate* derived uilng the
two-stage aodel for each bloat My are ihown In parenthese*.
Abbreviations: see Table V-1.
-------
TABLE V-3
VARIANCE OF LOGio RELATIVE POTENCY BY
AGENT AND ESTIMATION METHOD
PAH
B[b]F
B[k]F
B[ghi]P
IND
D[ah]A
One-Hit
0.2456
0.0598
0.0381
0.1433
0.0500
Multistage
0.1351
0.0905
0.0602
0.2161
0.2133
Two-Stage
0.02243
0.0759
0.0952
0.2294
0.0956
aSignificantly different from the estimate obtained using the
one-hit model at the p « 0.05 level using a standard two-sided
-F- test:
0.2456
F(3,3) - 9.28 < - 10.96
0.0224
Abbreviations: see Table V-l
V-9
-------
models are shown in Table V-4 for all PAHs that have three or
more independent estimates. Also displayed are the geometric
means of each PAH among methods of estimation and for methods
of estimation among PAHs. The geometric means for the same
estimation methods indicate that on average the three
approaches yield comparable results, with the two-stage method
slightly more conservative (i.e., yields a higher estimate of
risk). Selecting what is considered to be the most reliable
study apparently yields results that tend to be more
conservative than combining estimates.
An additional comparison of relative potencies from the
"best" studies (see Section IV) using the three estimation
methods is shown in Table V-5. For these estimates, the
two-stage approach always gives higher relative potencies than
the one-hi.t method and is close to or greater than that
obtained using the multistage method, except for D[a,h]A. This
comparison amplifies the assertions that the general two-stage
approach is conservative and that selective biases to .minimize
risk have not been introduced by subjective handling of data.
The effects of the various approaches and assumptions used
to obtain the linear terms for the B[a]P dose response models
are shown in Table V-6. For ingestion exposure, the two-stage
model allows for more curvature and is a point estimate of risk
so a lower value is obtained than from the other models as is
shown in the second line of Table V-6. The use of a data
subset with more consistent exposure and duration of follow up
actually increases the upper bound linear term
V-10
-------
TABLE V-4
GEOMETRIC MEANS OF RELATIVE POTENCIES BY
. AGENT AND ESTIMATION METHOD
PAH
B[b]F
B[k]F
B[ghi]P
IND
D[ah]A
Total
One-Hit
0.166
0.0290
0.0085
0.0585
0.823
0.0723
Multistage
0.228
0.0384
0.0346 •
0.0729
1.323
0.1239
Two-Stage
0.168
•B
0.0321
0.0266
0.1094
2.415
0.1305
Total
0.185
0.0329
0.0199
0.0776
1.380
0.1053
-Best" Study3
(Two-Stage)
0.140
0.066
0.022
0.232
1.110
0.1392
aRecommended relative potency
Abbreviations: see Table V-l
V-ll
-------
TABLE V-5
COMPARISON OF RELATIVE POTENCY ESTIMATES DERIVED FROM THE
•BEST" STUDIES USING THE TWO-STAGE, ONE-HIT, AND '
MULTISTAGE LINEAR 95\ UPPER BOUND METHODS
PAH
•B[a]P
B[a]A
B[b]F
B[k]F
B[ghi]P
CH
D[ah]A
IND
Two-Stage
1.0
0.145
0 . 140
0.066
0.022
0.0044
1.11
0.232
One-Hit
1.0
0.010
0.093
0.054
0.012
0.0023
1.046
0.146
95* Upper Bound
1.0
0.014
0.105
0.085
0.025
0
4.05
0.246
Abbreviations: see Table V-l
V-12
-------
TABLE V-6
EFFECT OF APPROACHES AND ASSUMPTIONS ON ESTIMATES
OF THE LINEAR TERM IN B[a]P DOSE RESPONSE MODELS
Route of
Exposure
Linear Term [mg/kg]'1
One-Hit Multistage (95%) Two-Stage
Modification (pointrBi) (upper boundrqi) (point:2AS)
Ingestion
Non-homogeneous
length of expo-
sure and follow-
up data used
Homogeneous
length of expo-
sure and follow-
up data used
Inhalation Equivalency in
air
High Dose
omitted
All data
included
Additional sur-
face area cor-
rection
High dose
omitted
11.53b
14.88
13.04
5.74a
0.917
0.789
0.301
0.453a
6.11b
aValues derived in this report
bEPA contractor's estimates
V-13
-------
from 11.53 to 13.04. Using a different data set from that used
by the EPA's contractor is alone thus not decreasing risk. The
" -o-stage approach using the more homogeneous data yields a 50%
overall reduction in the value obtained by EPA's contractor.
For inhalation exposure, the curvature of the dose
response relationship and use of point estimates give a 62%
reduction, 0.789 to 0.301, using the two-stage model as
compared to the upper bound approach. However, including the
high dose term and adjusting for duration of survival increase
the two-stage model estimate by 50%, from 0.301 to 0.453. The
major numerical difference between the methods stems from the
assumptions concerning what constitutes equivalent exposure
levels between species. The EPA contractor applied an
additional surface area correction to the data by expressing
the slope in mg/kg units. This approach is questionable and
increases the term by a factor of 3 /70/0.05 - 7.75, from
0.789 to 6.11. The reduction in potency after all adjustments
in the data and method have been made is 93% when the method
described in this report is used as compared to that derived by
EPA's contractor, changing from 6.11 to 0.453.
V-14
-------
VI. VALIDATION OF THE COMPARATIVE POTENCY APPROACH
This chapter describes how the appropriateness of the
two-stage model and the accuracy of the comparative potency
estimates for PAHs can be partially validated using
experimental data from the literature. Studies that used
combinations of PAHs to induce cancer in laboratory animals can
be used to show how tumor rates predicted by the two-stage
model and comparative potencies developed in this report
compare to the actual tumor rates observed in the experiments.
The criteria for a useful study for such a comparison are
discussed below, and a study in which PAHs were tested for
carcinogenesis using mice is evaluated. The model predicts
tumor rates reasonably well in this experiment at lower doses,
where deviations due to biological interactions are less likely.
An unbiased, sensitive, independent validation of' the
predictions of the low-dose additive two-stage or any other low
dose extrapolation model is not possible using currently
available experimental techniques. The tumor rates predicted
by the model in the low dose range are far below those that can
be practically measured using animal bioassays. As a result,
direct experimental evidence for the model must be obtained at
exposure levels far in excess of those expected in the
environment. Use of high experimental levels of exposure,
however, can lead to biological interactions that would not be
expected to occur at environmental levels.
VI-1
-------
Among the more important interaction mechanisms
anticipated at high doses are exposure saturation and enzyme
saturation. In the first case, PAHs are applied to mouse skin
at such high levels that they accumulate on the surface and are
absorbed at relative rates that are lower than that at lower
doses. As a result, higher doses administered to the animal
are proportionally greater than that which is actually absorbed
and predicted to reach target sites, leading to overpredictions
of tumor risk at high doses. In the second case, if the levels
of metabolizing enzymes are insufficient to convert all of the
carcinogenic PAHs in a mixture to their reactive derivatives,
the number of^ieactive derivatives will be fewer when PAHs are
s
administered together as a mixture than would be predicted for
the sum of the PAHs had they been administered separately;
tumor rates would consequently be less than those predicted by
the hypothesis of dose-additiyity.
An additional factor that makes direct verification of the
predictions of the model difficult is that no experiments using
combinations of the specific chemicals used to evaluate
potential health risks at hazardous waste sites have been
conducted. As a result, validation is dependent upon the
combined exposure experiments that are available in the
literature. Such experiments do not include all the PAHs
anticipated at hazardous waste sites such as former
manufactured gas plants. In spite of these limitations, using
the low-dose additive two-stage model and relative potency
VI-2
-------
estimates to predict a carcinogenic response to multiple PAHs
can provide partial verification of the approach.
A useful exercise in terms of validating both the
comparative potency method as well as the indicator chemical
approach to estimating risks at hazardous waste sites would be
to use these approaches to estimate the cancer risks of some
environmental complex mixtures for which bioassay and detailed
analytical data are available. Appropriate information may
exist for mixtures such as combustion products, coal tar
extract, and cigarette smoke condensate. Such testing may
reveal that the toxicity of such mixtures is partly due to
components other than those commonly considered, which should
therefore be included.
1. Criteria for Useful Combined Exposure Studies
For a study to provide a useful evaluation of the
comparative potency method, two essential elements are
required. First, B[a]P must have been included in the assay
and have demonstrated a dose response relationship. Second,
two or more PAHs must have been tested simultaneously. In
addition, an experiment in which high-dose interaction did not
affect the tumor response is desirable. Results where a
quadratic dose-response relationship exists up to the highest
response levels demonstrate the absence of high-dose
interaction under the assumed model. Evaluating an experiment
in which high-dose interaction occurred would require
additional information and assumptions concerning the nature of
VI-3
-------
the mechanisms of interaction. Sufficiently detailed
biological information with which to postulate such mechanisms
of interaction are not likely to be available. In this
situation, an empirical (non-biologically based) dose-response
modeling approach may be appropriate as a first approximation.
One study has been identified that contains sufficient
information to allow an attempt to verify the basic model.
This experiment is discussed and evaluated in the next section.
2. Carcinoaenicitv of Combined PAHs
Schmahl et al. (1977) investigated the carcinogenicity of
combinations of PAHs as part of a study to evaluate automobile
exhaust. Eleven PAHs were selected (not all carcinogens),
s
combined in proportions thought to reflect those in automobile
exhaust, and applied twice weekly to the backs of NMRI mice.
Animals were killed if a carcinoma appeared at the site of
application, or were observed until their natural deaths.
table VI-1 shows the PAHs tested and their relative proportions
by weight in the mixtures. Table VI-2 shows the results of tl.»
experiments. The tumor rates observed in the experiments can
be compared to those predicted using the comparative potency
method. In order to do so, a dose-response relationship for
B[a]P is derived in the next section and used to predict the
response to a mixture given its exposure in units equivalent to
B[a]P.
VI-4
-------
TABLE VI-1
TREATMENT GROUPS FOR CARCINOGENESIS
EXPERIMENT USING COMBINED PAHs
Dose (yg)
Control: Acetone
Benzo[a]pyrene
Mixture 1:
Benzol a]pyrene
Dibenz[a,h]anthracene
Benz[a]anthracene
Benzo[b]fluoranthene
Mixture 2:
Total
as solvent
1.0
1.0
0.7
1.4
SLJ.
4.0
1.7
1.7
1.2
2.4
6.8
3.0
Phenanthrehe
Anthracene
Fluoranthene
Pyrene
Chrysene
Benzo[e]pyrene
Benzo [ghi ] perylene
Total
27.0
8.5
10.8
13.8
1.2
0.6
3.1
65.0
81.0
25.5
32.4
41.4
3.6
1.8
9.3
195.0
243.0
76.5 *
97.2
124.2
10.8
5.4
27.9
585.0
729.0
229.5
291.6
372.6
32.4
16.2
83 .7
1755.0
SOURCE: Schmahl et al. (1977).
VI-5
-------
TABLE VI-2
RESULTS OF CARCINOGENESIS EXPERIMENT
USING COMBINED PAHS
Application
Dose (ng)a
Number of Animals. With Carcinoma
Effective Number of Animals
Observed
Control
Benzo[a]pyrene
Mixture 1
Mixture 2
—
1.0
1.7
3.0
4.0
6.8
12.0
65.0
195.0
585.0
1755.0
0/81
10/77
25/88
43/81
25/81
53/88
63/90
1/85
0/84
1/88
15/86
(13%)
(28\)
(53\)
(31%)
(60%)
(70%)
( 1%)
Predicted
0.5
9.0°
22.7
45.2
28.5
59.6
86.7<=
0.8
x s!e
27.8
aThis amount was applied twice weekly to the backs of NMRI mice.
bP(x)-l-exp-[.006116(1+3.52X)2]
cSignificant deviation from expected at p » 0.01 using X2
goodness-of-fit test
Source: Schmahl et al. (1977)
VI-6
-------
3. Derivation of the BfalP Dogg-Response Model
In Section III of this report the following quadratic form
of the dose-response model for B[a]P was derived:
P(x) - 1-exp C-A(l+Sx)2],
where A is the background transition rate parameter and S is
the exposure-related transition rate parameter. The control
data from the Schmahl et al. (1977) study were used to estimate
A, employing the Bayesian procedure discussed previously.
Given a tumor rate of 0/80 in the control group, the Bayesian
estimate of A was obtained from the following relationship:
N
(.5)/(80.+ l) - 1-exp-A or A - 0.006116.
The parameter S was estimated using an approximate likelihood
method that yielded a relative transition rate of S - 3.52.
The dose-response model thus has the numerical form:
P(x) - l-exp-[0.006116(1+3.52x)2]. (VI-1)
The predicted number of carcinoma-bearing animals was obtained
by multiplying the tumor rate obtained from equation (VI-1)
times the effective number of animals in each dose group. The
results of these calculations are shown in the last column of
Table VI-2. A standard X2 goodness-of-fit test yielded the
result X2 - 0.677 with 4-2 » 2 degrees of freedom, which has
VI-7
-------
a "p" value of 0.73 associated with it and therefore indicates
an adequate fit.
Since the quadratic model fits the B[a]P data well in the
observed dose range of 0 to 3 pg, high-dose interactive
effects are not expected to be important factors in that
range. The predictions that are the most consistent with
low-dose additivity would therefore be for mixtures composed of
PAHs with a combined dose equivalent to 3 ng of B[a]p' or less.
4. Predictions of Tumor Rates due to Exposure to PAH Mixtures
In order to predict the tumor rates expected as a result
of exposure to combined PAHs, the potencies of the PAHs in each
mixture in the Schmahl et al. (1977) study were calculated in
terms of B[a]P equivalents; these are shown in Table VI-3. The
lowest experimental dose was used for the calculations/ as well
as the relative potencies obtained in Section IV. Since the
relative proportions of PAHs in each mixture were constant for
all exposure levels, the B[a]P equivalent units at higher
exposure levels were obtained by simple proportional
multiplication. These B[a]P exposure units were substituted
into equation V-l and multiplied by the effective number of
animals in each dose group to obtain the predicted number of
animals with carcinomas. The predicted numbers were then
compared to the observed numbers, which are shown graphically
in Figure Vl-1.
VI-8
-------
TABLE VI-3
CALCULATIONS OF B[a]P EQUIVALENT EXPOSURE UNITS
Relative Dose Equivalent
Potency3 (ng) B[a]P
Dose
Mixture 1
Benzo[a]pyrene
Dibenz[a,h]anthracene
Benzo[a]anthracene
Benzo[b]fluoranthene
Mixture 2
Phenanthrene*3
Anthracene13
Fluoranthene*5
Pyreneb
Chrysene
Benzo [e] pyrene*3
Benzo[ghi]perylene
1.000
1.110
0.145
0.140
Total
0.0
0.0
0.0
0.0
0.0044
0.0
0.022
1.0
0".'7
1.4
Q.9
4.0
27.0
8.5
10.8
13.8
1.2
0.6
Total
65.0
1.000
0.777
0.203
Q.126
2.106
0.0
0.0
0.0
0.0
0.00528
0.0
Q.Q6S2
0.07348
NOTE: Doses equivalent to B[a]P for other exposure levels were
obtained by using simple ratios:
Dose (pg)
Equivalent
B[a]P Dose
Mixture 1
Mixture 2
4.0
6.8
12.0
65.0
195.0
585.0
1755.0
2.106
3.580
6.318
0.07348
0.22044
0.66132
1.98396
•From Tale IV-26
^Considered noncarcinogenic by IARC (1983)
VI-9
-------
FIGURE VI-1
COMPARISON OF OBSERVED8 AND PREDICTED TUMOR RATES
FOR MICE EXPOSED TO PAH MIXTURES
Probability of
Tumor Response
1.0
—«• Dose Response Curve Based on the Model:
P(x)-1 - «xp - [0.006116(1*3.52 x)2 ]
to Mbrtur* 1
to Mfartur* 2
O Obf»rv»d
Comb4n*d
aObserved txunor rates are based on Schnahl et al. (1977)
VI-10
-------
The following observations can be made from the
information given in Table Vl-2. The two lower Mixture 1
exposure levels yielded predictions that were slightly larger
but statistically consistent with the observed numbers of
tumor-bearing animals. The highest dose led to an
overprediction of the observed number by a significant amount.
However, exposure at this level was more than twice that of the
dose range used for B[a]P when administered alone and as a
result is likely to have led to a high-dose interaction
phenomenon. The relevance of this value to low-dose
extrapolation is therefore questionable. For Mixture 2, the
predicted values were greater than the observed values. This
result may have been due to high-dose interactions or, more
N
probably, overestimation of the potency of benzo[ghi]perylene.
It should be emphasized that the predicted values were obtained
without using the observed values in any manner.
5. Summary
Schmahl et al. (1977) tested B[a]P and two mixtures of
both carcinogenic and noncarcinogenic PAHs for carcinogenicity
when applied to the backs of mice. For each dose-group it was
possible to predict tumor rates without using the experimental
data from this specific study in any manner. The method
yielded good estimates for combinations of the carcinogens
benzo[a]pyrene, dibent[a,h]anthracene, benz[a]anthracene, and
benzo[b]fluoranthene in the same dose range observed for B[a]P
alone (Mixture 1). Using potency estimates for chrysene and
Vl-11
-------
benzo[ghi]perylene, the effects of the agents in Mixture 2
(weak carcinogens or noncarcinogens) were systematically
overestimated. In general, the method tends to overestimate at
high exposure combinations and give good results at more
moderate levels. As a result, the approach may be viewed as
conservative for the mixtures evaluated, which in turn enhances
its overall credibility.
VI-12
-------
VII. REFERENCES
ADRIAENSSENS, P.I., WHITE, C.M., and ANDERSON, M.W. 1983.
Dose-response relationships for the binding of
benzo[a]pyrene metabolites to DNA and protein in lung,
liver, and forestomach of control and butylated
hydroxyanisole-treated mice. Cancer Res. 43:3712-3719
ANDERSON, E.L., and THE CARCINOGEN ASSESSMENT GROUP (CAG) OF
THE U.S. ENVIRONMENTAL PROTECTION AGENCY. 1983.
Quantitative approaches in use to assess cancer risk.
Risk Analysis 3:277-295
ARMITAGE, P., and DOLL, R. 1954. The age distribution of
cancer and a multi-stage theory of carcinogenesis. Brit.
J. Cancer 8:1-12
ARMITAGE, P. and DOLL, R. 1957. A two-stage theory of
carcinogenesis in relation to the age distribution of
human cancer. Brit. J. Cancer 11:161-169
ARMITAGE, P. 1985. Multistage models of carcinogenesis.
Environ. Health Perspect. 63:195-201
BELLAR, T.A., KICHTENBERG, J.J., and KRONER, R.C. 1974. The
occurrence of organohalides in chlorinated drinking
water. J. Am. Hater Works Assoc. 66:703-706
BERENBLUM, I., and SHUBIK) P. 1947. A new, quantitative
approach to the study of the stages of chemical
carcinogenesis in the mouse's skin. Br. J. Cancer
1:383-391
BINGHAM, E., and FALK, H.L. 1969. The modifying effect of
carcinogens on the threshold response. Arch. Environ.
Health 19:779-783
BOCK, F.G., and DAO, T.L. 1961. Factors affecting the
polynuclear hydrocarbon level in rat mammary glands.
Cancer Res. 21:1024-1029
BOREK, C., and SACHS, L. 1968. The number of cell generations
required to fix the transformed state. Proc. Natl. Acad.
Sci. USA 59:83-85
BRESNICK, E., MCDONALD, T.F., YAGI, H., JERINA, D.M., LEVIN, W.,
WOOD, A.W., and CONNEY, A.M. 1977. Epidermal hyperplasia
after topical application of benzol a]pyrene,
benzo[a]pyrene diol epoxides, and other metabolites.
Cancer Res. 37:984-990
VII-1
-------
BRYAN, W.R., and SHIMKIN, M.B. 1943. Quantitative analysis of
dose-response data obtained with three carcinogenic
hydrocarbons in strain C3H male mice. J. Natl. Cancer
Inst. 9:503-531
CHU, M.M.L., and CHEN, C.W. 1984. Evaluation and Estimation
of Potential Carcinogenic Risks of Polynuclear Aromatic
Hydrocarbons. Paper presented at the Pacific Rim Risk
Conference. Honolulu, Hawaii. Unpublished (available
from the authors, U.S. EPA Carcinogen Assessment Group)
COGLIANO, V.J. 1986. The U.S. EPA's methodology for adjusting
the reportable quantities of potential carcinogens.
Proceedings of Superfund '86 Conference: Management of
Uncontrolled Hazardous Waste Sites, Washington, D.C.
CONNEY, A.H., PANTUCK, E.J., HSIAO, K.C., KUNTZMAN, R., ALVARES,
A.P., and KAPPAS, A. 1977. Regulation of drug metabolic...
in man by environmental chemicals and diet. Fed. Proc.
36:1674-1652
COOPER, C.S., GROVER, P.L., and SIMS, P. 1983. The metabolic.,
and activation of benzo[a]pyrene. In Bridges, J.W., and
Chaseand, L.F., eds. Progress in Drug Metabolism. Vol.
7. John Wiley & Sons, New York. Pp. 295-395
CRADDOCK, V.M. 1971. Liver carcinoma induced in rats by
single administration of dimethylnitrosamine after partial
hepatectomy. J. Natl. Cancer Inst. 47:899
CREASIA, D.A., POGGENBURG, J.K., and NETTESHEIM, P. 1976.
Elution of benzo[a]pyrene from carbon particles in the
respiratory tract of mice. J. Toxicol. Environ. Health
1:967-975
DEUTSCH-WENZEL, R.P., BRUNE, H., GRIMMER, O., DETTBARN, G., and
MISFELO, J. 1983. Experimental studies in rat lungs on
the carcinogenicity and dose-response relationships of
eight frequently occurring environmental polycyclic
aromatic hydrocarbons. J. Natl. Cancer Inst. 71:539-544
DUMOUCHEL, W.H. and HARRIS, J.E. 1983. Bayes methods for
combining the results of cancer studies in humans and
other species. J. Amer. Statist. Assoc. 78:293-315
EISENSTADT, E., WARREN, A.J., PORTER, J., ATKINS, D., and
MILLER, J.H. 1982. Carcinogenic epozides of
benzo[a]pyrene and cyclopenta[cd]pyrene induce base
substitutions via specific transversions. Proc. Natl.
Acad. Sci. 79:1945-1949
VII-2
-------
ENVIRONMENTAL PROTECTION AGENCY (EPA). 1984. Health^ Effects
Assessment for Polycyclic Aromatic Hydrocarbons (PAH).
Environmental Criteria and Assessment Office, Cincinnati,
Ohio. EPA 540/1-86-013
ENVIRONMENTAL PROTECTION AGENCY (EPA). 1986a. Guidelines for
Carcinogen Risk Assessment. Federal Register 51(185):
33991-34003. September 24, 1986
ENVIRONMENTAL PROTECTION AGENCY (EPA). 1986b. Guidelines for
the Health Risk Assessment of Chemical Mixtures. Federal
Register 51(185):34013-34025. September 24, 1986
ENVIRONMENTAL PROTECTION AGENCY (EPA). 1986C. Health Effects
Assessment Summary Table. Revised Summary Tables for 58
Health Effects Assessment Documents (HEAs). May 6, 1986
FAHL, W.E., SCARPELLI, D.G., and GILL, K. 1981. Relationship
between benzo[a]pyrene-induced DNA base modification and
frequency of reverse mutations in mutant strains of
Salmonella typhimurium. Cancer Res. 41:3400-3406
FINNEY, D.G. 1964. Statistical Methods in Biological Assay.
Second Edition. Hafner Publishing Co. New York
s
GART, J.J. 1966. Alternative analyses of contingency tables.
J. Roy. Stat. Soc. B 28:164-179
GELBOIN, H.W. 1980. Benzo[a]pyrene metabolism, activation, and
carcinogenesis: Role and regulation of mixed-function
oxidases and related enzymes. Physiol. Rev. 60:1107-1166
GROVER, P.L., and SIMS, P. 1968. Enzyme-catalysed reactions
of polycyclic hydrocarbons with deoxyribonucleic acid and
protein in vitro. Biochem. J. 110:159-160
HABS, M., SCHMAHL, D., and.MISFELD, J. 1980. Local
carcinogenicity of some environmentally relevant
polycyclic aromatic hydrocarbons after lifelong topical
application to mouse skin. Arch. Geschwulstforsch.
50:266-274
HEIDELBERGER, C., and WEISS, S.M. 1951. The distribution of
radioactivity in mice following administration of 3,4-
benzpyrene-5-C14 and l,2,5,6-dibenzanthracene-9,10-C14.
Cancer Res. 11:885-891
HENNINGS, H., SHORES, R., WENK, M.L., SPAJK3LER, E.F., TARONE,
R., and YUSPA, S.H. 1983. Malignant conversion of mouse
skin tumors is increased by tumor initiators and unaffected
by tumor promoters. Nature 304:67-69
VII-3
-------
HIRAKAWA, T., ISHIKAWA, T. , NIMOTO, N., TAKAYAMA, S., and
KITAGAWA, T. 1979. Induction of enzyme-altered islands
in rat liver by a single treatment with benro[a]pyrene
after partial hepatectomy. Gann 70:393-394
HOFFMAN, D., and WYNDER, E.L. 1966. Beitray zur carcinogenen
Wirkung von Dibenzopyrenen. Z. Krebsforsch. 68:137-149
INTERNATIONAL AGENCY FOR RESEARCH ON CANCER (IARC). 1983. IARC
Monographs on the Evaluation of the Carcinogenic Risk of
Chemicals to Humans. Volume 32. Polynuclear Aromatic
Compounds. Part 1. Chemical, Environmental and
Experimental Data. World Health Organization, Lyon, France
JEFFREYS, H. 1961. Theory of Probability. 3rd ed. Oxford
University Press, Oxford
KADEN, D.A., KITES, R.A., and THILLY, W.G. 1979., Mutagenicity
of soot and associated polycyclic aromatic hydrocarbons to
Salmonella typhimurium. Cancer Res. 39:4152-4159
KALUNAGA, T. 1974. Requirement for cell replication in the
fixation and expression of the transformed state in mouse
cells treated with 4-nitroquinoline-l-oxide. Int. J.
Cancer 14:736-742
KOTIN, P., FALX, H.L., and BUSSER, R. 1959. Distribution,
retention and elimination of C14-3,4-benropyrene after
administration to mice and rats. J. Natl. Cancer Inst.
23:541-55
LAND, H., PARADA, L.F., and WEINBERG, R.A. 1983. Cellular
oncogenes and multistep carcinogenesis. Science
222:771-777
LASKIN, S., KUSCHNER, M., and DREW, R.T. 1970. Studies in
pulmonary carcinogenesis. In Hanna, M.G., Nettesheim, P.,
and Gilbert, J.R., eds. Inhalation Carcinogenesis:
Proceedings of the U.S. Atomic Energy Commission Symposia..
Series. Oak Ridge, Tennessee
LaVOIE, E.J., AMIN, S., HECHT, S.S., FURUYA, K., and HOFFMAN,
D. 1982. Tumour initiating activity of dihydrodiols of
benzo[b]fluoranthene, benzo[j]fluoranthene, and
benzo[k]fluoranthene. Carcinogenesis 3:49-52
LEE, P.N., and O'NEIL, J.A. 1971. The effect both of time arj
dose applied on tumor incidence rate in benzopyrene skin
painting experiments. Brit. J. Cancer 25:759-770
VII-4
-------
LEVIN, W., BRUENING, M.K. , WOOD, A.W. , CHANG, R.L., KEDZIERSKI,
B., THAKKZR, D.R., BOYD, D.R., GADAGINAMATH, G.S.,
ARMSTRONG, R.N., YAGI, H., KARLE, H.M., SLAGA, T.J.,
JERINA, D.M., and CONNEY, A.M. 1980. An enantiomeric
interaction in the metabolism and tumorigenicity of ( + )-
and <-)-benzo[a]pyrene 7,8-oxide. J. Biol. Chem.
255:9067-9074
LEVIN, W., WOOD, A., CHANG, R.L., RYAN, D., THOMAS, P.E., YAGI,
H., THAKKER, D.R., VYAS, K. , BOYD, C., CHU, S.L.-Y.,
CONNEY, A.M., and JERINA, D.M. 1982. Oxidative
metabolism of polycyclic aromatic hydrocarbons to ultimate
carcinogens. Drug Metab. Rev. 13:555-580
MAHLDM, D.D., WRIGHT, C.W., CHESS, E.K., and WILSON, B.W. 1984.
Fractionation of skin tumor-initiating activity in coal
liquids. Cancer Res. 44:5176-5181
MARSHALL, C.J., VOUSDEN, K.H., and PHILLIPS, D.H. 1984.
Activation of c-Ha-ras-1 proto-oncogene by in vitro
modification with a chemical carcinogen, benzo(a)pyrene
diol-epoxide. Nature 310:586-589
MISFELD, J. 1980. The tumor-producing effects of automobile
exhaust condensate and of diesel exhaust condensata. In
Pepelko, W.E., Danner, R.M., and Clarke, N.A., eds.
Health Effects of Diesel Engine Emissions: Proceedings of
an International Symposium. Vol. 2. Environmental
Protection Agency, Cincinnati, Ohio. EPA 600/9-80-057b.
Pp. 1012-1025
MITCHELL, C.E. 1982. Distribution and retention of
benzo[a]pyrene in rats after inhalation. Toxicol. Lett.
11:35-42
MOOLGAVKAR, S.H. 1986. Carcinogenesis modeling: From
molecular biology to epidemiology. Ann. Rev. Public
Health 7:151-169
MOOLGAVKAR, S.H. and KNUDSON, A.G. 1981. Mutation and
cancer: A model for human carcinogenesis. J. Natl.
Cancer Inst. 66:1037-1052
MOOLGAVKAR, S.H., and VENZON, D.J. 1979. Two-event models for
carcinogenesis: Incidence curves for childhood and adult
tumors. Math. Biosci. 47:55-77
NATIONAL ACADEMY OF SCIENCES (NAS). 1983. Polycyclic Aromatic
Hydrocarbons: Evaluation of Sources and Effects.
National Academy Press, Washington, D.C.
MEAL, J., and RIGDON, R.H. 1967. Gastric tumors in mice fed
benzo[a]pyrene: A quantitative study. Tex. Rep. Biol.
Med. 25:553-557
VII-5
-------
NEWBOLD, R.F., and BROOKES, P. 1976. Exceptional mutagenicity
of a benzo(a)pyrene diol epoxide in cultured mammalian
cells. Nature 261:52-54
NEWBOLD, R.F., WIGLEY, C.B., THOMPSON, M.H., and BROOKES, P.
1977. Cell-mediated mutagenesis in cultured Chinese
hamster cells by carcinogenic polycyclic hydrocarbons:
Nature and extent of the associated hydrocarbon-DNA
reaction. Mutat. Res. 43:101-116
NEWBOLD, R.F., BROOKES, P., and HARVEY, R.G. 1979. A
quantitative comparison of the mutagenicity of
carcinogenic polycyclic hydrocarbon derivatives in
cultured mammalian cells. Int. J. Cancer 24:203-209
PEREIRA, M.A., BURNS, F.J., and ALBRTY, R.E. 1979. Dose
response for benzo[a]pyrene adducts in mouse epidermal
DNA. Cancer Res. 39:2556-2559
PFEIFFER, E.H. 1977. Oncogenic interaction of carcinogenic and
non-carcinogenic polycyclic aromatic hydrocarbons. In
Mohr,V., Schmahl, D., and Tomatis, L., eds. Air Pollution
and Cancer in Man. IARC Scientific Publication No. 16.
Lyon, France. Pages 69-77
PHILLIPS, D.H., GROVER, P.L., and SIMS, P. 1979. A
quantitative determination of the covalent binding of a
series of polycyclic hydrocarbons to DNA in mouse skin.
Int. J. Cancer 23:201-208
REDDY, M.V., GUPTA, R.C. , RANDERATH, E., and RANDERATH, K.
1984. 32P-Postlabelling-test for covalent DNA binding
of chemicals in vivo: Application to a variety of
aromatic carcinogens and methylating agents.
Carcinogenesis 5:231-243
RIGDON, R.H., and REAL, J. 1969. Relationship of leukemia to
lung and stomach tumors in mice fed benzo[a]pyrene. Proc.
Soc. Exptl. Biol. N.Y. 130:146-152
RIGDON, R.H., and REAL, J. 1966. Gastric carcinomas and
pulmonary adenomas in mice fed benzo[a]pyrene. J. Natl.
Cancer Inst. 34:297-305
ROOK, J.J. 1974. Formation of haloforms during chlorination
of natural waters. Water Treatment Exam. 23:234-243
SCHMAHL, D., SCHMIDT, K.G., and HABS, M. 1977. Syncarcinogenic
action of polycyclic aromatic hydrocarbons in automobile
exhaust gas condensates. In Mohr, U., Schmahl, D., and
Tomatis, L., eds. Air Pollution and Cancer in Man. IARC
Publication 16. World Health Organization, Lyon, France.
Pages 53-59
VII-6
-------
SIMS, P., and GROVZR, P.L. 1974. Epoxides in polycyclic
aromatic hydrocarbon metabolism and carcinogenesis. Adv.
Cancer Res. 20:165-274
SLAGA, T.J., TRIPLETT, L.L., and NESNOW, S. 1980. Mutagenic
and carcinogenic potency of extracts of diesel and related
environmental emissions: Two-stage carcinogenesis in skin
tumor sensitive mice (SENCAR). In Pepelko, W.E., Danner,
R.M., and Clarke, N.A., eds. Health Effects of Diesel
Engine Emissions. Proceedings of an International
Symposium. Vol. 2. Environmental Protection Agency,
Cincinnati, Ohio. EPA 600/9-80-057b. Pp. 874-897
STOWERS, S.J., and ANDERSON, M.W. 1985. Formation and
persistence of benzo[a]pyrene metabolite-DNA adducts.
Environ. Health Perspect. 62:31-39
THORSLUND, T.W., BROWN, C.C., and CHARNLEY, G. 1987. The use
of biologically motivated mathematical models to predict
cancer risk due to environmental exposures to a
carcinogen. Risk Analysis 7:109-119
THYSSEN, J., ALTHOFF, J., KIMMERLE, G., and MOHR, U. 1981.
Inhalation studies with benzo[a]pyrene in Syrian golden
hamsters. J. Natl. Cancer Inst. 66:575-577
TONG, C., FAZIO, M., and WILLIAMS, G.M. 1980. Cell cycle-
specific mutagenesis at the hypoxanthine phosphori-
bosyltransferase locus in adult rat liver epithelial
cells. Proc. Natl. Acad. Sci. 77:7377
UMBREIT, T.H., HESSE, E.G., and GALLO, M.A. 1986.
Bioavailability of dioxin in soil from a 2,4,5-T
manufacturing site. Science 232:497-499
VAN DUUREN, B.L., SIVAX, A., SEGAL, A., ORRIS, L., and
LANGSETH, L. 1966. The tumor-promoting agents of tobacco
leaf and tobacco smoke condensate. J. Natl. Cancer Inst.
37:519-526
VOSTAL, J.J. 1983. Bioavailability and transformation of the
mutagenic component of particulate emissions present in
motor exhaust samples. Environ. Health Perspect.
47:269-281
WEINSTEIN, I.E., JEFFREY, A.M., LEFFLER, S., PULKRABEK, P.,
YAMASAKI, H., and GRUNBERGER, D. 1978. Interactions
between polycyclic aromatic hydrocarbons and cellular
macromolecules. In Ts'O, P.O.P. and Gelboin, H.V., eds.
Polycyclic Hydrocarbons and Cancer. Vol. 2: Molecular
and Cell Biology. Academic Press, New York. Pp. 3-36
VII-7
-------
WHITTEMORE, A. 1978. Quantitative theories of carcinogenesis ,
Adv. Cancer Res. 27:55-58
A., and KELLER, J.B. 1978. Quantitative theories
of carcinogenesis. SIAM Rev. 20:1-30
WIGLEY, C.B., NEWBOLD, R.F., AMOS, J. , and BROOKES, P. 1979.
Cell mediated mutagenesis in cultured Chinese hamster
cells by polycyclic hydrocarbons: Mutagenicity and DNA
reaction related to carcinogenicity in a series of
compounds. Int. J. Cancer 23:691-696
WYNDER, E.L., and HOFFMAN, D. 1959. A Study of tobacco
carcinogenesis. VII. The role of higher polycyclic
hydrocarbons. Cancer 12:1079-1086.
YANG, D., TANNENBAUM, S.R., BUCK I , G. , and LEE, G.C.M. 1984.
4-Chloro-6-methoxyindole is the precursor of a potent
mutagen (4-chloro-6-methoxy-2-hydroxy-l-nitroso-indolin-
3-one oxime) that forms during nitrosation of the fava
bean fVicia fava) . Carcinogenesis 5:1219-1224
VII-8
-------
APPENDIX A
This appendix contains a table that displays the maximum
likelihood parameter estimates, the 95\ upper bounds on the
linear terms, and X goodness-of-fit test values for the
multistage and one-hit models, respectively, along with p
values that indicate how well the models fit the observed tumor
data for each of the PAH carcinogenesis bioassays evaluated in
Sections III and IV.
A-l
-------
APPENDIX A
ESTIMATES OF PARAMETERS DCRIVED FROM A MULTISTAGE 01 ONE-NIT MODEL
Stud/
PAN
Model
Meat mi Upton (1M7)
•aP
Multistage (MS)
One-hit (ON)
Thyesen et el. (1961)
•eP
Multistage
One-hit
Deutech-Uenzel et at. (1963)
•aP
Multistage
Multistage
One-hit
One-hit
• tblF
Multistage
One-hit
KklF
Multistage
One-hit
I NO
Multistage
Multistage
One-hit
One-hit
• tghllP
Multistage
One-hit
Muter of
Dose Croupe qO
6
6
4
Is
4
5«
4
4
4
4
4
3a
4
3«
4
4
i Likelihood PeraMter EstlMtes
q2 q3 q4 qn
0
0
0
0
P
0
0
0
0
0
0
0
6.77E-03
0
6.77E-03
0
0
0
0
8. HE -03
0
3.44E-02
2.07E«00
2.93E-01
2.60E«00
2.44CHW
0
2.26E-01
8.43E-02
1.30E-01
2.49E-01
3. SAC -01
2.50E-01
3.UE-01
2.77E 02
3.00C-02
0 0
4.40E-03
1.01E*00 0
9.20E*00
2.74E-01 2.42E-02
1.43C 02 0
0 0
0
0
0 -
6.4SE-04 0
p-v«lue
7.23E-03
1.59E-02
2.96E-02
5.71E 02
3.21E*00
2.29E*00
3.JK»00
3.J5£«00
2.41E-01
3.66E-01
1.89E-01
1.95E-01
3.40E-01
5.63E-01
3.40E-01
5.63E 01
5.77E-02
5.77E 02
1.05
4.20
0.62
2.71
2.7
<0.001
3.23
3.08
0.12
2.23
0.67
0.83
2.17
0.75
2.16
0.75
0.2
0.19
0.90
0.37
0.43
0.10
0.10
-1
0.20
0.08
0.73
0.33
0.41
0.66
0.34
0.39
0.34
0.39
0.65
0.91
-------
APPEMHX A (continued)
Study
PAN
Hotel
laVole et •(. (1962)
• (•IF
Multistage
One-hit
l(b)F
Multistage
Multistage
One-hit
One-hit
•uaber of
Ooee CroitM
Kulu likelihood Parameter Ettlaatee
ql q2 q3 q* qn
x2 p-velue
Multistage
Multistage
One-hit
One-hit
•tur
Multistage
One-hit
Hebe et el. (I960)
2
2
4
5a
4
5a
4
5a
4
3a
4
4
0
0
0
0
0
0
l.OSC-03
0
3.0SC-03
0
0
0
6.32C-02
6.32C-02
2.97E-02
J.91E-02
2.57E-02
3.91E-02
5.9U-01
9.07t-03
'5.9IC-OJ
9.071-03
1.AX-03
1.A3C-03
0
0
0
0
0
Multistage
Multistage
One-hit
One-hit
l(b)F
Multistage
Multistage
One-hit
One-hit
•(Jir
Multistage
One-hit
4
3a
4
3a
4
3a
4
3a
4
4
0
0
0
0
0
0
0
0
0
2.48C-01
0
2.49E-01
2.90E-01
3.2K-03
0
4.32E-02
2.32E-02
6.05E-03
6.05E-03
0 0
1.28C-01
0 9.34C-04
4.95E-03
,
0 0
9.90E-02
9.90E-02
3.UC-02
S.UC-02
3.UC-02
S.MC-02
8.60C-03
1.33C-02
8.60E-03
1.33E-02
2.3AC-OI3
2.56E-03
3.11E-01
1.93C-01
3.12E-01
3.86C-01
2.66C-02
3.87E-02
4.74E-02
4.08E-02
1.25E-02
1.25E-02
<0.001
<0.001
10.3
2.23
10.3
2.23
2.01
0.6
2.01
0.6
1.93
1.93
7.36
1.28
7.37
5.77
O.U
<0.001
9.56
0.49
0.09
0.09
-1
-1
0.06
O.U
0.06
O.U
•(0.001
O.U
<0.001
O.U
0.38
0.38
0.25
0.26
0.025
0.016
0.71
0.92
0.008
0.48
0.95
0.95
-------
APPENDIX « (continued)
Study
PAI
Nodal
•ate at al. (I960)
tf
NultUtag*
One-hit
•Mr
Mult I state
One-hit
•0
Multistage
Cm-hit
Melffer (1977)
• (•IP
Nultlttate
Mult(state
One-hit
One-hit
DlahlA
Mult Istate
Mult(State
Ont-hlt
Om-hlt
• InghM and Folk (1969)
• (•IP
Multistage
Ont-hlt
• (•1A
Nultlit*9«
One-hit
of
DOM Croup*
LUtllhood
EitlMtt*
4
4
4
4
4
4
0
0
5.UE-OJ
S.Ui-01
S.2ME-03
S.2K-03
0
2.33C-01
0
0
0
0
7
6m
7
ec
7
6«
5
5
5
5
0
0
0
2.16£*01
4.0IC-06
8.19C-U 2.68E-02 0
7.42C-02 3.39C-02 0
8.1H-02 2.60C-02
7.42E-02 3.39C-02
1.271-01 2.50C-02 0
1.12C-OI 3.92C-02 •
1.27E-01 2.MC-02
1.02E-01 3.42C-02
0
0
0
0
»2 p-v«(u«
1.44E-02 2.16E-01
1.UE 02 2.I8E-OI
4.16C-03
5.29E-03
$.171-03
5.m-03
S.3K-03
S.3K-03
0 0(45»<^) 3.02E-02
0 O(ojS) 3.8o£-02
3.02E-02
3.B6C-02
0 0«j5»q&) 2.BK-02
0 0(q5) 4.S3E-02
2.6ft -02
4.54E-02
3.37E»o 2.96t»01
4.20E»01
0 4. 10€ 01
4.10E-01
0.005
0.9
4.02
4.02
4.3
4.3
46.1
18.3
46.1
18.3
100.4
55. 7
100.4
55.7
0.001
1.21
3.17
3.17
0.98
0.64
0.26
0.26
0.23
0.23
<0.001
0.001
<0.001
0.001
0
-------
APP6KHI A (continued)
Ul
Stud/
PAN
Notfal
VarOuurvt tt •!. (1966)
Ont-hlt
CM
NultUt»o«
One-kit
• tblF
Nultlit»9*
Om-hlt
NultUtagt
Om-hlt
•oftawi end Uyndtr (1966)
• (•IP
Nuttlstag*
Om-hlt
•>
Nultl«t»fl«
Multistage
Om-hlt
Om-hlt
Om-hlt
•offMn and Uynder (1966)
• (•IP
NultUt*g«
One-hit
•u*er of
DOM Croups
2
2
2
2
2
2
2
2
3
3
S
4
5
4
3
3
2
2
<*>
2.886-01
2.886-01
2.886-01
2.886-01
2.886-01
2.886-01
2.556-01
2.556-01
0
0
0
0
0
0
0
0
6.906-02
6.906-02
N
ql
1.696*02
1.696*02
1.32
1.32
2.01
2.01
0
0
3.436*01
3.436*01
1.18
0
1.18
2.08
3.366-01
3.366-01
6.16
6.16
lull
«»
0
0
0
0
0
Uktllhood Para
q3
eter 6stlMtet
0
3.106*02
x2 p-value
4.866*19
4.866*19
2.23
2.23
3.43
3.43
2.376-01
2.376-01
4.906*01
4.906*01
1.82
1.43
1.82
3.80
1.23
1.23
8.97
8.97
<0.001
<0.001
<0 001
<0.001
<0.001
<0.001
0.14
0.14
0.32
0.32
9.36
0.911
9.36
4.27
2.04
2.04
<0.00t
<0.001
-1
-1
-1
-1
-1
-1
0.70
0.70
0.57
0.57
0.025
0.634
0.025
0.118
0.15
0.15
-1
-1
-------
APPENDIX A (continued)
Stud?
PAN
Hof fern and Wynder (IMA)
HO
Multistage
Ont-hit
• IghlJP
Multistage
One-kit
Wynder and Hoffswi (19W)
• (•IF
Multistage
Multistage
One-hit
Ont-hlt
DlahlA
Multistage
One-hit
CM
Multistage
One-hit
•ryan and Stilafcln (1MS)
• lelP
Kultlttaa*
One-hit
DUMA
Nultlttaga
One-hit
b Highest dote group oaltted.
HuAer of
Dote Croupe
2
2
2
2
1
4
S
4
3
3
2
2
13
13
13
13
<*
6. 906-02
6.90C-02
6.90E-02
6.90C-02
0
0
0
0
0
0
0
0
2.61C-02
2.61E-02
0
2.4SE-01
qi
4.5K 01
4.5J£ 01
3.19C-02
3.19C-02
0
0
1.511*02
1.89E«02
9.15£»01
1.986*02
5.1U 01
5.1U-01
1.72
1.72
7.B7E-01
7.87E-01
NeJtleui L
92
3.766*04
3.476*04
1.396*04
0
0
Likelihood Peraeeter Ettleatee
«|3 ojt 71
ql-
*2 p-velue
0<<
-------
APPENDIX B
This appendix contains a table for each of the PAH
carcinogenesis bioassays evaluated in Sections III and IV of
this report that shows the doses of PAHs administered, numbers
of tumor-bearing animals observed, and the numbers of
tumor-bearing animals that were predicted from the relative
potencies derived using the multistage or one-hit models.
B-l
-------
Table B-l
Data Used to Estimate the Dose-Response R«lationship
Between Ingested B[o]P and Forestomoch Tumors
PAH
8[a]P°
Dose
(mg/kg/day)
0
0.13
1.30
3.90
5.20
5.85
Number of
Mice Exposed
289
25
2*
37
40
to
Number of
Observed
0
0
0
0
1
4
Tumor Bearing Animals
Predicted
Multistage One-Hit
0 0
<0.001 0.027
0.001 0.256
0.374 1.17
1 . 68 1 . 68
2.97 1.88
Abbreviations: B[a]P, benzo[a]pyrene.
Source: Neal and Rigdon (1967).
B-2
-------
Table B-2
Data Used To Estimate the Dose-Response
Relationship Between B[a]P and Respiratory Tract Tumors
PAH
B[a]P°
Exposure Rate
-------
Table B-3
Data Used to Estimate the Dose-Response Relationship
Between Injected (lung) B[a]P and Selected
PAH* and Epidermoid Carcinoma*
PAH
Exposure Levels
in mg
Number of
Rats Exposed
Number of Tumor Bearing Animals
Observed Predicted
Multistage One-Hit
B[o]P°
B[b]F
B[k]F
I NO
B[ghi]P
0
0.1
0.3
1.0
0
0.1
0.3
1.0
0
0.16
0.83
4.15
0
0.16
0.83
4.15
0
, 0.18
0.83
4.15
35
35
35
35
35
35
35
35
35
35
31
27
35
35
35
35
35
35
35
34
0
4
21
(b)
0
0
1
9
0
6
3
12
0
3
8
(b)
0
0
1
4
0
4
21
0
0.097
0.875
9.03
0
0.482
2.38
12.12
0
1.94
8.95
0
0.156
0.812
4.03
0
7.57
18.15
0
0.783
2.30
7.09
0
•> 0.723
3.18
11.29
0
1.94
8.95
0
0.167
0.860
3.98
0 Abbreviations: B[a]P, benzo[a]pyrene; B[b]F, benzo[b]fluorantnene;
B[k]F benzo[K]fluoranthene; ZNO, indeno(1,2.3-cd)pyrene;
B[ghi]P, benzo[ghi]perylene
b Highest dose group omitted from analysis because of lacK of fit.
Source: Deutch-Wenzel et al. (1983).
B-4
-------
Table B-4
Data Used to Estimate the Dose-Response Relationship Between
Skin Painted B[a]P and Selected PAHs and Squamous Cell Papilloma
PAH Total Initiating Number of
Dose (MO) Mice Exposed
B[b]F°
B[J]F
B[k]F
B[a]P
0
10
30
100
0
30
100
1000
0
30
100
1000
0
30
20
20
20
20
20
20
20
20
20
20
20
20
20
20
Number of Tumor Bearina Animals
Observed Predicted
Multistage One-Hit
0
9
12
(b)
0
6
11
(b)
0
1
3
15
0
17
0
6.47
13.80
0
4.77
11.93
0
0.955
3.01
16.09
0
17
0
6.47
13.80
0
4.77
11.93
N 0
0.955
3.01
16.09
0
17
a Abbreviations: B[b]F, benzo[b]fluoronthene; B[j]F, benzo[j]fluor-
anthene; B[k]F, benzo[k]fluoranthene; B[a]P, benzo[a]pyrene
b Highest dose group omitted from analysis because of lack of fit.
Source: LoVoie et al. (1982).
B-5
-------
Table B-5
Data Used to Estimate the Dose-Response Relationship Between
Skin-Painted B[a]P and Selected PAHs and Tumors
PAH
PAH Dose
(pg/onimol)
Number of
Mice Exposed
Number of Tumor Bearing Animals
Observed Predicted
Multistage One-Hit
B[a]Pa
B[b]F
B[J]F
CP
B[k]F
IP
0
1.7
2.8
4.6
0
3.4
5.6
9.2
0
3.4
5.6
9.2
0
1.7
6.8
27.2
0
3.4
5.6
9.2
0
3.4
5.6
9.2
80
34
35
36
80
38
34
37
80
38
35
38
80
34
35
38
80
39
38
38
80
36
37
37
0
8
24
(b)
0
2
5
(b)
0
1
1
2
0
0
0
3
0
1
0
0
0
1
0
0
0
10.48
22.12
0
2.11
4.89
0
0.774
1.17
2.06
0
0.001
0.044
2.96
0.410
0.200
0.195
0.195
0.421
0.189
0.195
0.195
0
13.23
19.46
0
2.88
^ 4.14
0
0.774
1.17
2.06
0
0.134
0.550
2.33
0.410
0.200
0.195
0.195
0.421
0.189
0.195
0.195
0 Abbreviation*: B[o]P, benzo[a]pyrene; B[b]F, benzo[b]fluoronthene;
B[J]F, benzo[J]fluoranthen«; CP, cyclopentadieno(ed)pyren«;
B[k]F benzo[k]fluoranthene; IP. indeno(1,2.3-cd)pyr«n«
b Highest dose group omitted from analysis because of lack of fit.
Source: Hobs et ol. (1980).
B-6
-------
Table B-6
Data Used to Estimate the Dose-Response Relationship Between
Injected (subcutaneous) B[o]P and D[ah]A and Sarcomas
PAH
Oose
Number of
Mice Exposed
Number of Tumor Scoring Animols
Observed Predicted
Wultistoqe One-Hit
B[o]PQ
D[oh]A
0
3.12
6.25
12.05
25.0
50.0
100.0
0
2.35
4.7
9.3
18.7
37.3
75.0
600
100
100
100
100
100
100
600
100
100
100
100
100
100
42
9
33
61
37
77
(b)
42
37
39
44
66
65
(b)
42.94
16.47
24.87
39.20
60.18
82.92
58.04
17.82
24.87
37.27
56.60
79.23
42.94
16.47
24.87
39.20
60.18
82.92
58.04
17.62
24.87
V37.27
56.60
79.23
a Abbreviations: B[a]P, benzo[a]pyrene; 0[ah]A, dibenz[oh]anthracene
b Highest dose group omitted from analysis because of lack of fit.
Source: Pfeiffer (1977).
B-7
-------
Table B-7
Doto used to Estimate the Dose-Response Relationship Between
Skin Painted B[a]P and B[a]A and Malignant Tumors
PAH
B[a]A
Concentration
(*)
Number of
Mice Exposed
0
0.002
0.02
0.2
1.0
20
32
18
52
29
Number of Tumor Bearing Animals
Observed Predicted
Multistage One-Hit
B[a]P°
0
0.00002
0.0002
0.002
0.02
20
18
21
20
12
0
0
0
0
5
0
<0.001
<0.001
0.001
5.00
0
0.008
0.091
0.81*7
4.22
0
0
1
3
B
0.287
0.473
0.335
1.80
6.01
0.287
0.473
0.335
1.80
6.01
Abbreviations: B[o]P. benzo[a]pyrene; B[a]A, benz[a]anthracene.
Source: Bingham and Folk (1969).
B-8
-------
Table B-8
Data Used to Estimate the Dose-Response Relationship between
Skin Painted B[o]P and Selected PAHs and Carcinomas
PAH
B[o]P°
CH
BCb]F
B[mno]F
Dose
(mo)
0
0.150
0
1
0
1
0
1
Number of
Mice Exposed
20
20
20
20
20
20
20
20
Number of
Observed
5
20
5
16
5
18
5
4
Tumor Bearing Animals
Predicted
Multistage One-Hit
S 5
20 20
S 5
16 16
5 5
18 18
S
4.3 4.5
4.5 4.5
0 Abbreviations: B[a]P, Benzo[o]pyrene; CH, Chrysene; B[b]F,
Benzo[b]fluoranthene; B[mno]F, Benzo[mno]fluoranthene.
Source: Van Duuren et al. (1966).
B-9
-------
Table B-9
Data Used to Estimate the Dose-Response Relationship
Between Skin Painted B[a]P, IP, and B[ghi] Tumors
PAH
B[a]P°
IP
B[ghi]P
Dose
0
0.05
0.1
0
0.01
0.05
0.1
0.5
0
0.05
0.1
Number of
Mice Exposed
100
20
20
100
20
20
20
20
100
20
20
Number of
Observed
0
17
19
0
0
0
6
(b)
0
1
0
Tumor Bearinq Animals
Predicted
Multistage One-Hit
0 0
16.40 16.40
19.35 19.35
0 0
0.006 0.411
0.760 1.97
5.33 3.75
0 0
0.333 0.333
0.661 0.661
0 Abbreviations: B[a]P. benzo[a]pyrene; IP, indeno(1,2,3-cd)pyrene;
B[ghi]P, benzi[ghi]perylene.
b Highest dose group omitted from analysis because of lack of fit.
Source: Hoffman and Wynder (1966).
B-10
-------
Table B-10
Data Used to Estimate the Dose-Response Relationship Between
Skin Painted B[a]P and Selected PAHs and Tumors
PAH
B[o]Pa
IP
B[ghi]P
Total Dose
(mg)
0
0.25
0
0.25
0
0.25
Number of
Mice Exposed
30
30
30
30
30
27
Number of
Observed
2
2k
2
5
2
2
Tumor Bearing Animals
Predicted
Multistage One-Hit
2 2
2* 2k
2 2
5 5
2 2
2 2
Abbreviations: B[a]P, benzo[a]pyrene; IP, indeno(1,2,3-cd)pyrene;
B[ghi]P. benzo[ghi]perylene
Source: Hoffman and Wynder (1966).
B-ll
-------
Table B-ll
Data Used to Estimate the Dose-Response Relationship Between
Skin Painted B[a]P and Selected PAHs and Carcinomas
PAH
Dose
(% concentration)
B[a]P°
D[ah]A
CH
0
0.001
0.005
0
0.001
0.01
0.1
0
1
Number of
Mice Exposed
20
29
30
20
20
20
20
20
20
Number of Tumor Bearing Animals
Observed
0
0
19
0
2
18
(b)
0
8
Predicted
Multistage One-Hit
0
1.07
18.29
0
2
18
0
8
0
4.07
10.08
0
3.59
17.24
0
8
0 Abbreviations: B[a]P, benzo[a]pyrene; D[ah]A. dibenz[ah]anthracene
CH. chrysene.
b Highest dose group omitted from analysis because of lack of fit.
Source: Wynder and Hoffman (1959).
B-12
-------
Table B-12
Data U«ed to Estimate the Dose-Response Relationship Between
Subcutaneously Injected B[a]P and Selected PAHs
and Spindle-Cell Carcinomas
PAH Dose
(mg)
B[a]P° 0
0.00195
0.0078
0.0156
0.031
0.062
0.125
0.25
.5
1.0
2.0
4.0
8.0
D[ah]A 0
0.00195
0.0078
0.0156
0.031
0.062
0.125
0.25
0.5
1.0
2.0
4.0
8.0
Number of
Mice Exposed
160
81
40
19
16
20
19
21
19
20
19
19
21
160
79
40
19
21
20
29
21
21
22
19
20
21
Number of
Observed
4
2
0
0
0
4
15
14
19
18
19
16
20
4
2
6
6
16
20
21
19
20
22
19
17
16
Tumor Bearing
Animals
Predicted
Multistage One-Hit
4.12
2.55
1.55
.981
1.22
2.49
4.08
7.70
11.18
18.52
18.41
18.98
21.00
34.73
17.24
8.87
4.31
4.95
5.09
6.68
7.49
9.91
14.16
15.92
19.33
20.97
4.12
2.35
1.55
.981
1.22
2.49
4.08
7.70
11.18
16.52
18.41
18.98
21.00
34.73
17.24
8.87
4.31
4.95
5.09
6.68
7.49
9.91
14.16
15.92
19.33
20.97
a Abbreviations: B[a]P, benzo[a]pyrene; 0[ah]A. dibenz[ah]anthracene
Source: Bryan and Shimkin (1943).
B-13
-------
* I V 1*11 Sl.'-Rnu
rw n •_»
Ii kn >v»>•* »»•) (/• —vinj »EU>
P»» (
fedflp v
C«pl>- f«il f>r' «0U
nod (r>X-/i«9 MftUl
Mlnr-y MCU
nth
h/t
Owl* rod pevuinj »cu
act* *ul prrv*t/inj tBU
• t'^nnl
m«l>/- of HM in ilt
!tiiU/> ./rl 0»y if!
t>f rl '.l«y t'.i
CliyUt
Hm dv UB
i Ojyl,.i
N/n 1.1 ^ Lb«
K«J (f.TVilt] M>U4 0/»l I
MKW/MJU Ivkr >»»V I-
Ii.k*- SnV !•••
fr/rl Cl^ U»
(k/ r.l Cl v U- •
flK '
fin ^«
'jW» l.*W
nn '>Htf i
l.i
I.I
1.1
I.I
I.I
•;• to 4 Kr: 19 .'; r
•:• i-.' * trj ii ;•?"
ul'. • p. .<) (
•i i» u/: r
7U 4|.r;
•: l.< -4 bri
I Kr
•I (.»
it
l< .'.' I. If:
ii r c o./1;
i< ;•? c i. R
K ; r IP.
aw,
fCiiii.) tin;
•; lo 4Ur> M-.7C l.'f.
JC'ut.l
7 u -4 l^j !>•;? C O.A
^4.^:(irt.) KO«
tfl I c^>. Hi C
7 U 4 t-rj 117.' C l.t
I K» K7:C n.s;
;4'iC.it.) C. TV,
ilVMK
', t ''f. ')'.
s ii
'( .'I'!' 'l.llllj I/.HY
.« i fn p.iiift i,^
10 vi''1)
10
!inr
K) :
'-I "
II. 1
:M
un.?
rO HI; O.'IU) I,'J)Y
t/,.4
Jll.S
in.b
n.i'j
b.l
l.S
'/It O.illl) l.t«lY
tf: ll.'VK l/tDY
M7 n.ur i/twY
7ii o.rtn i/toY
iv ».un i/tst
711 O.HD l/t«iY
I* l'.U>4 \,\ft1
\n o.fw i/t«Y
I.T) o.ii) i/t«
7 4) O.OU) |/tNY
'?. 4| O.UI6I I/On
iio: 4i ii.ota i/tnv
«.7 41 O.OI6B 1/W
•X) i> ').!>/» U.S. (H. I- .'I
i MO .iiki u.s. !?i 11. ;j
I .-^i u^ u.s. ir« 11.7)
i io> iv» f-S. (n> in "i
I
I
I .fO
I
I
I
i-c ys
1C 180
1C
rl r rl |l. (
H.F.I. (iVi
U.S. [»* (l'«.7|
tr-tw -t ii. I
Sir.
«.r.i.
(jriw •! il. (IV- •
l>'«J>».y«l i 'lUlp I ''-I •
U.S. (I* tll-7)
*.M. (1*7,
vt
U.S.
U.S. tWll'.V)
».l.l.
i> \*n
i (.K l'«v)
tWlnriSrUll (H/"
Cuvcr & Sir. (1<£7|
I
o c :«j
OC 740
of.yn (dv») c
OC W> ( I-*/)
ocs,^
I) C IJO
-------
\/
rue ^»cv i
. Cljy lea
t/fC/>U *<.( (rowing M:U Flnf rjl»O» IM
y..«t oil o»;'•.!•»» $oii/«ii a *t u.
fouil
Cliyl-Mi
t>r*li
0 v l'-
iUrt ,1 f.'« m
•••jrs «/ l'-«4 in
. i'.i/» • .( r'H in
.W-.-( fM in
fK -WCVl'*l
r;<
f.l
• i',
IV.'.
u',
•2 la -4 hri Ii. .T r I.
r, .«.
» : KtU •:*. 18 -.?
i.l IV H
( NWU V I) W
I' ft
1S-A
rtjs
t.l IV .^
•2 ta 4 bjrt UK? I U.;
•1 br W-C C «
6 I*N« SI) Ui
•Itr
U2C
•2U -4br» I? 2C O.n
i'.;',
•J U -4 br, K..7 c | (8
i IS
is :*
•2 u 4 |^: |i /2l \£
•? to 4 Lri K tfC !<:
•it* is;?( 17:
fiD tup.
.VI
M t .T.
.vc
Illf.
in i!
/.«
14 V'-1'1)
w 1'"
O.'M
i;
G.IJM
17
u
IM.2
i\
:??
fi.l
i'.i
12.)
I.M
M
s&
M
14
.'*>
1.4
) 4
14
W.'.
V.I
l-i.S
•4
IU»
O.ifll.5
M.S
UiS
n.s
i
O.MS
I
s
lu.d
M.!
in ?
.Vi
l.u
IM..'
(.••« U.u'-in l/t»r
M» n.Jil/ l/U*
*t u.'UI/ !/O>»
r'i o.m;j i^y
:•" . o.nc i.-t»<
4.T) fl.in? i,t»at
r* fl.ur* i,^
:4) o.uxs i/o«
>4 O.OIO 1/tBY
."0* O.OU4 ^WY
I<"S O.fbil 1/tnv
l«^ 0.ftJ4 l;t»«
147 G.V
O.UIT
.'.fl 0.iC» i
1C Ur ^ tt il. (1^1 1 «
1C «J^.«J «l i.i. P> (14V)
«t ii. ti''*D •
-l jl.
1C f>riw cl jl.
i r to (^i) M*: (ji rw)
1 (WcHr «l il. (IV)) •
1C VS (i(Vfi) BfloC (* •/«/)
1C tr*w cl il. (W-Nl •
| Crtw .1 jl. (19.^1 •
1C r.r Hr M *l. (lW«i «
1 C*>w ?l il. (IV>. •
1 XI l»/i U.S. 0* (I
| |ij) d^ U.S. CM (I
1C
1C
1C
IS
I
i:
1C
1C
1C
1C
i 'i
i .•<
I I'C4v-
I
0 «1) (Jv.)
I
1C
I
I
I
If
1C
I C
0 *UUvsl
o c:
o c:
o c;«(dv-i
OC VSl-Jvj)
U t S/-, (.Jw)
U.S. 0* (IW)
Sin (I'M)
S>a (ISfitj
•At«d (I il (|V|) •
M.S. W(l*/)
rt )l (IVI) •
:\. t,id rt jl (IVII •
''f ?-rxft (IAS)
'y».1 5 ',IM (\%>\
('.».«.•» S :>
C'l-i; (.H 1 »I7|
1. la. l> Hi/)
-------
KDifl
Cr»o:ol< m.'-t pT'
IHIJUK:
mam* s
«M
(fl -^tritc- Slui)t Mr.U
FCF K«yl (Starving »j;U
Gr*>wl« *»•• prufvioj »nU
W »cd pr» /rving »uU
»nt<
fVrt <••
f«j»l fuel «>:U
iO fuel »t.U nth
Hi ittrc of fVM in soil:
Ki>t>r* of PiM in
of M in
•<•!
(Vr-rt Cl >y lo
Oirjnl Cl«y
flit V»CY
Dumt llay
ftrrt
SAWVLOW
SWVLCW1
Hc£o11'« fine
FlIC r
FDC S
6.1
6.1
6.1
6.1
6.1
6.1
up.
-2 to 4 ur-.
•1 bv
•1 Kar
~'i U> -4 b*i
•2 to -4Kr;
•lb»
i>:-.? f;
'j-a
13-22 C
I'M? C
v:o
18-22 C
18 -# C
K-A1 C
1.0*:
10 q/UJ>j
*:
o.r;
10 .]/1(nJ
0.75;
1.35
1.>.
•2 lo -4
18-22 C 0.7?
7.9 KHOUlC^p. X) C
7.9 SlHDKCf. IOC
7.9 60-10) I C-v- * C
M 5 fi»ld cip. 18 22
6ftUft S/)
6WHH Vll
1.8 1AO
IV
no
17
u:
S3
;M
H
n..i
o.v
o.n
4t>
21
S3
1.2
13
21
9.94
9.94
9.94
106.2
167
112
(C5
f.30'
MS
PCS
270
7JO
ISt)
ID
173
93
(6
(6
73
67
y
ft
CIO
uo
vo
\w
1210
X7
O.OTOJ I/WY
O.ODII 1/TWY
O.OCt2 1/WY
O.IIC4 1/WY
O.OtW 1/OfiY
O.ftJM 1/W
O.UH4 1/OflY
O.OT63 1/WY
0.1X66 1/trtY
0.0071 I/WY
0.1/C 1/WY
O.UtK? 1/WY
0.«* I/WY
0.01'D 1/WY
0.0103 1/WY
0.0107 1/WY
0.0081 iqAij
O.Ott •I'Vii
0.01X mtfuL
0.044) »^
O.ODU i^Vq
0.1W) wAq
l.ttdv- U.S. CPU (1987)
H f.75 (dv-) W* (J< W)
lUO.lv- "•'>• EPfl (1W7)
Kfldv- U.S. D>ft( 1087)
.03
1C
1C
-C
•C
-C
C
•C
-C
0-C 240 (dv-)
0-C 240
0 C 2C
0-C toO (dv-1
0 C W
OCS7S
U.S. tffl(lW)
U.'>. EPA (IWj
U.S. 0-A(lJb7)
Sin (1132)
Si*i (1932)
Sin (19B2)
Sir. (VfQ)
Sin (1%6)
Sin (19B2)
U.S. ETB(IW)
Sin (19U)
Sin (1XP)
SSir.
; SIK (iw)
W 1987)
X IW)
-------
ICDIA
Sfliir«
PCP »»»J -ft tztrvin)
Crt-KM* »U »nd pretervinj wste
PCP «f.J («••!<
Slqp Oil (->te »<«J preil>« foiiil fuel «cU «i(h
Rcfintry »cU
Crttcot* »<>d prturvin »»U
fine t
Ittrtir* of PPM in >jils
bf rfW in Vilj
of PfiH in vlir< of PW in unh
NiiUr* of PflH in j.>ilj
C/K&fle •• (Ml.)
*
-2to-4Sr,
J*(at.)
24.K (vil.j
80 i f.?ld <
8) * c^>.
7.9
18 2 C
18-KC
I622C
W
0
K
18-22 C 1
'
•'J fr*vr-/if»j
IVfollt* fine
13-22
15-25
1525
JK
ac
IOC
20C
»{.
80 ': field c^. IP .1'
C Dffi
10^100,
INI if W.
4.1
21.?
4]
%
26
210
5.3
67.S
47
171
3.1
27
lfi.5
3V
27
73
V>
27
40
106.2
50)
5
5
Aft
nu
so
,f
JS7
1ft). 2
wo
irui
(.71)
410
)X
.rw
ioj
161
IX)
116
09
•fi
78
77
77
71
70
'A
5i
42
41
40
10.5
5.5
3?3
KID
171.
470
'ri
'WMIS
NA 1/trtY
O.OOD1 1/WY
O.ttn? i/fay
o.orrt I/WY
o.mi I/WY
O.GU17 1/WY
o.fa-i i/wv
U.'IC4 1/WY
O.IUK 1/WY
O.lfl4) 1/WY
O.ClW 1/WY
O.OD6 1/WY
0.007 1/WY
0.0071 1/WY
o.aw I/WY
0.009 1/WY
O.tfD 1/WY
O.OS6 I/WY
G.OM9 I/WY
0.012 1/WY
0.0132 1/WY
0.01&5 1/WY
0.017 1/WY
0.0171 1/WY
O.OtGO 1/WY
o.i;toi/wY
0.0031 1/WY • R
0.101)!
o.iait)
0.1179 I/WY
O.C70»v1
O.W
0.37-P
,.t
100 djv-:
280 d«^
2a)d*/i
280 dv-
230 d>/:
100 d»/:
2SO.lv:
100 dv
100 d-
'•75 (day i)
JBOdiy:
1-C 575 (!>/»)
1-C
1
1 10) dv:
1C
1
1-C 180
1-C
1-C
0400
OC240
0-C 240
0 400
0-C 240
Of. ;,75
0-C f,75
ii c m
(197r~.)
U.S. tf«(lW)
U.S. EW(1W)
U.S. [PMIW)
U.S. fj« (1137)
U.S. EM(W)
U.S. tM(l«7)
U.S. £W(lcrt (19B5)
S Sin (11J7)
i-e Reprt (\W>)
j^r SSi§3 (
Befrf (M 1W)
Peltr (J< 1»7)
Re W (X IG07)
-------
»GIA amnTf 3)iipH (Win «r>n*£ rpfwtf IW.WFF. IHIIUH: t^ain crtcnci WIMIS RDtima nwrux
Cr*;*.te «c. I (reviving ml* Du-jnt Cljy U» -2 to -4 t-ri W ?2 C I.E 1.4 2CUJ 0.0.103 1/tflY 1 280 fcyj U.S. tPfl (1W)
FINE 'MH l«lh 6 HfH*f 5-JO 10 ^/KMj 5 1821 0.0n04 1/fHY 1C V5 (dxri) Kcf-C (X !98/)
ft/ant C1.vU« -Ib* 1?22C O.K 2 iAJ O.U313 1/WY 1 10) v«*l 4-^ '4t O.OD49 1/WY 1-C Sir,'(r»2)
I re o-pu^J y \fl O.CD5J I/DAY 1C Sin (1*2)
Kre (-v"^ ' >47 U'J O.OOM 1/WY 1-C Sin (\V2)
br< cMfxxol 93 99 0.(T)70 I/WY 1-C Sin (IT82)
G.wp1
-------
tG'IA
Cr«is.>U rod prewrvinj
"efinrry
Winery *j:l«
"PI "i(«r«tor Sluij.
"•firvry
IVi-anl C1»y Im
ft/«l Clay lo*
Grant Clay l->»
Kdun Srty lo*
Grant Clay IMI
lidun >«ndy [a*
>Cliylo4i
> Sand/ LOM
**»» Clay lav
Kidun S«>iy loji
liiwnjylaji
8.1
8.1
8.1
8.1
8.1
8.1
•2 k -4 Ur j K-2? C
•? lo 4 fnri 13 ?2C l.r:
?U) -4UTJ IJ.Tlf , (j
•2 lo 4 Lrj W-fl C 0 /'<
W.«(Mt.) aw
•|b» IS-L'C l.K
».S (Ml.) 2j w
K(nl.) -nfuj-
C.IVZ
'UW WLrllfT BTCWT WOiirs
°-; 130 O.OCI |^y
0-^ I/O 0.0)41 i/tnv
O-1) 170 O.OTM2 l/ftw
9-l IB 0.006 1/tHY
«-5 21) 0.054 1/ttY
5-6 « 0.007 VtWY
5-? W 0.01 I/WY
;-; 63 0.011 1/ttY
8-5 63 0.011 l/lOY
" 62 0.0112 1/OAY
>5-5 M O.OL^ l/ttv
DJPlliM
?BO d»,i n.s.
190 J^ U.S.
2* 'bya U.S.
A.P.I. (1987)
lll'dayj M.S. [?A (I'.W)
A.P.I. (K«7)
A.P.I. (1*7)
A.P.I. (19S7)
A.P.I.
TO 'lays U.S. EPfl
A.P.I.
-------
REFERENCES
Bauer, James E., Douglas G. Capone. 1985. Degradation and
Mineralization of the Polycyclic Hydrocarbons Anthracene and
Naphthalene in Intertidal Marine Sediments. Applied and
Environmental Microbiology, 50(1):81-90.
Bossert, Ingeborg, Wayne M. Kachel, Richard Bartha.
of Hydrocarbons During Oily Sludge Disposal in Soil.
Environmental Microbiology. 47(4):763-767.
1984. Fate
Applied and
Bulman, T. L. , S. Lesage, P. J. A. Fowlie, M.D. Webber. 1985.
The Persistence of Polynuclear Aromatic Hydrocarbons in Soil.
Pace Report No. 85-2.
Calocerinos & Spina. 1984
Investigations, Step 2. August
Calocerinos & Spina. 1985
Investigations, Step 3. March
Calocerinos £> Spina. 1984.
Investigations, Step 1. March
Harbor Point
Harbor Point
Harbor Point
Chiou, C. T. , Schmedding, D. W., Manes, M. 1982,
Organic Compounds in Octonol-Water Systems.
Technol., 16(1):4-9
Property Land
Property Land
Property Land
Partitioning of
Environ. Sci.
Coover, Mervin P., and Ronald C. Sims. 1987. The Effect of
Temperature on Polycyclic Aromatic Hydrocarbon Persistence in an
Unacclimated Agricultural Soil. Hazardous Waste & Hazardous
Materials, 4(1):69-81.
Dibble, J. T., R. Bartha. 1979. Effect of Environmental
Parameters on the Biodegradation of Oil Sludge. Applied and
Environmental Microbiology. 37(4):729-739.
Dzombak, 0. A., and R. G. Luthy. 1984.
Polycyclic aromatic Hydrocarbons on
137(5):293-306
Estimating Adsorption of
Soils. Soil Scienc—
Edison Electric Institute. September, 1984.- Handbook on
Manufactured Gas Plant Sites. Utility Solid Waste Activities
Group, Superfund Committee.
Environmental Research and Technology. June, 1983. Land Treatment
Practices in the Petroleum Industry. The American Petroleum
Institute.
Fu, J. K., and" R. G. Luthy. 1986. Effect of Organic Solvent on
Sorption of Aromatic Solutes onto Soils. J. Env. Engrng. ASCE.
112(2):346-364
-------
Gardner, W. s., R. F. Lee, K. R. Tenore, and L. W. Smith. 1979.
Degradation of Selected Polycyclic Aromatic Hydrocarbons in
Coastal Sediments: Importance of Microbes and Plychaete Worms.
Water, Air, and Soil Pollution II, 339.
Groenewegen, D., and H. Stolp. 1976. Microbial Breakdown of
Polycyclic Aromatic Hydrocarbons. Tbl. Bakt. Hyg. I. Abt: Orig.
B162, 225.
Haug, R. T., and Tortorici, L. D. 1986. Composting Process Design
Criteria, Part 4. BioCycle. Nov/Dec:34-39
Herbes, S. E. and L. R. Schwall. 1978. Microbial Transformations
of Polycyclic Aromatic Hydrocarbons in Pristine and Petroleum-
Contaminated Sediments. Applied Environ. Microbiol. 35: 306.
Herbes, S. E. 1981. Rates of Microbial Transformations of
Polycyclic Aromatic Hydrocarbons in Waters and Sediments in the
Vicinity of a Coal-Coking Wastewater Discharge. Applied Environ.
Microbiol. 41, 20.
JTC Environmental Consultants. 1984. Factors -Affecting Odor
Release During Sewage-Sludge Composting. Washington Suburban
Sanitary Commission. January.
Kuter, G. A., Hoitink, H. A. J., Rossman, L. A. 1985. Effects of
Aeration and Temperature of Composting of Municipal Sludge in a
Full-Scale Vessil System. Journal, WPCF. 57(4):309-315
Leo, A., Hansch, C. , Elkens, D. 1971. Partition Coefficients and
Their Uses. Chemical Reviews. 71(6):525-616
Loehr, R. C. , and Jacques, R. B. 1987. In-Place Detoxification of
Contaminated Soils. In Print.
Loehr, R. C. , Ronald Sims. 1987. The Land Treatability of
Appendix VIII Constituents Present in Petroleum -Refinery Wastes:
Laboratory and Modeling Studies.
Mackay, D., and W. Y. Shiu. 1977. Aquous Solubility of Polynuclear
Aromatic Hydrocarbons. J. Chen. Eng. Data. 22(4):399-402
Mahmood, R. J., and Sims, R. C. 1986. Mobility of Organics in Land
Treatment Systens. 112 (2):236-243
McCoy, L. H., Ed. 1986. Determining the Distribution of Organic
Contaminants Between Air, Water, and Soil. The Hazardous Waste
Consultant. Jan./Feb: 1.20-1.24
McCoy, L. H., Ed. 1985. Mobility and Degradation of Common
solvents in Ground Water. The Hazardous Waste Consultant.
May/June: 1.21-1.23
-------
Medvedev, V. A., and V.' D. Davidov. 1972. Transformation o|
Individual Organic Products of the Coke Industry in Chernozemic-
Soils. Pochvovedenic 11(22).
• Miller, F. C. and Finstein, M. S. 1985. Materials Balance in the
Composting of Wastewater Sludge as Affected by Process Control
Strategy. Journal, WPCF. 57 (2):122-127
Parker, L. V., and Jenkens, T. F. 1985. Removal of Trace-Level
Organics by Slow-Rate Land Treatment. Water Res. 20(11):1417-1426
Reid, R. C., Prausnitz, J. M., Sherwood, T. K. 1985. The
Properties of Gases and Liquids. McGraw-Hill, New Jersey, pp.
629-677
Poglazova, M.N., G. E. Fedoseeva, A. J. Khesina, M. N. Meissel,
and L. M. Shabad. 1967. - Further. Investigations of the
Decompostion of Ben(a)pyrene by Soil Bacteria. Doklady Akademii
Nauk SSSR. 176: 1165.
Remediation Technologies, Inc. 1987. Technology Description of
Land Treatment. Prepared for: The Gas Research Institute.
N
Schmidt, J. W. , Simovic, L., Shannon E. 1981. Natural Degradation
of Cyanides in Gold Milling Effluents. Presentation at the Gold
Mining Industry Seminar
Shabad, L. M. : A Ya Khesina, H. P. Schubak, and G. A. Smirnov
1969. Carcinogenic Hydrocarbon Benzo(a)pyrene in the -Soil. J.
Nat. Cancer Inst. 47: 1179.
Simovic, L. , and Snodgras, W. J. 1985. Nateral Removal of Cyanic.s
in Gold Milling Effluents - Evaluation of Removal Kenitics. Wat_r
Poll. Res. J. Canada. 20(2):120-134
Sims, R., et.al. 1984. Review of in-Place Treatment Techniques for
Contaminated Surface Soils - Vol. 2. NTIS. EPA-540/2-84-003b
Utah Water Research Laboratory. Permit Guidance Manual on
Hazardous Waste Land Treatment Demonstrations, November.
Sims, R. , and Bass, J. 1984. Review of in-Place Treatr.._nt
Techniques for Contaminated Surface Soils - Vol. 1. NTIS.
EPA-540/2-84-003a
Sims, R. C., M. R. Overcash. 1983. Fate of polynuclear aromatic
compounds (PNAs) in soil-plant systems. Residu Reviews, 88: 1-68.
Sims, R. C., J. L. Sims, D. L. Sorensen, W. J. Doucette, and L. T.
Hastings. 1987. Waste-Soil Treatability Studies for Four Complex
Industrial Wastes: Methodologies and Results, Volumes 1 and 2.
Sisler, F. D., and C. E. Zobell. 1947. Microbial Utilization of
Carcinogenic Hydrocarbons. Science 106: 521.
Witherow, J. L., and Bledsoe, J. L. 1286. Design Model for the
Overland Flow Process. Journal WPCF. 58 (5):318-386
-------
ATTACHMENT D
-------
EQUATIONS USED TO CALCULATE RISK LEVELS
Model Used to Estimate Target Levels of Carcinogenic Pah in Soil;
Model Incorporates Decay of PAH in Soil
Let:
C(t) - cone in soil at time t (ppm)
C(0) = cone in soil at t = 0
t = time measured in years
Shalf = half life of decay in surface soils in years
T = period of exposure in years starting at t = 0
C(T) - average concentration over exposure interval
from 0 < t < T.
Then:
(t/Shalf)
C(t) = C(0)*0.5
C(t) = C(0) * exp {-(0.693*t)/Shalf}
Over interval 0 < t < T, the average concentration
C(T) = C(0)/T * Int exp {-(0.693*t)/Shalf)dt
0—>T
After integration and simplification:
C(T) - {C(0)*Shalf/(T*0.693)) [1- exp(-0.693*T/Shalf)]
C(T)/C(0) - FRAC= (1/T * Shalf/0.693) [l-exp(-0.693T/Shalf)]
then:
C(0) =» C(T)/FRAC
-------
FOR SIMPLE SCENARIOS
A. Exposure Estimate
Effective ADD(life) = dirt * conc/1.OE+6 * 1/int
* dur/life * matrix
where:
dirt = arot of dirt ingested per event (mg)
cone =* concentration of chemical in dirt (ppm)
int = interval between events (e.g.,30 ** once every 30 days)
dur =» duration over which exposure takes place (yrs)
life = (lifetime of average person =» 70 years)
matrix = matrix effect associated with chemical being
administered within a soil matrix as compared to the
method by which chemical was administered in the
experimental tests used to determine toxicity end points.
In the present set of calculations the conservative
assumption of matrix » 1 has been made.
B. Lifetime Risk Calculation
. Risk = (ADD(life.) * Wt. Avg. CPF)/ body wt
where:
Wt. Avg. CPF was estimated based on the observed relative
concentrations of the carcinogenic PAH compounds at the
site and the relative potencies of these compounds via the
oral intake route. The weighted average potency was
estimated to be 1.6661 (mg/kg/day)-l.
Body Wt « average body weight of exposure group
C. Target Level Calculation: Set level equal to 1.0 E-6 (one in
one million)
Cone - (1.0 E+6 * 1.0 E-6) * Int * life * body wt
dirt * dur * matrix * Wt. Avg. CPF
Here Cone is calculated as the average concentration of
chemical in soil that would yield an incremental lifetime
risk level of one in one million.
-------
B'ecause the concentration of PAH will decrease with time due to
biodegradation, we want to calculate the initial concentration in
soil C(0) that would result in an average concentration C(T)
where this is set equal to Cone. Recall that:
C(0) = C(T)/FRAC
C(0) - Conc/{(l/T * Shalf/0.693) [l-exp(-0.693T/Shalf)]}
More Complicated Scenarios: 1 through 70 years - unrealistic
worst-case scenario
Total lifetime risk » Risk (1-6 yrs) + Risk (6-11 yrs) +
Risk (11-70 yrs)
Perform these operations for l.OE-6 target risk as follows:
cone = {[(1.0E+6*1.0E-6)*life]/(Wt.Avg. CPF * matrix)} * ^
{[(intA * wtA)/(dirtA*durA)] +(intB * wtB)/(dirtB*durB)] +
[(intC * wtC)/(dirtC*durC)]}
where A refers to 1-6-year old period
B refers to 7-11 year old period
C refers to 12-70 year old period
Caveat: for this time varying case, the biodegradation rate
should be incorporated mathematically as an integral. The
simplification used here introduces a relatively small
error.
-------
REMEDIATION
TECHNOLOGIES INC
224- ? ~2-s A.9 S
Ke--.. .'-A 5SC32
v'2C6; 372-C2--
Dr. Russell D. Walker
Florida DER
2600 Blair Stone Road
Tallahassee, FL 32399-2400
Dear Dr. Walker and Mr. DeAngelo:
March 23, 1988
Mr. Tony DeAngelo
U.S. EPA, Region IV
Waste Management Division
345 Courtland Street
Atlanta, GA 30365
At a meeting I had with Dr. Walker on March 15, 1988, Dr.
Walker requested that we explicitly state the assumptions used in
modeling soil ingestion risks presented in my memorandum titled
"Action Levels for the Live Oak Site", dated February 17, 1988.
The following assumptions were used in developing the risk
factors presented in Table 5 of that memorandum:
Carcinogenic
Factor
Potency 1.4713 (mg/kg/day)
-1
Average Soil
Concentration at Time 0
Matrix Factor for Soil
Ingestion
Frequency of Ingestion
Degradation half-life
Amount of Soil
Ingested/Event
55 ppm and 7.6 ppm respectively for
the treatment area and the site area
based on a composite of the upper two
feet of soil.
0.3
Once every three days
0.5 years
100 mg for ages 0-5, 50 mg for ages 6-
11 and 5 mg for ages 11-70.
The basis for these
memorandum. Dr. Walker
residential exposure risks
Amount of Soil
Ingested/Event
Degradation Half-Life
assumptions- were identified in the
requested that we also model the
using the following assumptions:
100 mg for ages
ages 11-70.
0.5 years, 1 year
0-11 and 25 mg for
and 1.5 years.
Concord. Massachusetts - Pittsburgh. Pennsylvania - Fort Collins. Colorado — Austin. Texas
-------
These results are presented in Table 1. Changing the amount
of soil ingested from ages 6-70 does not change the results
originally presented in Table 5. Increasing the half-life
increases the risk only marginally and is still within 10~6 for
the treatment area and 10~7 for the site area. This sensitivity
analysis demonstrates that a treatment goal of 100 ppm of carcin-
ogenic PAH is protective of human health and is an appropriate
criteria for the site.
It should be noted that the risk levels presented in Table 1
are for unrestricted site development and for surficial soils.
Therefore, the requirements in the ROD which pertain to: (a) deed
restrictions and (b) soil cover on the treatment area are not
necessary components of the final remedy.
I trust the above information satisfactorily addresses your
question regarding the proposed action levels. Should you
require any further clarification please do not hesitate to give
me a call.
Sincerely,
John Ryan ,
Principal ^
>' •
-------
TABLE 1
LIFETIME SOIL INGESTION RISKS
ASSOCIATED WITH UNRESTRICTED RESIDENTIAL DEVELOPMENT
AFTER TREATMENT OF CONTAMINATED SOILS
TO 100 PPM TOTAL PAH CARCINOGENS3
HALF LIFE SOIL INGESTION/EVENT RISK LEVEL
(years) (ing/event) Treatment Area Site Areab
0.5 100 mg for ages 0-5, 1.0 x 10~6 1.4 x 10~7
50 mg for ages 6-11
and 5 ng for ages 12-
70
0.5 100 mg for ages 0-11, 1.0 x 10~6 1.4 x 10~7
25 mg for ages 12-70
1.0 100 mg for ages 0-11, 1.5 x 10~6 2.1 x 10"7
25 mg for ages 12-70
1.5 100 mg for ages 0-11, 2.0 x 10~6 - 2.8 x 10~7
25 mg for ages 12-70
a) Note: The residential scenario assumes that houses are constructed
on the site immediately after treatment is completed and that a
person spends his entire life on this site (from birth to 70 years).
This scenario is extremely conservative and highly unlikely given
the demographics of the area.
b) Weighted average of wood storage area, plant area, treatment area
and the overall site.
-------
REMEDIATION
TECHNOLOGIES INC
•<
-------
Please let me know if you have any additional questions,
Best regards,
John Ryan
Principal
JR:ct
' »
cc: T. D«Ang«lov
K. Burke
K. Paulsen
S. Hugenberg
J. Rodes
-------
BIOREMEDIATION OF CONTAMINATION
BY HEAVY ORGANICS
AT A
WOOD PRESERVING PLANT SITE
By
Ronald J. Linkenheil
Remediation Technologies, Inc.
Fort Collins, Colorado
Thomas J. Patnode
Glacier Park Company
Seattle, Washington
-------
BIOREMEDIATION OF CONTAMINATION BY HEAVY ORGANICS
AT A WOOD PRESERVING PLANT SITE
Ronald J. Linkenheil
Remediation Technologies, Inc.
Fort Collins, Colorado
Thomas J. Patnode
Glacier Park Company
Seattle, Washington
ABSTRACT
On-site treatment was chosen as the closure alternative for
a creosote impoundment at a Superfund site in Minnesota. This
. • « , '
alternative was identified in the feasibility study as the most
cost effective source control measure for the site. The
effectiveness of using .land treatment technology to detoxify
contaminated soils at the site was demonstrated in pilot, scale
studies. Results of these studies were used to develop design
criteria for a full scale treatment facility.
A lined 3-acre treatment facility was constructed in 1985 to
treat 10,000 c.y. of contaminated soils 'and sludges from the
creosote impoundment. The facility has been .successfully operated
by ReTec since 1986 achieving greater than 90 percent reduction of
polynuclear aromatic hydrocarbons (PNAs) during the first year of
operation. This paper summarizes results from the first year of
treatment and demonstrates the effectiveness of the full scale
system. Aspects of construction and start-up of the full scale
facility are also reviewed.
INTRODUCTION
Wastewaters from a creosote wood preserving operation have
been sent to a shallow, unlined surface impoundment for disposal
-------
since the 1930's. The discharge of wastewater to -the disposal pond
generated a sludge which is a listed hazardous waste under the
Resource Conservation and Recovery Act (RCRA). Due to groundwata.
contamination of the shallow aquifer at the site by PNAs, the
State of Minnesota nominated the site for listing on the Superfund
National Priorities List in 1982. Since 1982 numerous remedial
investigation activities have been undertaken to determine the
nature and extent of contamination at the site. Based on the
results of these studies and extensive negotiations, the Minnesota
Pollution Control Agency (MPCA), the U.S. Environmental Protection
Agency (EPA), and the owner of the facility signed a Consent Order
in March 1985 specifying actions to be taken at the site.
In general terms, the remedial actions selected by the MPCA
and EPA involve a combination of off-site control measures and
.source control measures. The off-site controls involve a series of
gradient control wells to capture contaminated ground water. Th'
source control measures include on-site biological treatment of the
sludges-and contaminated soils and capping of residual contaminants
located at depths greater than 5 feet. Costs for on-site treatment
and capping were estimated to be $59/ton.
PILOT SCALE STUDIES
Before the on-site treatment alternative was implemented,
bench scale and pilot scale studies were conducted to define
operating and design parameters for the full scale facility.
Several performance, operating, and design parameters were
evaluated in the land treatment studies. These included:
-------
o Soil characteristics;
o Climate;
o Treatment supplements;
o Reduction of gross organics and PAH compounds;
o Toxicity reduction;
o Effect of initial loading rate;
Effect of reapplication;
Three different loading rates were evaluated in the test
plot studies: 2 percent, 5 percent, and 10 percent BE
hydrocarbons. The soils used in the pilot study consisted of a
fine sand which was collected from the upper 2 feet of the RCRA
impoundment. The soil was contaminated with creosote constituents
consisting primarily of PNA compounds. Total PNAs in the soil
ranged from 1000 to 10,000 ppm, and BE hydrocarbons in the
contaminated soil ranged from approximately 2 to 10 percent by
weight.
Because the natural soils are fine sands and" extremely
permeable, it was decided that the full scale system would include
a liner and leachate collection system to prevent possible leachate
break through. To simulate the proposed full scale conditions, the
pilot studies consisted of five lined, 50 foot square test plots
with leachate collection. The studies were designed to maintain
soil conditions which promote the degradation of hydrocarbons.
These conditions included:
o Maintain a pH of 6.0 to 7.0 in the soil treatment zone;
o Maintain soil carbon to nitrogen ratios between 50:1 and
25:1; and
o Maintain soil moisture near field capacity.
Hydrocarbon losses in the test plots were measured using
benzene as the extraction solvent. The analysis of BE hydrocarbons
provides a general parameter which is well suited to wastes
-------
containing high molecular weight aroraatics such as creosote wastes
Reductions of BE hydrocarbons were fairly similar between all the
field plots. Average removals for all field plots over four mont^
were approximately 40% with a corresponding first order kinetTC
constant (k) of 0.004/day.
The reduction of PNA constituents was monitored by measuring
decreases in 16 PNA compounds. The following compounds were
monitored in the test plots:
2 Rings 3 Rings 4. 5. and 6 Rings
Naphthalene Fluorene Fluoranthene
Acenaphthylene Phenanthrene Pyrene
Acenaphthene Anthracene Benzo(a)anthracene
Chrysene
Benzo(j)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene
Dibenzo(a,h)anthracene
Benzo(g,h,i)perylene
Indeno(l,2,3,c,d)pyrene
Greater than 62 percent removals of PAHs were achieved in_
all the test plots and laboratory reactors over a four montn--
period. PAH removals for each ring class are shown below:
o 2 ring PAH: 80-90 percent
o 3 ring PAH: 82-93 percent
o 4+ ring PAH: 21-60 percent
o Total PAH: 62-80 percent
Table 1 summarizes first order rate constants and half-life
data for BE hydrocarbons and PNA compounds for the 5 and 10 percent
BE hydrocarbon test plots. With the exception of the 4 and 5 ring
PNAs, the table shows that the kinetic values are approximately
equal for the 5 and 10 percent loading rates. In the case of the 4
' and 5-ring compounds, the 5 percent loading rate resulted in higher
kinetic rates for these compounds as compared to the 10 percent
-------
TABLE 1
COMPARISON OF PILOT SCALE KINETIC DATA
AT TWO INITIAL LOADING RATES
Benzene
Extractable
2-Ring PAH
3-Ring PAH
4-Ring PAH
TOTAL PNAs
First Order Rate
5% Plot
0.003
0.023
0.016
0.004
0.009
Constant fdav'1)
10% Plot
0.003
0.023
0.016
0.001
0.008
Half-life
5% Plot
231
30
43
173
N
77
(davs)
10% Plot
231
30
43
693
87
-------
loading rate. This difference may have been due to more 2-ring and
3-ring compounds being available to soil bacteria at the 10 percent
loading rate. These compounds may be preferentially degraded
soil bacteria.
OPERATING AND DESIGN CRITERIA
The pilot scale studies were successful in developing
operating and design criteria for a full scale system. These
criteria are summarized below:
o Treatment period can be extended through October
o Soil moisture should be maintained near field capacity
o Soil pH should be maintained between 6.0 and 7.0
o Soil carbon:nitrogen ratios should be maintained between
25:1 and 50:1
o Fertilizer applications should be completed in small
frequent doses
o Initial benzene extractable hydrocarbon contents of 5 to 1C ._
are feasible
o Waste reapplication should occur after initial soil
concentrations have been effectively degraded
o Waste reapplication rates of 2 to 3 Ib of benzene
extractable per cubic foot of soil per 3 degradation months
can be effectively degraded
The studies suggest that all the loading rates tested are
feasible. First order rate constants were fairly similar between
»
all the test plots although the intermediate loading rate (5%
benzene extractable hydrocarbons) nay demonstrate a slightly higher
removal of high molecular weight PAH compounds. The higher loading
rates, however, showed the greatest mass removals. The selection
of an initial loading rate should balance additional land area
-------
requirements against time requirements for completing the treatment
process. Moderate loading rates (5%) will result -in a faster
detoxification whereas higher loading rates will decrease land area
requirements.
CONSTRUCTION AND START-UP OF FULL SCALE SYSTEM
Construction of the full scale system involved preparation
of a treatment area within the confines of the existing RCRA
impoundment (Figure 1). The treatment area was constructed on top
of the impoundment to avoid permitting a new RCRA facility. If the
facility was located outside the impoundment, then a Part B permit
would have to be obtained before the treatment facility could be
constructed. By locating the treatment area within the confines of
the impoundment, the treatment system was considered part of
closure of the impoundment. This enabled us to fast track the
clean-up and avoid the delays associated with permitting a new RCRA
unit.
The principal construction activities at the site involved:
o Preparation of a lined waste pile for temporary storage of
the sludge and contaminated soil.
o Removal of all standing water in the impoundment.
o Excavation and segregation of the sludges for subsequent
free oil recovery.
o Excavation of approximately 3-5 feet of "visibly"
contaminated soil from the impoundment and subsequent
storage in the lined waste pile.
o Stabilization of the bottom of the impoundment as a base for
the treatment area.
o Construction of the treatment area including installation of
a 100 ml HOPE liner, a leachate collection system and 4 feet
of clean backfill.
-------
~—£- -V
1 Sit" pla for ^n- :
fa\ tnt >s n
-------
o Installation of a sump for collection of the stormwater and
leachate.
o Installation of a center pivot irrigation system.
As previously discussed, a lined treatment area was
constructed because the natural soils at the site are highly
permeable. A cap also was needed for the residual contaminants
left in place below the liner. Therefore, the treatment area liner
serves two functions at the site. The first function is to provide
a barrier to leachate from the treatment area. The second function
is to provide a cap over the residual contaminants that were left
in place.
The treatment area was constructed on top of the existing
waste water disposal pond after all contaminated materials were
removed. The surface area for treatment is approximately 125,000
\
' square feet. Containment berms with 3 to 1 slopes enclose the
treatment area and prevent surface run off from leaving the site.
The treatment area is lined with a 100 mil HOPE membrane
(Figure 2). The base of the liner slopes 0.5 percent to the south
and west. A sump with a 50,000 gallon capacity is located in the
southwest corner of the treatment area. A layer of silty sand
ballast 18 inches thick was placed on top of the treatment area
liner. A 6 inch gravel layer was placed on top of the ballast.
This layer serves as a leachate collection system and as a marking
layer for land treatment operations.
The leachate collection system includes 2 foot wide leachate
collection drains at 100 foot centers (Figure 2) . The drains are
filled with gravel and perforated pipe to carry leachate from the
collection system to the sump. The drains were wrapped in filter
-------
SE CT|QM C -C
'c 1
/ n. »
-—-j -^ ^
_, $*< ««•
1 c I —^
T37T
AltEA
QSTAtL
peTAIL
'or
De
T_,at ht -m-^o —
-------
fabric to prevent clogging. Two feet of uncontaminated sand was
placed above the leachate collection system. This layer of sand
serves as an initial mixing layer for the contaminated soils and is
the treatment zone for the full scale system.
Water in the leachate collection sump is discharged by
gravity flow to a manhole and is automatically pumped via a lift
station to a 117,000 gallon storage tank. Water in the storage
tank is recycled back to the treatment area via a spray irrigation
system. Water in excess of irrigation requirements is discharged
to the municipal wastewater treatment plant.
Construction of the waste pile and treatment area was
completed in October 1985. In late April 1986, a center pivot
irrigation system was installed and 120 tons of manure were spread
x
in the treatment area. Manure loading rates were based on
achieving a carbon:nitrogen ratio of 50:1. In addition to
nitrogen, the manure provides organic matter which enhances
absorption of the hazardous waste constituents.
In May 1986, a 3 inch lift of contaminated soil was applied
to the treatment area. The target loading rate for start-up was a
BE hydrocarbon concentration of 5 percent. The soil was mixed
(rototilled) with 3 inches of native soil to achieve a treatment
depth of 6 inches. This application involved approximately 1200
cubic yards of sludge and contaminated soil. Table 2 summarizes
start-up data for the full scale facility.
The treatment area is irrigated almost daily due to dry
weather during the summer months. Irrigation needs are determined
from soil tensiometer readings, soil moisture analyses, and
-------
TABLE 2
SUMMARY OF START-UP DATA (5/23/86)
Parameter
Average
Benzene Extractables, % 53000
TOC.ppm 29710
TKN. ppm 1367
Ammonia, ppm 2.37
Total Phosphorus, ppm 522
Total Potassium, ppm 502
pH 7.66
Polynuclear Aromatic Hydrocarbons (PAH), pom:
Naphthalene 1148
Acenaphthylene 21
Acenaphthene 1082
Total 2-Ring PAH 2251
Fluorene 1885
Phenanthrene 4190
Anthracene 3483
Total 3-Ring PAH 9558
Fluoranthene ^ 1575
Pyrene 958
Benzo(a)anthracene
and Chrysene 837
Total 4-Ring PAH 3370
Benzofluoranthenes 368
Benzopyrenes 294
lndeno(123cd)pyrene 111
Dibenzo(ah)anthracene 100
Benzo(ghi)perylene 106
Total 5-Ring PAH's 979
Total PAH's
16159
-------
precipitation and evaporation records. Typical irrigation rates
range from 1/4. inch to 3/8 inch per application. This application
rate keeps the soils in the cultivation zone moist without
saturating soils in the lower treatment zone. Maintaining soil
moisture near field capacity was determined to be a key operating
parameter in the pilot scale studies.
PERFORMANCE OF THE FULL SCALE FACILITY
Benzene extractable (BE) hydrocarbons and 16 polynuclear
aromatic (PNA) compounds are being monitored to evaluate the
performance of the facility. Figure 3 shows the BE hydrocarbon
concentrations measured in the Zone of Incorporation (ZOI) during
the first year of treatment. BE hydrocarbon concentrations
decreased approximately 60 percent over the firs't year of
operation. Most of the decrease occurred during the first 120 days
(May through September). Little decrease in BE hydrocarbon
concentrations was observed during the Fall and Winter months.
Figures 4 and 5 show PNA concentrations measured in the
treatment facility during the first year of treatment. Figure 4
summarizes data for 2-ring and 3-ring PNAs. Figure 5 summarizes
data for the 4-ring and 5-ring compounds. Greater than 95 percent
reductions in concentration were obtained for the 2 and 3 ring
PNAs. Greater than 70 percent of the 4-ring and 5-ring PNA
compounds were degraded during the first year of operation.
With the exception of anthracene, all the 2-ring and 3-ring
compounds were degraded below or near detection limits after 90
days of treatment. Greater than 92 percent of the anthracene
-------
g>
'5
C
0)
u
0
.Summer
Months
Winter
Months
60 120 180 240 300
Days after First Waste Application
360
Figure 3. BE Hydrocarbons Degradation vs Time
-------
Concentration, ppm
-»• ro co -u u
o o o o c
o o o o c
D O O O O C
1 . ! ,'!••' 1 . 1 . 1
1
K
>*
'*
•^M
1
H Naphthalene
E3 Acenaphthene
D •Fluorine'"1*
E3 Phenanthrene
D Anthracene
0
90
180
360
Days after First Waste Application
Figure 4. 2-Ring and 3-Ring PNA Degradation vs Time
-------
2000
E
Q.
Q.
C
I 1000H
4_«
c
o>
u
c
o
o
Fluoranthene
Pyrene
Benzanthracene
and Chrysene
Benzofluoranthene
Benzopyrene
0 90 180 360
Days after First Waste Application
Figure 5. 4-Ring and 5-Ring PNA Degradation vs Time
-------
present in the waste was degraded during the first 90 days of
treatment. Similarly, most of the 4 and 5 ring removals occurred
during the first 90 days of treatment. This was expected because
the warmest weather occurred during the this period.
Table 3 shows average PNA removals measured in the pilot
scale studies and compares them with the full scale removal
efficiencies. Full scale removal efficiencies were higher than
test plot removal efficiencies for every PNA ring class and BE
hydrocarbons. However, it must be noted that the full scale
facility operated for 360 days compared to only 126 days for the
test plot units. Table 3 also presents average half-life data fqr
both the test plots and the full scale unit. Full scale half-lives
were consistently, in the low end of the range of half-lives
N
reported for the test plot units.
In summary, the rate and amount of PNA degradation is
proportional to the number of rings contained by the PNA compounds
(Figure 6). The 2-ring and 3-ring PNAs degraded most rapidly. The
4-ring and 5-ring PNAs degraded at slower rates, however, these
compounds are strongly adsorbed to soils and are immobilized in the
treatment zone of the facility. Table 4 summarizes water quality
data for the leachate collection system of the facility. Only
acenaphthene and fluoranthene were detected in the drain tile */ater
samples. Concentrations for these two compounds were near
analytical detection limits.
CONCLUSION
The data developed during this project has shown that on-
-------
TABLE 3
COMPARISON OF FULL SCALE AND
TEST PLOT REMOVALS
AVE. PERCENT REMOVAL AVE. HALF-LIFE (DAYS)
Parameter - Full Scale1 Test Plots2 Full Scale Test Plots
2-Rmg PAHs
3-Ring PAHs
4- and 5-Ring PAHs
Total PAHs
BE Hydrocarbons
95
95
72
90
60
93-95
83 - 85
32 - 60
65-76
35-56
<45
45
115
65
150
29-33
46-49
95-226
61-83
106-202
1 Removal eficiency calculated attar 193 days d treatment
2 Removal efloency calculated after 126 days o< treatment.
-------
E
Q.
O.
2
c
O)
o
c
o
o
10000
8000-
6000 -
4000 -
2000-
2-Ring PAH
3-Ring PAH
D 4-RingPAH
H 5-Ring PAH
0 90 180 360
Days after First Waste Application
Figure 6. PNA Degradation by Ring Class
-------
TABLE 4
DRAIN TILE WATER QUALITY
Concentration, ppb
Compound June 1986 August 1986 October 1986
Naphthalene <1 <1 <1
1-Methylnaphthalene <1 <1 <1
2-Methylnaphthalene <1 <1 <1
Acenaphthylene <1 <1 <1
Acenaphthene <1 3.7 2.7
Fluorene <1 <1 <1
Phenanthrene <1 <1 <1
Anthracene <1 <1 <1
Fluoranthene <1 . 2.1 1.4
Pyrene <1 <1 . <1
Benzo(a)anthracene <1 <1 <1
Chrysene <1 <1 <1
Benzdfluoranthenes <5 <1 <1
Benzopyrenes <5 <1 <1
lndeno(123cd)pyrene <5 <1 <1
Dibenzo(ah)anthracene <5 <1 <1
Benzo(ghi)perylene <5 <1 <1
-------
site treatment of creosote contaminated soils is' feasible. Based
on the data developed in pilot scale studies, a conservative design
for a full scale system was developed and constructed. The fu)
scale unit has matched or surpassed the performanace of the piloc
scale unit in degrading creosote organics. The advantages of on-
site treatment are that it reduces the source of contaminants at
the site in a very cost effective manner. In addition, it
satisfies the developing philosophical approach that EPA has to on-
site remedies and it reduces the liability of the owner/operator
due to off-site disposal.
-------
ATTACHMENT C
State Letter of Concurrence
-------
ATTACHMENT D
Letter from DOI re Natural Resources Survey
-------
United States Department of the Interior
OFFICE OF THE SECRETARY
WASHINGTON, D.C. 20240
ER-84/552
Mr. Gene Lucero, Director
Office of Waste Programs Enforcement
Environmental Protection Agency
Washington, D.C. 20460
Dear Mr. Lucero:
Pursuant to your request we have conducted a preliminary natural resources survey at
the Brown Wood Preserving site, Live Oak, Suwannee County, Florida, to determine
whether any natural resources under the trusteeship of the Secretary of the Interior have
been, are being, or have the potential to be affected by releases of hazardous substances
at the site.
Our survey revealed that there are no lands, minerals, waters, or Indian resources under
T-terior's trusteeship in the vicinity of the site. Although certain fish and wildlife,
including endangered species, under our trusteeship inhabit the vicinity, there is no evi-
dence of damages to these resources at the site itself. Moreover, the probability of off-
site damages is remote.
Accordingly we would be willing to grant a release from claims for damages to natural / «/
resources under the trusteeship of the Secretary of the Interior in regard to releases cf I
hazardous substances at the Brown Wood Preserving site.
Sincerely,
Blanchard, Director
Environmental Project Review
Janet Farella
-------
ATTACHMENT E
Transcription re the Public Meeting on October 9, 1987
-------
Date:
Location:
Minutes of the Public Meeting
December 09, 1987
Live Oak City Hall, Live Oak, Florida
2 to 5 PM
The Proposed Remedy for the Brown Vfood Preserving
National Priorities List Site, Live Oak, Suwannee
County, Florida
Transcribed from cassette tape by Tony DeAngelo, 12/16/87
DeAngelo: This is the public meeting for the Brown Wood Superfund Site which
is located over on Sawmill and Goldkist. If you have not signed
in, please do so so that we have a reading on who attended the
meeting.
First, I'd like to introduce the regulatory people here. My name
is Tony DeAngelo. I'm an engineer and the Superfund Project
Manager and work out of the EPA Regional office in Atlanta. This
gentleman (points to his right) is Mike Henderson. He works for
the Office of Congressional and External Affairs at the Regional
office in Atlanta. He's our PR man. This gentleman here (points
to his left) is Charles Rooks. He's the attorney who works out of
the Regional office and is assigned to this case.
This gentleman is John Ryan frcm ReTech.
AMAX and the Brown Foundation.
He is a consultant for
Also, we have Cindy Hilty, a Supervisor from the FDER in Tallahassee.
Russ Walker, PhD, who's the State Project Manager who works on the
Site. The other gentleman here I don't know. If he would introduce
himself,....
Joe Rodes: I'm Joe Redes.
DeAngelo: The purpose of this meeting is to present our proposed remedy for
the site and to solicit any comments from anybody and how our
. proposed remedy might be modified. If anyone here has not been
down to the library, I encourage you to go down to library and at
least browse through the documents down there. Under Section 113
of the Law, the Superfund Law, there's what we call an administrative
record placed in that repository down there. From time to time new
documents will be sent down to the repository,
-------
- 2 -
_ .-Angelo: At this time I would like to present a brief description of the
Superfund process so that you can get a general idea of how the
Superfund process works and where we are in the process.
The initial step in the process is the location and identification
and grading of a large number of sites nationwide as regards their
potential or probability to cause hazard or to be hazardous to
the public health, welfare and the environment. We have a system
whereby we do preliminary assessments and site investigations
and state grants are given for that purpose. Out of this large
number of sites, about over 20,000 at this time in the nation,
we cull those which have a very high priority. These are subjeci 3
to the Hazard Ranking System scoring, a score between 0 and 100.
It goes through a very detailed verification process in Washington;
and from there the sites are verified to have that hazard and are
placed on the National Priority List. Brown Wood is one of those
sites that was placed on the List. At this time there are about
1,000 sites nationwide on the National Priority List. Brown Wood
ranks approximately in the middle or perhaps a little below middle
on the List. From this large number of sites that we get from
culling down, we get those sites that we focus a lot of attention on..
They are treated in a very special manner. There's a process w!._reby
we do what we call a remedial investigation/feasibility study which
is essentially a study which defines the nature and extent of the
contamination and then comes up with alternatives for cleanup.
From ther we go to what we call a decision document or the record of
decision. At that time EPA's comments and the public's comments
and the potentially responsible parties' comments are concentrai 1
and we come up with a remedy which is approved by EPA. From there
we go into what we call the remedial design/remedial action stage
where we design the actual remedy and then the remedial action
carries it out. After that the site is once again subjected to an
investigation, mainly of files, and we come up with a deletion
report. That is submitted to Headquarters and then finally.hopefully
the site is deleted from the National Priority List and is deemed to '
be safe. Where Brown Wood is at the moment is: we have a remedial
investigation/feasibility study done. And a record-of-decision
hopefully will be signed by the end of this month (December) and at
that time we will begin negotiating with the potentially responsible
parties, AMAX and the Brown Foundation and any other potentially .
responsible parties we can bring to bear; and see if they want to
implement the remedial design/ remedial action. You might ask why
is this present removal activity going on out there
being undertaken under Section 122 (e)(6) of the Superfund Law w,,ich
indicates that this type of activity can go on if authorized by the
President. The activity is in concert with the present proposed
draft remedy and isn't deemed to be hazardous or causing any kind
of endangerment to the public. Unless anyone has any questions
-------
- 3 -
DeAngelo: about this, what I've said already, I'll,
?r course most of the people here understand the basic history of
the site. It was a wood preserving ,,,,,,.
It basically operated for 30 years from 1948 through 1978. The
preservatives used were creosote and pentachlorophenol, the
acronym for that is PCP. There are several areas where the
creosote and the PCP have contaminated the surface soils ; and
there is a lagoon out there which has creosote sludges in it.
As far as we know at this time there's no significant groundwater
contamination. Chiefly due to the probable existence of a layer
of clay underneath the sludge in the lagoon and also because
there are a great many channels under the site which allow the
groundwater to move rather rapidly. Since the last release over
10 years ago most of the contamination
EPA's proposed remedy for the site consists of several parts.
Number 1: we propose that the sludges in the lagoon and any
other sludges that we find be removed and pretreated or treated
and be taken to a facility at Emelle, Alabama, or Pinewood,
South Carolina. Two: we propose that a test demonstration using
one or two cells using contaminated surface soil
Three: we also propose that
Four: .«•«••
I believe that pretty well describes what the EPA proposes; so
so I'd like to open it up to questions or comments.
(No questions or comments forthcoming.)
Okay.
I think that the present removal action is probably very appropriate
at this time. There are several reasons for that. Frankly, we at
the Regional office have received some flak for authorizing this
activity. However, there are overlaps of window of opportunity
that we can take advantage of at the present time, the chief one is
the dry season in this part of the South. It will cost a great deal
more if we operate in the wet season and it will present greater
problems
Number two, there is a difficult regulatory situation at the present
time. Under the Resource Conservation and Recovery Act, and
Amendments to that Law, or additions to that, there is a gradual
Land Ban going into effect whereby only certain wastes can be taken
from these sites and sent to hazardous waste facilities. It's a
gradual phase-in. There have been variances given so that certain
wastes go in at certain time periods. This type of waste here,
creosote wastes, can go in only for a
very short period of tine. When that window closes, given the
-------
- 4 -
present situation, then it can no longer be done. I was told by
Headquarters within the last couple of days that virtually all
this type of waste will eventually have to be incinerated on-site
or off-site. The type of land disposal whereby excavation and
removal was accomplished will be a thing of the past
Are there any questions? (Pause) Any questions?
Cindy Hilty: With regard to the Land Ban, what does it say exactly? RCRA
facilities will no longer be allowed to accept creosote wastes
What does it exactly cover?
Tony DeAngelo:
I gave documents to Russ (Walker)
What you will find is that it becomes difficult to know what to do
with the wastes given the lead times that you have. It's going to
become difficult to plan ahead given these varying windows of
opportunity.
Chuck Rooks: Let me just clarify that: when Tony said that there was some flak
on conducting the removal operation. Nobody is questioning the
environmental safety or anything like that. It was simply a
matter of how it was handled at the Regional office and it doesn't
present any kind of environmental harm we think
that the is appropriate and everyone agrees with that.
Dennis Price(FDER local liaison):
Tony DeAngelo:
Does anyone have any questions about the enforcement process?
Exactly how things are going to operate after the record-of-
decision is signed.
Someone: I'm sure we do.
Tony DeAngelo:
Well, all right, I want to say a few words about that. There are.
two sides to the Superfund process in terms of not only immediate
removals, chemical spills and that kind of stuff, but also in
terms of remedial or long-term activity at sites like Brown Wood.
The one side is what we call the Fund lead side. In that area
EPA people act in concert with EPA contractors to accomplish
such activities. On the other side we have the enforcement area.
That is where EPA technical people and legal personnel get together
and meet with people we feel are potentially responsible under the
Law for what's occurred at the site and try to work out agreements,.
-------
- 5 -
to accomplish cleanups It's generally
a -ncre laoorious process.
In the case of Brown Wood, if we had not done anything, not authorized
any activity within the last couple of nx>nths, in other words, if
that removal had not been done, then it would probably be at least
one wnole year before anyone would be out on the site again doing
any kind of cleanup work. Under the new law we have to go through
a consent decree process. A consent decree is an agreement.between
the potentially responsible parties and EPA wherein the potentially
responsible parties agree to do certain work within certain time
frames. That Decree is passed on by a judge in a Federal District
Court. The Decree is not presented by a Regional attorney, but by
the Department of Justice. So you can see that the mechanisms are
set up so that it's probably going to take quite a bit longer in
order to come to an agreement to actually do the actual cleanup.
I would appreciate any questions
I'm willing to take a shot
(Unintelligible.)
Cindy Hilty: (She says something about looking forward to seeing something from
Retech.)
(Unintelligible.)
Tony DeArrgelo:
Do you have any? Okay, if there's no other questions we will stay
around awhile to answer any questions you might have. I appreciate
y'all coming and I look forward to seeing you in the future. Thankyou.
-------
TABLE IV-17
CARCINOGENIC ACTIVITY OF PAHs ON MOUSE SKIN
Compound3
B[a]P
D[ah]A
CH
AN
PY
FL
Dose
(% Concentration)
0.001
0.005
0.01
0.001
0.01
0.1
1.0
10.0
10.0
0.1
Incidence of
Papillomas
<*)
43
73
95
30
95
90
45
0
0
0
Incidence of
Carcinomas
3
63
95
30
90
75
40
0
0
0
Abbreviations: B[a]P, benzo[a]pyrene; D[ah]A,
dibenz[ah]anthracene; .CH, chrysene; AN, anthracene; PY, pyrene;
FL, fluoranthene.
Source: Wynder and Hoffmann (1959).
IV-28
------- |