Unred States
Environments1 Protect on
Agency
Research and
Development
THE EVALUATION AND ESTIMATION OF POTENTIAL CARCINOGENIC
RISKS OF POLYNUCLEAR AROMATIC HYDROCARBONS (PAH)
EPA/600/D-89/049
OKEA-C-147
January 1985
Final
Prepared for
Presentation at the 1984 Chemical Congress
of Pacific Basin Societies
Prepared by
Carcinogen Assessment Group
Office of Health and
Environmental Assessment
Washington DC 20460

-------
NOTICE
This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.

-------
OHEA-C-147
January 1983
Final
THE EVALUATION AND ESTIMATION" OF POTENTIAL CARCINOCENIC
RISKS OF POLYNUCLEAR AROMATIC HYDROCARBONS (PAH)
Margaret M.L. Chu
Chao W. Chen
Carcinogen Assessment Group
Office of Health and Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
40i M Street, S.W.
Washington, DC 20460

-------
EVALUATION AND ESTIMATION OF POTENTIAL CARCINOGENIC RISKS OF
POLYNUCLEAR AROMATIC HYDROCARBONS*
MARGARET M.L. CHU, CHAO W. CHEN
Carcinogen Assessment Group, Office of Health and Environmental
Assessment, Office of Research and Development, U.S. Environ-
mental Protection Agency, 401 M Street, S.W., Washington, D.C.
20460.
INTRODUCTION
The evaluation and estimation of potential risks of human
exposures to hazardous chemicals such as polynuclear aromatic
hydrocarbons (PAHs) can be useful in the setting of permissible
levels of hazardous chemicals in the workplace and the environ-
ment, or in the setting of regulatory priorities. The hazard
potential can be evaluated by considering all of the parameters
related to the fate, effects, and dose-response characteristics
of the hazardous substance in question. Important parameters
for assessing such hazards are chemical structure; physical-
chemical properties; mechanisms of action; chemical, biological,
and environmental transformation and transport; and toxicity
indices. Some of these parameters can be estimated directly
from experimental data, while others may be estimated indirect-
ly through the use of modeling techniques.
Risk can be conceptualized as a composite function of the
hazard and exposure potentials. In order to estimate the poten-
tial risk of human exposure to a hazardous chemical, exposure
parameters such as level, duration, frequency, and route are
needed. Exposure parameters can be derived using monitoring or
modeling results. The ability of the chemical to induce the
potential effect (i.e., a "potency" factor) can then be coupled
to the magnitude of exposure to give an estimate of the poten-
tial risk.
The assessment of human cancer risk is a complicated
scientific undertaking. It relies heavily upon available data,
*The views expressed in this paper are those of the authors and
not necessarily those of the U.S. Environmental Protection
Agency.
I

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
scientific assumptions, and judgments to bridge data yaps.
Tnus there is great uncertainty in every step of the process.
The extent and the form of a risk assessment also depends
upon the uses for wnich it is designed.
This paper presents the general framework of current
approaches useful in the assessment of potential carcinogenic
risks; problems associated with these approaches with empnasis
related to the assessment of specific individual ^AHs; and,
finally, alternatives that can be developed as more data gaps
are filled in the near future.
EVALUATION OF POTENTIAL CARCINOGENIC HAZARDS
Substances suspected of Deing carcinogenic hazards can be
evaluated by considering available chemical, Diological, and
toxicologic data. This process, called hazard identification
(25), relies heavily on three types of information: (1) epide-
miologic and clinical studies of human populations, (2) long-
term experimental animal studies, and (3) short-term _in_ vivo
and i_n vitro tests, comparative metabolism, pharmacokinetics,
and other biochemical and mechanistic studies, including struc-
ture-acti vi ty correlations.
Tne types and volume of information available, and their
contribution to such assessments, vary from compound to com-
pound. The weignt-of-evidence approach can be used to organize
this information in formulating a judgment of the potential
carcinogenic hazard of the compound at hand (I, 18, 40, 41).
Tne weight-of-evidence is defined as tne degree of evidence for
carcinogenicity in numans, and not the relative carcinogenic
activity or "potency" of the agent.
Tne most complete form of weight-of-evidence determina-
tion is made from a consideration of the validity, quality ana
relevance of each epidemiologic and long-term animal study, as
well as all short-term toxicologic, biological, chemical, and
mechanistic information.
Epidemiologic Studies
Epidemiologic studies can, under certain conditions, pro-
vide direct evidence of the association of increases in tumor
incidence in humans with exposure to specific chemicals. Con-
sistent results in independent studies, freedom from bias and
confounding factors, reliable exposure data, sufficient follow-
up time, and high levels of statistical significance are impor-
2

-------
POTENTIAL CARCINOGENIC RISKS CF PAHs
tant factors fading to increased confidence in the determina-
tion o* a causal relationship.
Long-Tern Anirral Studies (Carcinogenesis Bioassays)
While ^uman data provide the most direct evidence for the
cardnogenicity of a compound, usuaTy such data either c!o not
exist or are inadequate. In tne absence of human data, reliance
is placed on information from long-tern animal studies in
assessing the potential carcinogenic ns< to humans. It should
be noted that information compiled and evaluated by the Inter-
natioral Agency for Research on Cancer (I ARC ) shows that chemi-
cals or groups of chenicals that are krown to be human carci-
nogens and have been tested appropriately, produce cancer ir
animal.s. "h^s certainly supports the use of animal data as
indicators of potential human carcinogenicity.
The general factors considered in evaluating the carcino-
genicity of chemicals using animal studies are the induction of
rare tumors, the earlier induction of tumors, and the induction
of higher incidences of tumors when compared to control animals.
Confidence in the results of animal experiments is gained when
the carcinogenic effects have Seen confirmed in repeated experi-
ments, and have been observed in different strains or species,
in different dose groups or sexes, or in Tultiple organs or
tissues, with high degrees of malignancy and dose-related
t rends.
Short-Term Tests, Structure-Activity Relationships, and Metabo-
lic, Pharmacokinetic, and Mechanistic Studies
Results from short-term tests, structure-activity analy-
ses, and metabolic, pharmacokinetic , and mechanistic studies
are currently being used as supportive evidence in modi-
fying judgments based on epidemiologic and long-term animal
studies. By far the largest volume of information recently
generated is in the area of short-term tests for mutagenicity.
While the IARC has not formally incorporated mutagenesis testirg
results in its classificatior system, an evaluation scheme
for analyzing mutagenicity testing data has been developed (18,
19). Short-term in vi vo carcinogeneses testing (sometimes
referred to as " 1imited-bioassay") such as skin-painting expe-
riments with mice, and mouse-skin initiatior-promotion studies,
are not given the same status as the conventional carcinogene-
sis bioassay. Structure-activity analyses and metabolic, phar-
macokinetic, and mechanistic studies are evaluated on a case-
by-case basis, since their availability is variable.
3

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
The results of the analysis of epideiniologic evidence and
evidence from long-term animal studies are combined to determine
the human carcinogenic hazard potential of the chemical being
evaluated. Tnis determination is modified on the basis of data
fron short-term tests and other supportive information (18, 41)
ESTIMATION OF POTENTIAL CARCINOGENIC RISKS
After qualitative evaluations of the data bearing on a
chemical's ability to induce a carcinogenic effect, and the
relevance of such data to humans, it is desirable to estimate the
magnitude of the potential nunan risks. Two additional catego-
ries of data are usually needed to provide such an estimate: (1)
dose-response data from which a "potency factor" can be derived;
(2) information on the number of humans and the types, levels,
and durations of potential human exposures. The results of
coupling the potency factor with the magnitude of exposure will
provide a numerical estimate of potential human risk.
This paper is concerned mainly with estimating the carci-
nogenic potential of a compound. The evaluation of potential
human exposures and the estimation of risks based on potential
exposures will not be considered here.
Estimating Carcinogenic Potency
The carcinogenic potency of a compound can be defined as
the probability of an individual's developing cancer in his or
ner lifetime following exposure to a unit aose, if the unit
dose is sufficiently small.
The carcinogenic potency of a known or suspect carcinogen
cannot be estimated with accuracy because it is not possible to
determine the shape of the dose-response curve beyond experi-
mental exposure levels. In the absence of knowledge regarding
the shape of the dose-response curve, the multistage model is
used for low-dose extrapolation to provide an upper-bound esti-
mate of carcinogenic potency when animal bioassay data are used.
The reasons for selecting tne multistage model and using the
upper-bound estimate are given in the following section. When
human data are used, the procedure for estimating carcinogenic
potency, and the accuracy of such estimates, depend on the
availability and the quality of the data. The data reported in
an epidemiologic study may range from a simple relative risk
estimate associated with a rough estimation of average exposure
to a full report on each individual in the cohort, including
information such as age, cause of death, detailed work history,
4

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
smoking habits, and length of exposure.
Choice of the Extrapolation Model
Because the procedure for estimating risx fron human data
varies depending on the availability ana quality of data, we
concentrate our discussion on the procedure for estimating car-
cinogenic potency where animal bioassay data are used. Several
dose-response models are available for low-dose extrapolation.
These include the probit, the multi-hit, the logit, and the
nultistage models. These models are generally statistical in
character, and are not derived from biological arguments, except
for the multistage model, which has been used to support the
somatic mutation hypothesis of carcinogenesis (3, 42, 43). The
mam difference among these models is the rate at which the
response function, P(d), approacnes P(0) as dose d decreases.
For instance, the probit model would usually predict a smaller
risx at low doses than the multistage model because of the
difference of the decreasing rate in the low-dose region.
However, it should be noted tnat one could always artificially
make the multistage model have the same or even greater rate of
decrease as the probit nodel by transforming the dose rate
and/or by assuming that sone of the parameters in the nultistage
model are zero. This, of course, is not reasonable without
knowing, a priori, what the carcinogenic process for the agent
is. The multistage model is used for the extrapolation because
it is the most general model, with other models approximating
some form of tne multistage model according to the values of
the parameters. Although the multistage model appears to be a
reasonable (at least the most general) model to use, the point
estimate generated from tne model is not usea because a ques-
tion remains as to the shape of the dose-response curve beyond
the experimental exposure level. Therefore, the upper-bound
estimate of the carcinogenic potency is derived when animal
bioassay data are used. This upper-bound estimate can be taken
as a plausible estimate if the true dose-response curve is
actually linear at low doses. Upper-bound estimation means
that the risks are unlikely to be higher but could be lower if
the compound has a concave dose-response curve or if there is a
threshold at lower doses. The other reason why the upper-bound
estimate is used instead of the point estimate is that, in some
cases, the point estimate is extremely unstable, depending on
where the lowest experimental dose is, while the upper-bound
estimate is much more stable.

-------
POTENTIAL CARCINOGENIC RISKS OF ^AHs
THE EVALUATION OF CARCINOGENIC POTENTIAL AND ESTIMATION
OF CARCINOGENIC POTENCIES OF SOME PAhs
Epidemiologic Evidence for Carcinogenicity of PAHs
Hunans are exposed to PAHs in the form of complex mixtures
rather than single compounds. For this reason, njman data for
exposure to specific PAHs are not available. The evaluation of
human evidence for carcinogenic risks of exposure to PAHs thus
must rely largely on experimental evidence from aninal studies.
Where available, human data can be classified as follows,
using the IARC criteria (18). Sufficient evidence for carcino-
genicity in humans requires the finding of causal association
between chemical exposure and cancer in humans on the basis of
analytical epidemiologic studies. Limited evidence indicates
that a causal relationship is credible but tnat alternative
explanations cannot be excluded, and inadequate evidence indi-
cates that there are few pertinent data, that the data do not
show association, or that tne data do not exclude chance, bias,
or confounding.
Eviaence from Animal Studies
Very few long-term animal studies have been conducted on
PAhs (with the exception of benzo(a)pyrene). Table 1 summa-
rizes the results of the studies in which PAHs were admini-
stered to animals orally.
PAHs were the first class of compounds snown to be carcino-
genic in experimental animals (17, 19). Although numerous
routes of administration, in several animal species, have seen
used in the study of benzo(a)pyrene, the majority of the animal
studies on other PAHs have been mouse skin assays. The IARC
stated that data from mouse skin assays may contribute to
sufficient evidence of carcinogenicity because certain PAHs
initially established as carcinogenic by application to mouse
sxin have been shown to produce malignant tumors at other sites
following administration via other routes (19). Table 2 gives
the results of the evaluation of the animal evidence for carci-
nogenicity of the 30 non-substituted PAHs evaluated oy the IARC
that have been shown to occur in the environment.
Using IARC criteria (18), animal evidence is classified
as sufficient when a carcinogenic effect is observed in more
than one strain or species, in more than one experiment, or via
more than one route of administration; or in which the degree of
6

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
TABLE 1.
CARCINOGENICITY OF PAH BY ORAL ADMINISTRATION,
I
PAHd
Speci es
Dose
Route of
admi ni strati on
Tumorigenie effects
B[a]P
Mouse
0.2 mg in poly-
ethylene glycol
Intragastric
14 tumors of the forestomach in
5 animals out of 11

Rat
(Sprague-Dawley;
age 105 days)
2.5 mg per day
Oral
Papillomas developed in the eso-
phagus and forestomach in 3 out
40 animals

Hamster
2-5 mg biweekly
Intragastri c
5 stomach papillomas developed in
67 animals treated for 1-5 months;
7 papillomas and 2 carcinomas in
18 animals treated for 6-9 months;
5 papillomas in 8 animals treated
for 10-11 months

Hamster
500 ppm
Dietary
(4 days per
week for up
to 14 months)
12 tumors (2 esophagus, 8 fore-
stomach, 2 intestinal) in 8
ani mals
DB[a,h]A
Mouse
9-19 mg
(total dose)
Dietary
(5-7 months)
Forestomach tumors in 7 of 22 sur-
vivors after one year; one tumor
was a carcinoma
aAbbrevi ations for PAHs are as follows: B[a]P = Benzo[a ]pyrene; DB[a,h]A = Di benzo[a ,h]arithracene;
B[a]A = Benzo[aJanthracene.
7
(continued on the following page)

-------
POTENTIAL
TABLE 1.
CARCINOGENIC RISKS OF PAHs
(conti nued)
PAH
Species
Dose
Route of
adini ni strati on
1
Tumorigenic effects
DB[a,h]A
(cont.)
Mouse
(A backcross)
Mouse
(DBA/2)
0.4 mg per day
0.76-0.85 mg
per day
Oil emulsion
(drinking water
replacement)
Oil emulsion
(drinking water
replacement)
11 papillomas of the forestomach
in 20 animals within 406 days
Pulmonary adenomatosis in all 27
survivors at 200 days; 24 animals
had alveologenic carcinomas; 16
had hemangio-endothelimoas; 12 of
13 females had mammary carcinomas
2 pulmonary adenomatoses seen
among 25 controls

Mouse
(Swiss,
male)
1.5 mg in
polyethylene
glycol
Oral
(single dose)
Papillomas of the forestomach in 2
out of 42 animals within 30 weeks

Mouse
(BALB/c,
female)
15 mg total dose
in almond oi1
Intragastric
(twice weekly
for 15 weeks)
Mammary carcinomas in 1 out of 20
intact animals and 13 out of 24
pseudo-pregnant animals
B[a]A
Mouse
0.5 ing in
mineral oil
Stomach tube
(single dose)
No tumors in 13 mice in 16 months
(continued on the following page)
8

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
TABLE 1. (continued)



Route of
i
PAH
Species
Dose
administration
Tumorigenic effects
B[a]A
Mouse
0.5 mg in
Stomach tube
Papillomas in 2 out of 27 mice; no
(cont.)

mineral oil
(8 or 16 admi-
tumors in mineral oil group



ni strati ons




at 3- to 7-day




i ntervals)


Mouse
1.5 mg as a
Stomach tube
Lung adenomas in 56 of 59;

(B6AF1/J)
3% solution
(15 times in
hepatomas in 38 of 59;


in metho-
five weeks)
papillomas of the stomach in 2


celaerosol OF





2 times
Lung adenomas in 1/ of 20;



3 days apart
hepatomas in 16 of 20




Controls:




Lung adenomas in 10 of 59;




hepatomas in 2 of.59
SOURCE: IARC Monographs, Vol. 3, 1973.
9

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
TABLE 2.
DEGREE Of EVIDENCE OF CARCINOGENICITY FOR EXPERIMENTAL ANIMALS
OF NON-SUBSTITUTED PAHs.

Degree of evidence

of carcinogenicity
Chemi cal
for experimental animals
Anthanthrene
Limi ted
Antnracene
No evidence
Benz[a]anthracenea
Suffi ci ent
Benzo[b]fluoranthene
Suffi ci ent
Benzo[j]f1uoranthene
Suffi ci ent
Benzo[k]f1uoranthene
Sufficient
Benzo[ghi]fluoranthene
Inadequate
3enzo[a]fluorene
Inadequate
3enzo[b]f1uorene
Inadequate
Benzo[c]fluorene
Inadequate
Benzo^ghi]pery1ene
Inadequate
Benzo[c]phenanthrene
Inadequate
Benzo[a]pyrenea
Suffi cient
Benzo[e]pyrene
Inadequate
Chrysene
Limi ted
Cyclopental[cdjpyrene
Limi ted
Dibenz[a,c]antnracene
Limi ted
Dioenz[a ,n]anthracenea
Suffi ci ent
Di benz[a, 1Janthracene
Limi ted
Di benzo[a,e]f1uoranthene
Limited
Di benzo[a,e]pyrene
Sufficient
Dibenzo[a,n]pyrene
Suffi ci ent
Qibenzo[a,i Jpyrene
Suffi cient
Di benzo[a,1]pyrene
Suffi ci ent
F1uoranthene
No evidence
F1uorene
Inadequate
Indeno[l,2,3-cd]pyrene
Suffi ci ent
Perylene
Inadequate
Phenanthrene
Inadequate
Pyrene
No evidence
aChemicals which also have
oral and/or inhalation studies.
SOURCE: Adapted from IARC
Volume 33, 1984,
10

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
tumor incidence, site, type or 1atency-shortening is unusual.
A limited-evidence classification signifies limitation in the
quality or reporting of the studies, and a limited number of
species or strains tested or experiments performed. Evidence
is judged to be inadequate when the results of the studies
cannot be interpreted as showing the presence or absence of a
carcinogenic effect. A no-evidence category is used for chemi-
cals with no observed carcinogenic effects in several animal
studies.
Evidence from Short-Term Tests
The extent of short-term tests on individual PAHs varies
(19). A summary of the I ARC's conclusions is tabulated in
Table 3. Tne IARC first proposed and considered the results of
short-term tests in making an overall evaluation of carcinogenic
risk of chemicals to humans in 1982 (18). The scheme proposed
by the IARC is as follows:
i.	Sufficient evidence: When there were a total of at
least three positive results in at least two of tnree test
systems measuring DNA damage, mutagenicity, or chromosal anoma-
lies. When two of the positive results were for the sane
biological endpoint, they had to be derived from systems of
different complexity.
ii.	Limited evidence: When there were at least two posi-
tive resuIts, ei ther for different endpoints or in systems
representing two levels of biological complexity.
iii.	Inadequate evidence: When there were too few data
for .an adequate evaluation, or when there were contradictory
data.
iv.	No evidence: When there were many negative results
from a variety o7 short-term tests with different endpoints,
and at different levels of biological complexity. If certain
biological endpoints are not adequately covered, this is indi-
cated.
Overall Evidence for Carcinogenicity
Table 4 is a summary of the overall evidence for carcinoge-
nicity of the 30 non-substituted PAHs, incorporating human, ani-
mal and short-term test results. Two categorization schemes
(18, 41) are used in the evaluation. For the IARC grouping
scheme, Group 1 (human carcinogen) is reserved for compounds
with sufficient evidence from epidemiologic studies; Group 2
11

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
TABLE 3,
DEGREE OF EVIDENCE IN SHORT-TERM MUTAGENICITY TESTS OF NON-SUB-
STITUTED PAHs EVALUATED BY AN
IARC WORKING GROUP IN FEBRUARY
1983.

Degree of

evidence in snort-
Chemi cal
term mutagenicity tests
Antnanthrene
Inadequate
Anthracene
No evidence
Benz[a]anthracene
Sufficient
Benzo[b]fluoranthene
Inadequate
Benzo[j]fluoranthene
Inadequate
Benzo[k]fluoranthene
Inadequate
3enzo[ghi]fluoranthene
Inadequate
Benzo[a]fluorene
Inadequate
Benzo[b]f1uorene
Inadequate
Benzo[c]fluorene
Inadequate
Benzo[ghi]perylene
Inadequate
3enzo[c]phenanthrene
Inadequate
Benzo[a]pyrene
Suffi cient
Benzo[e]pyrene
Limited
Chrysene
Limited
Cyclopenta[cd]pyrene
Suffi cient
Di benz[a,c]anthracene
Sufficient
Di oenz[a,h]anthracene
Sufficient
Di benz[a,jjanthracene
Inadequate
Dibenzo[a,e]fluoranthene
No data
Dibenzo[a,e]pyrene
Inadequate
Dibenzo[a,h]pyrene
Inadequate
Dibenzo[a,i]pyrene
Inadequate
Di benzo[a,1Jpyrene
No data
Fluoranthene
Limited
Fluorene
Inadequate
Indeno[l,2,3-cdjpyrene
Inadequate
Perylene
Inadequate
Phenanthrene
Li mi ted
Pyrene
Limited
SOURCE: IARC Monographs, Vol. 32, 1983.
12

-------
('() r t N TIAL CARCINOGENIC RISKS (It PAIIs
I ABLE
SUMMARY UP THE DEGREE 01 EVIDENCE FROM HUMAN, ANIMAL AND SH0R1 - TERM IE5IS OF NON-SIIBSI 1IIJUI) PAIIs
KVAUiatf.d by AN I ARC WORKING GkCTuF IN FEBRUARY lWI
PAH
Degree of evidence
Human	of carcinogenicity
evidence for experiment a I animals
Degree of
evidencp in
short -t i'rni
mutageni c i ty
tests
OvpraI I
evidence of
(.arc i nogeni c i ty
based on
IAR(.
TP7T
groups" groups'
Ant hanthrpne
No
data
Limit ed
Inadequate
3
C
Anthracene
No
data
No evidence
No evidence
--
E
Renzf a Janthracene
No
data
Suf fIcient
Suf flei ent
211
R2
Renzolbjfluoranthene
No
data
Suf f1c ient
Inadequate
2B
B2
Renzolj ]fluoranthpnp
No
data
Suf f i c i ent
Inadequate
21!
B2
Renzo[k]fluoranthene
No
data
Suf f i c lent.
Inadequate
2R
l!2
Benzolghljfluoranthene
No
data
Inadequate
Inadequate
3
I)
Renzof a ]f1uorene
No
data
Inadequate
Inadequate
3
D
Benzo[bJf luorene
No
data
Inadequate
Inadequate
3
1)
Benzolc ]fluorene
No
data
1 nadegnate
Inadequate
3
1)
Benzo[gh1 Jperytene
No
da! a
Inadequate
Inadequate
3
1)
Renzo[c Jphenanthrene
No
data
1 nadequate
Inadequate
3
0
Benzo[a Jpyrene
No
data
Suf f leienL
Sn f f1c i ent
2R
B?
Benzofelpyrene
No
data
1nadequate
L imi ted
3
U
Chrysene
No
data
Li mi ted
Llmi ted
3
C
Cyr1oppnta[cd ]pyrpne
No
data
L i mi ted
Suf f icient
3
C
l)ihenz[ a ,r. Janthracene
No
data
Limlted
Suf f ic ient
3
C
D1benz[a ,h]anthrar.pnp
No
data
Suf f 1 c. ient
Suf f icient
2R
R2
Dibpnz[a,jjanthracene
No
data
Limi ted
Inadtqqate
3
C
Dibenzo[a ,ejfluoranthene No
data
LI ml ted
No data
3
C
Di hen zo[.a ,e]py rene
No
data
Suf f icient
Inadequate
2B
R2
0i henzo[a ,h]pyrpnp
No
data
Suf fi ci ent
Inadequate
2R
B2
l)ibpnzo[a, i ]pyrene
No
data
Sufficient
Inadequate
2R
R2
Dibenzo[a,1 ]pyrenp
No
data
Suf f i c ient
No data
2B
R2
Fluoranthpnp
No
data
No evidence
Limi ted

D
F luorene
No
data
Inadequate
Inadequat e
3
D
1ndeno[ 1, 3-cd ]pyrenp
No
data
Suf f ir i ent
Inadequate
2R
R2
Perylpne
No
data
Inadequate
1 nadequate
3
0
Phenanthrenp
No
data
Inadequate
Limi ted
3
1)
Pyrene
No
data
Nn evidence
L imi ted
--
n
¦''I ARC ('.roups: "1^ Tiuma~n ca rc.1 nogen ~ ?A anTTTR,- prohab I e human carcinogen; 3, carcinogenicity to humans
cannot lie c I ass i f i pd .
''EPA Groups: A, human rare i nogpn; B1 and R2, probable human carcinogen; C, possible human carcinogen;
I), not classified; T, no evidence for human carcinogenicity.
SOURCE: IARC Monographs, Vol. 32, 19H -J.
II

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
(probable human carcinogen) is subdivided into 2A (at least
limited human evidence) and 26 (sufficient evidence from animal
studies). Group 3 (not classified) includes chemicals with
limited, inadequate, or no evidence in animal studies.
The proposed EPA grouping scheme for categorizing the over-
all evidence is similar to the IARC grouping scheme. The pro-
posed EPA Group A is equivalent to IARC Group 1, and EPA Groups
61 and 32 are equivalent to IARC Groups 2A and 23. In the EPA
scheme, Group C is restricted to chemicals with limited animal
evidence, Group 0 is for cnemicals with inadequate human and/or
animal data, and Group E is for chemicals with no human and/or
animal evidence.
Estimation of Carcinogenic Potency of PAHs
The first paper on the relative potencies of a series of
PAHs was published in 1939 (14). In this study, the investi-
gator collected the results of mouse-skin carcinogenesis tests
on PAHs and derived a method to compare potencies. The carci-
nogenic potency index is commonly referred to as the Iball index
(percent tumor incidence x 100/mean latency period in days).
The deficiencies of tnis index are that it does not reflect the
dosage administered, and it assumes tnat the tumor response is
linearly related to age, while it is known that tumor response
is exponentially related to age.
Inhalation and ingestion are important routes of numan ex-
posure to PAHs. It is desirable to estimate potency factors
for these routes of exposure. For benzo[a]pyrene, because data
are available from inhalation and oral routes of administration,
potency estimates can be derived by means of the data in Tables
5 and 6.
Using data from Table 5 and the linearized multistage model
(1), the carcinogenic potency of B[a]P by oral exposure is esti-
mated to be qf = 11/(mg/ky/day). The value qf is the 95%
upper confidence limit of the linear component q\ in the multi-
stage model
P(d) = 1 - exp[- q^d - q2d^ ... - qKdk]
Under the multistage model, the cancer risk p(d) at a con-
stant exposure d can be calculated by p(d) = q\ x d when d is
sufficiently small and when q^ / 0. However, the maximum
likelihood estimate (MLE) for the linear component q\ is very
unstable and may be estimated to be zero even if the true dose-
response model contains both non-zero linear and higher order
14

-------
TABLE 5
B[a]P POTENCY BY ORAL ROUTE:
PAPILLOMA/CARCINOMAd
INCIDENCE RATE OF STOMACH
Dose
(ppm in diet
Dose	Incidence rate of
mg/kg/day) stomacn papi1loma/carcinoma
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
aThis table contains only those groups that are comparable with
respect to age at exposure, number of days exposed, and age
ki1 led.
SOURCE: Neal and Rigdon, 1967.
TABLE 6.
B[a]P POTENCY BY INHALATION: INCIDENCE RATE OF RESPIRATORY
TRACT TUMORS IN SYRIAN GOLDEN HAMSTER.
Dose
Incidence
(ng/Ti3)
rates
0
0/27
2.2
0/27
9.5
9/26
46.5a
13/25
aBecause of the higher mortality rate in the highest dose
group, the data from this group is excluded in the calcu
lation of potency.
SOURCE: Thyssen etal., 1981.
15

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
polynomial terms. This can easily be seen by observing that
two response models (one containing only a quadratic term and
the other containing both linear and quadratic terms) can
adequately fit a set of experimental tumor incidence data.
This is because the dose-response model is dominated by higher
order polynomial terms at the high-dose range, as in the expe-
rimental data. However, a dose-response model that contains a
linear component predicts significantly larger risk, at low
doses, than does a model without a linear component, ^ince
the upper confidence limit for the linear component,	is
always positive and its estimate is "robust," the value qj is
used to represent the carcinogenic potency of a compound. At
low doses, the risk is calculated by qf x d.
Using data from Table & and the linearized multistage
model (1), the carcinogenic potency ot B[a]P by inhalation
exposure is estimated to be q| = 4 x 10"'/{ug/m
The carcinogenic potency of otner PAHs can be estimated by
reference to the potency of benzo[a]pyrene as a function of the
relative potency index using mouse-skin painting data. The
potency of PAHj can be expressed as the following equations:
,3 \
(1) potency PAH} (oral)
= potency
potency
PAHi
Bla]P
(skin)
(ski n
potency of
3[a]P (ora
(2) potency PAH^
inhala-) = potency PAHi
tion	potency B[a]P
skin) potency of
(skin) x B[a]P (inha
1ati on
Listed below are seven PAHs for which sufficient informa-
tion is available to apply this approach. The results are
presented in Tables 7 and 8.
	PAH (abbreviation)
1.	Benzo[a]anthracene (B[a]A)
2.	Benzo[a]pyrene (B[a]P)
3.	Chrysene
4.	Benzo[k]fluoranthene (B[k]F)
5.	Dibenzo[a,h]anthrancene (DB[a,h]A)
6.	Inaeno[l ,2,3-c ,djpyrene (I[1,2,3-c,d]P
7.	Benzo[b]f1uoranthene (B[b]F)
Carcinogenicity
evidence from
animal stuaies
Sufficient
Suffi cient
Limi ted
Suffi ci ent
Sufficient
Sufficient
Suffi ci ent
16

-------
POTINIIAI CARC I NOlil NIC RISKS OF PAIIs
TAIll F 7.
INDICES USED TO RANK 7 PAHS.
PAHs
B[ a IP
B[aP
B[a |P
Q|	f'lO an(' I .I-.
ed[o and 951- C.I.
References
4/0	2.98 * 10"3	153.49
(8.6? * 10"'', 5.98 x 10-1)
9.3? x 10"*	(14)
(6.54 x lO"4, l.2l x ID-')
435
1.43 x lO"3	67.62
(3.68 x lO"4, l.43 x lO"3)
l.7l x l(H	(40)
(8.si x in-4, 2.5/ x in--3)
N.A. N.A.
20.83
Dill a ,h JA	299.62 6. 34 x lO'4	292.81
( 3.24 x 10"4 , 9.41 x lO"4)
BLklF
N.A. N.A.
I.43 x 10"?	(5)
(4.24 x K)"3, 2.44 x lO"2)
6.16 x 10"4	(45)
(3.28 x lO"4, 9.04 x II)-'1)
Nn point est.
0.30	(n.35, none)
(46)
B[h JF	35.64 5.0 x 10"3	11 .57	1 .29 x 10"2	(46)
(2.75 x 10"3, 7.25 x 10"3)	(8.54 x lO'3, 1.73 x lO"?)
Chrysenp
0.53 0.35
(0.23, 0.47)
0.88	0.21
(0.11, 0.31)
(45)
B(.a JA	N.A. N.A.
(0.34, 1.12)
0.28
0.73
(0.34, 1.12)
(5)
l[1,2,3,-c,dJP N.A	N.A.
(0.08, 0.22)
1.16	0.15
(0.08, 0.22)
(45)
Remarks: 1. Qi is the 45% upper confidence limit for the linear component in a multistage model and
F0j(j is the dose level corresponding to the 10% incremental tumor response when time-to-
tumor data are used. lhe values of c|J and ed\o are similarly defined when incidence data
are used.
2.	N. A. : Not aval I al> I e.	k '
3.	Since there is only one dose group for Chrysene, ()j is calculated hy usimi Kaplan-Meier
survival analysis and the assumption that the control group has a ?ci o response.
I 7

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
TABLE 8.
RELATIVTPOTENCY IN REFERENCE TO B[ajP AND 95% CONFIDENCE LIMIT
Compounds
Based on Q|
(or q})
compound/B[a]P
Based on EDxq (°r ed10)
B[a]P compound and 95%
confidence limits

DB[a ,h]A
0.69
2.26
(0.56, 5.36)

B[b]F
0.08
0.29
(0.07, 0.65)

Chrysene
1.22 x 10-3
4.09 x 10"3
(1.06 x 10-3, 7.12 x
10"3)
I[L,2,3-c,d]P
1.71 x lO-2
1.14 x lO"2
(5.09 x 10"3, 2.63 x
10-2)
B[a]A
1.34 x lO"2
1.96 x 10"2
(5.48 x 10"3, 4.96 x
10*2)
B[k]F
4.44 x 10"3
none
none

Remarks:
~ ~
1.	When available, Qj and EDjq are preferred to and ed^Q
in deriving the relative potency,
2.	The confidence limits of the ratio are constructed using
Geary's theorem.
18

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
The appropriateness of this approacn, however, is uncer-
tain at the present time. Further investigation and analysis
using information from mechanistic, pharmacokinetic, and macro-
molecular binding studies may provide additional insight for the
estimation of oral and inhalation potencies using s '< i n-pai nti ng
carcinogenicity data.
LIMITATIONS OF CURRENT ASSESSMENT APPROACHES
AND POTENTIAL ALTERNATIVES
While there is general agreement within tne scientific
community about the general approach (15, 16, 29) that should
be used in carcinogen risk assessment, judgments and assump-
tions made to fill data gaps are often controversial. With
the advancement of research, tne scientific data base will grow
and the uncertainties involved in risk assessment will be
reduced. Ideally, testing data based on the knowledge of the
mechanism of carcinogenesis would increase the certainty of the
results in assessing human risks. However, our current under-
standing of the mechanism of carcinogenesis is very limited.
Major advancements have been made in the understanding of the
initiation steps of carcinogenesis, but events related to
promotion and progression to tumor formation are largely unex-
plored. Thus, assessments of carcinogenic risks to humans are
based on observational and correlative data and plausible
assumptions and judgments. The following sections summarize
some of the limitations of current assessment approaches and
present a forward look at potential alternatives.
Evaluation of Potential Carcinogenic Hazards
(1) Epidemiologic studies. While epidemiologic studies can
provide direct evidence For carcinogenicity in humans, the
following limitations are inherent in such studies:
a)	They are expensive and time-consuming.
b)	Their sensitivity may be limited because of small num-
bers of persons exposed.
c)	Individuals in such studies have often been exposed to
several carcinogenic compounds.
d)	The duration of follow-up may not be sufficiently long
for cancer manifestation.
e)	Such studies provide associative evidence only, and
usually provide limited information on exposure for
specific agents.
19

-------
kUlENTIAL CARCINOGENIC RISKS OF PAHs
(2)	Rodent Long-Term Carcinogenesis Bioassays. Although dt
present the carcinogenesis bioassay is the most reliable method
for carcinogen identification, the bioassay presents certain
limitations in assessing human carcinogenicity.
a) Besides being expensive, the standard bioassay (24)
of a single compound requires about 500 rodents and
takes two years for the experiment and at least an
additional year for pathology, statistical analysis,
and preparation of the report.
d) Such studies provide data only at two dose levels and
a control 1evel.
c)	Because of tne inherent insensitivity of these studies,
they often involve high doses which are jsually several
orders of magnitude above potential human exposures.
Thus, they do not provide linearity and threshold in-
formation on dose-response behavior at low doses.
d)	Negative results in such studies are difficult to in-
terpret because only small numbers (50) of animals are
tested per dose, and the species tested are usually
only rats and nice.
(3)	Short-Term Mutagenesis Tests. The use of these tests to
predict animal and human response relies heavily upon our
knowledge about the mechanisms of carcinogenesis, which is very
incomplete at this time.
a)	Mutagenesis testing is based on the observation that
many carcinogens either directly or indirectly gene-
rate electrophilic metabolites that bind to DNA. Thus,
one of the weaknesses of the test system is the re-
quirement of an activation system, which contributes
to the variability of test results, depending on the
source of the activation factor.
b)	Since no single test will detect all potential carci-
nogens, it is necessary to develop an appropriate bat-
tery of short-term tests based on the predictive values
of specific tests with specific classes of compounds.
c)	Short-term tests for the study of agents which affect
the carcinogenesis process by mechanisms other than
genotoxicity are not available.
d)	The carcinogenesis process progresses in multiple
stages. Short-tern mutagenicity tests, with the excep-
tion of cell transformation assays, detect only initi-
ation activities.
20

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
Exposure Evaluations
The'exposures measured by selective monitoring measure-
ments or estimated through modeling nay not correlate well
with actual experience in larger populations. There is uncer-
tainty in monitored levels of chemicals at trace levels in the
workplace or environment. In addition, high-risk populations
are usually not identified, Monitored data generally tend to
be crude estimates relying on modeling in which assumptions have
to be made.
Tne relationship of exposure level to biological dose de-
pends on information such as absorption and excretion; environ-
mental transformation; bioaccumulation, and thus bio-magnifi-
cation via the food chain; frequency, duration, intensity and
route differences; and chemical interactions. Such data are
usually unavailable in exposure evaluations.
Estimation of Carcinogenic Risks
The major limitation in the estimation of potential carci-
nogenic risks to humans is the lack of a mechanistic understand-
ing of the process of carcinogenesis, and tne paucity of scien-
tific data that can be used in quantitative evaluation. This
limitation contributes to the nigh degree of uncertainty asso-
ciated with the estimation of potential human carcinogenic
risks.
Some of the key contrioutors to the uncertainty of risk
estimates are: uncertainty of low-dose behavior because the
snape of the dose-response curve is unknown; uncertainty about
the relevance of availaole biological data with respect to
humans; uncertainty of species differences in sensitivity to
carcinogens; and uncertainty about the relationship between
exposure level and biological effective dose.
Limitations of Current Approaches to the Estimation of the Car-
cinogenic Risks of PAHs
As stated above, in estimating carcinogenic risks, heavy
reliance is placed on human epidemiologic studies and the re-
sults of long-term animal carcinogenesis testing. Since, in the
case of specific PAHs, human data are not available, reliance
must be placed on long-term animal studies. However, of the 30
nonsubstituted PAHs evaluated, only benzo[a]pyrene, benz[a]an-
threne and dibenzo[a,h]anthracene have been tested by the oral
or respiratory route. Benzo[a]pyrene, because it has received
the most thorough study, has frequently been used as a surrogate
21

-------
potential carcinogenic RISKS OF PAHs
PAH in the estimation of carcinogenic risks of mixtures contain-
ing DAHs . This approach has added uncertainty in that other
PAHs can be either more or less potert than benzo[a]pyrene.
Uncertainties also exist with relation to tne potential chemi-
cal , biocheTical, and toxicologic interactions among components
in the mixture.
The Tajor routes of human exposure to PAHs are via the gas-
troi ntesti nal tract and the respiratory tract (26). However,
animal experiments have been nerforned with skin-painting and
subcutaneous injection as routes of administration. The vali-
dity of extrapolating from one route of administrati on to
another is uncertain. The relative potency approach presented
in this paoer seems reasonable, particularly for ranking the
hazards of the different PAHs. Its applicability to setting
permissible exposure levels, however, is highly uncertain.
A Forward Look at Potential Alternatives
The magnitude of the limitations and uncertainties rela-
ting to the form of the dose-response curve can be easily seen
in reference to the inability of the 1 EDqi " oioassay experiment
with 2-AAF to resolve these problems. The current approach
emphasizes the use of animal bioassay data; short-term test
results take a supportive role.
However, short-term j_n vi tro biochemical and toxicologic
studies could provide information about mechanisms of actio0,
metabolic pathways, nacromolecular (DNA, RNA, and protein) ad-
duct formation, DNA repair, and cell proliferation. Sone of
these studies can be performed at dose levels much closer to
potentia- human exposures, and all can be performed in shorter
time periods than the conventional carcinogenesis bioassay.
Additional research should be performed to determine if short-
tern tests are viable alternatives.
(1) Mutagenesis Tests. The attractiveness of _in_ vitro tests
as potential alternatives for long-term carcinogenesis tests is
demonstrated by the proliferation of the mutagenesis test. The
genetic toxicology (Gene-Tox) program of EPA has conducted panel
reviews of the validity of the different test systems. The re-
sults are published in several volumes of "Mutation Research."
The International Commission for Protection Against Environmen-
tal Mutagens and Carcinogens (21) has reviewed the potential of
using mutagenesis as an indicator of carcinogenesis. Recently,
the National Toxicology Program convened an ad-hoc panel to
review their testing procedure, including the potential uses of
the different short-tern mutagenesis tests and macromolecular
22

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
binding studies (27). 3artsch, Tomatis, ana Malaveille (4)
reviewed the literature and qualitatively and quantitatively
compared~mutagenic and carcinogenic activities of chemicals. It
could be concluded that currently mutagenesis tests are appro-
priate both for screening chemicals under development and for
screening existing chemicals for further testing. They could be
used in biological monitoring of exposure to mutagenic carciino-
gens. Qualitatively, carcinogenesis and mutagenesis seemed to
correlate wel 1 for a few PAHs (anthracene, benz[a]anthracene,
oenzo[a]pyrene, diDenzo[a,h]anthracene) that are adequately stu-
died in both systems. Further research is needed to fill data
gaps before potential quantitative relationships can be esta-
blished- The lack of quantitative correlation can be illus-
trated by the results of the two PAH studies that are reviewed
below.
Combs et a!. (7) measured tne liver microsome-mediated
mutagenicity of 35 PAHs which are derivatives of cyclopentaphen-
anthrene and chrysene, using Aroclor-pretreated rats and S.
typhimurium TA100 strain. The results were compared with car-
cinogenicity expressed as Iball indices, using results from
skin-painting experiments in mice. The authors reported little
quantitative correspondence between mutagenic activity and car-
cinogenic potency. However, Huberman and Sachs (13) found that
the carcinogenicity of 10 PAHs paralled .their mutagenicity,
measured as fl-azaquirnine or ouabain resistance, in cell-media-
ted mutagenicity assays on Chinese hamster V79 cells co-culti-
vated with lethally irradiated rat embryo cells for metabolic
activation.
(2) Pharmacokinetics and Metabolism. Pharmacokinetic studies
(both experimental and computational) and metabolism studies
provide parameters for deriving a human biological effective
dose from exposure data. These studies can be performed at
lower doses and more dose levels and dose-rate schedules than
the carcinogenesis bioassay, and can provide important insights
about dose-response behavior at low exposure levels in humans.
In 1980, Anderson et al. (2), published a general scheme for
the incorporation of pharmacokinetics in low-dose risk estima-
tion for chemcial carcinogenesis. However, adequate data for
quantitative analysis are usually unavailable. Recently, Ram-
sey and Anderson (34) developed physiologically based pharmaco-
kinetic models. These models may have promise for estimating
human tissue dose.
Metabolic studies on PAHs are numerous. E.C. Miller (23)
first reported evidence of metabolic activation of PAHs in 1951.
Sne found covalent binding'of metabolites of benzo[a]pyrene when
23

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
that substance was applied to mouse skin. Since then, many
investigations of PAH metabolites, particularly metabolites
benzo[a]pyrene, and their ability to bind DNA, RNA, and protein,
have been reported. Cooper et al > (9) have extensively re-
viewed tne metabolism and activation of benzo[a]pyrene. Other
reviews also contain sections on the metabolism of PAHs (8, 26,
33, 36). A reading of these reviews wi1• show the complexity
of the metabolic processes involved in the activation of a PAH
to its ultimate carcinogenic intermediate. Because of this
complexity, mathematical modeling studies of PAH metabolites as
a function of exposure levels are not available. With the
physiologica11y based model developed by Anderson and Ramsey
(34) and Hoel et al . (11), it may be useful to re-review the
metabolic and pharmacokinetic studies on benzo[a]pyrene and
investigate the feasibility of such modeling.
(3) Macromolecular Binding. Macromolecular birding, particu-
larly to DNA, as a quantitative indicator in the process of
chemical carcinogenesis, has been most extensively reviewed by
Lutz (22).
Carcinogen metabolite DNA binding provides the most direct
measure of the biological effective dose, Carcinogen-DNA adduct
can be an.effective biological monitoring tool for human expo-
sure. Factors limiting such use are the small quantity of DNA
present in cells and uncertainty as to the appropriate tissue
or body fluid to be used for monitoring. An additional limita-
tion is that this measure is only useful for studying carcino-
gens that act through a genotoxic mechanism.
The importance of DNA binding for assessing the carcino-
genicity of PAHs to mouse skin was suggested by Brooks and Law-
ley in 196^ (6). They found a correlation between the carcino-
genicity of several PAHs to mouse skin and the covalent binding
of these hydrocarbons on mouse skin DNA. This finding is
supported by the results of Goshmand and Heidelberger (10) witn
10 PAHs on DNA binding and carcinogenesis in mouse skin. Since
then, numerous studies have demonstrated benzo[a]pyrene metabo-
lite-DNA adduct formation in different in vi tro animal and human
systems (for review see Conney [8], Cooper [9J, Perera [30], and
Selkirk [17]).
Analytical techniques are available to quantify benzo[a]py-
rene metabolite-DNA adducts by irmunologic methods (12, 31, 32).
Monoclonal antibodies were developed to benzo[a]pyrene dioi-
epoxide-DNA adducts (33). The use of these antibodies and immu-
noassays makes possible the detection of femtonole levels of
carcinogen in pg quantities of DNA (i.e., 1 adduct/10^ nucleo-
tides) (12, 35).
24

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
At this level of sensitivity, human tissue or body fluids can be
monitored for exposure.
In summary, while molecular biochemical studies cannot
currently replace conventional toxicologic testing, they provide
information that can be used for better estimates of biological
effective dose and dose-response cnaracteristics at low levels
of exposure, and for improvement in interspecies scaling.
In spite of these improvements, it would be unrealistic to
expect the development of a complete set of risk assessment in-
formation for all chemicals in the near future. What is possi-
ble is the development of complete sets of information for
specific exposures that are well studied in both humans and
animals. These examples could be used to derive a more realis-
tic set of quantitative parameters for chemicals predicted to
have similar carcinogenic mechanisms. Benzo[a]pyrene should be
a good model compound for developing a set of parameters for the
assessment of other carcinogenic PAHs.
25

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
REFERENCES
1.	Anderson, E.L., and the Carcinogen Assessment Group of the
U.S. Environmental Protection Agency (1983): Quantitative
approaches in use to assess cancer risk. Risk Analysis,
3:277-295.
2.	Anderson, M.W., Hoel, D.G., and Kaplan, N.L. (1980): A
general scheme for the incorporation of pharmacokinetics in
low-dose risk estimation for chemical carcinogenesis: Exam-
ple—vinyl chloride. Toxicol. Appl. Pharmacol., 55:154-
161.
3.	Armitage, P., and Doll, R. (1954): The age distribution
of cancer and a multistage theory of carcinogenesis. Br.
J. Cancer, 8:1-12.
4.	Bartsch, H., Tomatis, L., and Malaveille, C. (1982): Qual-
itative and quantitative comparisons between mutagenic and
carcinogenic activities of chemicals. In Mutagenicity:
new horizons in genetic toxicology (Heddle, J.A., ed.).
Acadenic Press, New York, NY.
5.	3ingham, E., and Falk, C. (1969): The modifying effects
of carcinogens on the threshold response. Arch. Environ.
Health, 19:779-783.
6.	Brooks, P., and Lawley, P.D. (1964): Evidence for the
binding of polynuclear aromatic hydrocarbons to the nucle-
ic acids of mouse skin: relation between carcinogenic power
of hydrocarbons and their binding to deoxyribonucleic
acid. Nature (London), 202:781-784.
7.	Coombs, M.M., Dixon, C., and Kissonerghis, A.M. (1976):
Evaluation of the mutagenicity belonging to the benz[a]
anthracene, chrysene and cyclopenta[a]phenanthrene series,
using Ames's test. Cancer Res., 36:4525-4529.
8.	Conney, A.H. (1982): Induction of microsomal enzymes by
foreign chemicals and carcinogenesis by polycyclic aromatic
hydrocarbons: G.H.A. Clowes Memorial Lecture. Cancer Res.,
42:4875-4917.
9.	Cooper, C.S., Grover, P.L., and Sims, P. (1983): The me-
tabolism and activation of benzo[a]pyrene. In Progress in
Drug Metabolism, Vol. 7 (Bridges, J.W., and Chasseaud,
L.F., eds.). "John Wiley and Sons, New York, NY, pp.
295-396.
10.	Goshman, L.M., and Heidelberger, C. (1967): Binding of
tritium-labeled polycyclic hydrocarbons to DNA of mouse
skin. Cancer Res., 27:1678-1688.
11.	Hoel, D.G., Kaplan, N.L., and Anderson, M.W. (1983): Im-
plication of nonlinear Kinetics on risk estimation in
carcinogenesis. Science, 219:1032.
12.	Hsu, I.C., Poirier, M.C., Yuspa, S.H., Grunberger, D.,
Weinstein, I.B., Yolken, R.H., and Harris, C.C. (1981).
26

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
Measurement of benzo[a]pyrene-DNA adducts to enzyme immuno-
assays and radioimmunoassays. Cancer Res., 41:1090-1095.
13.	Huberman, E., and Sachs, L. ( 1977): DNA binding and its
relationship to carcinogenesis by different polycylcic
hydrocarbons. Int. J. Cancer, 19:122-127,
14.	Iball, J. (19357= The relative potency of carcinogenic
compounds. Am. J. Cancer, 35:188-190.
15.	Interagency Regulatory Liaison Group (IRLG) (1979): Sci-
entific basis for the identification of potential carcino-
gens and estimation of risks. J. Natl. Cancer Inst., 63:
243-268.
16.	Interdisciplinary Panel on Carcinogenicity (1984): Cri-
teria for evidence of chemical carcinogenicity. Science,
225:682-687.
17.	International Agency for Research on Cancer (IARC) (1973):
Certain polycyclic aromatic hydrocarbons and heterocyclic
compounds. [ARC Monographs on the Evaluation of the Car-
cinogenic Risk of Chemicals to Humans, Vol. 3. Lyon,
France.
18.	International Agency for Research on Cancer (IARC) (1982):
Chemicals and industrial processes associated with cancer
in humans. IARC Monographs on the Evaluation of the
Carcinogenic Risk of Chemicals to Humans, Supplement 4.
Lyon, France, pp. ~T-2Tt
19.	International Agency for Research on Cancer (IARC) (1983):
Polynuclear aromatic compounds, Part 1: Chemical, environ-
mental , and experimental data. IARC Monographs on tne
Evaluation of the Carcinogenic Risk of Chemicals to Humans,
Vol. 32. Lyon, France.
20.	International Agency for Research on Cancer (IARC) (1984):
Polynuclear aromatic compounds, Part 2: Carbon blacks, mi-
neral oils, and some nitroarenes. IARC Monographs on the
Evaluation of the Carcinogenic Risk of Chemicals to Humans,
Vol. 33. Lyon, France.
21.	International Commission for Protection Against Environmen-
tal Mutagens and Carcinogens (ICPEMC) (1982): Mutagenesis
testing as an approach to carcinogenesis. Committee 2 Fi-
nal Report, ICPEMC Document 43-1982-3.3.
22.	Lutz, W.K. (1979): _In_ vivo covalent binding of organic
chemicals to DNA as a quantitative indicator in the process
of chemical carcinogenesis. Mut. Res., 65:289-356.
23.	Miller, E.C. (1951): Studies on the formation of protein-
bound derivatives of 3,4 benzpyrene in the epidermal frac-
tion of mouse skin.- Cancer Res., 11:100-108.
24.	National Cancer Institute (NCI) (1976): Guidelines for
carcinogen bioassay in small rodents. Carcinogenesis Tech-
nical Report Series, No. 1. DHEW Pub. No. (NIH)76-801.
27

-------
4
POTENTIAL CARCINOGENIC RISKS OF PAHs
25.	National Research Council (1983): Risk assessment in the
Federal Government: managing the process. Prepared by the
Committee on the Institutional Means for Assessment of Risk
to Public Health, Commission on Life Sciences. National
Academy Press, Washington, D.C.
26.	National Research Council (1983): Polycyclic aromatic
hydrocarbons: evaluation of sources and effects. Prepared
by the Committee on Toxicology and Environmental Health
Hazards, Commission on Life Sciences. National Academy
Press, Washington, D.C.
27.	National Toxicology Program (NTP) (1984): Report of the
ad hoc panel on chemical carcinogenesis testing and evalu-
ation. National Toxicology Program Board of Scientific
Counselors. U.S. Government Printing Office, 1984-421-
132:4726.
28.	Neal, J., and Rigdon, R.H. (1967): Gastric tumors in mice
fed benzo[a]pyrene: a quantitative study. Texas Reports on
Biology and Medicine, 25:553-557.
29.	Office of Science and Technology Policy (OSTP) (1984):
Chemical carcinogens: review of the science and its asso-
ciated principles. Federal Register, 49:21595-21661.
30.	Perera, F., and Weinstein, I.B. (1982): Molecular epidemi-
ology and carcinogen-DNA adduct detection: new approaches
to studies of human cancer causation. J. Chron. Qis., 35:
581-600.
31.	Perera, F.P., Poirier, M.C., Yuspa, S.H., Nakayama, J., Ja-
retzki , A., Curnen, M.M., Knowles, O.M., and Weinstein, I.
B. (1982): A pilot project in molecular cancer epidemio-
logy: determination of benzo[a]pyrene-UNA adducts in ani-
mals and human tissues by immunoassays. Carcinogenesis,
3:1405-1410.
32.	Poirier, M.C., Santella, R., Weinstein, I.B., Grunberger,
0.,	and Yuspa, S.H. (1980): Quantitation of benzo[a]py-
rene deoxyguanosine adducts by radioimmunoassay. Cancer
Res., 40:412-416.
33.	?ott, F., Oberdorster, B., Jacob, J., and Grimmer, G.
(1983): Behavior of PAH in the organism. In Environmental
Carcinogens: Polycyclic Aromatic Hydrocarbons, Chemistry,
Occurence, Bio- chemistry, Carcinogenicity' ["Grimmer, G.,
ed.). CRC Press.
34.	Ramsey, J.C., and Anderson, M.E. (1984): A physiological-
ly based description of the inhalation pharmacokinetics of
styrene in rats and humans. Toxicol. Appl. Pharmacol. 73,
1:159-175.
35.	Santella, R.M., Lin, C.D., Cleveland, W.L., and Weinstein,
1.B.	(1984): Monoclonal antibodies to DNA modified by a
benzo[aJpyrene diol expoxide* Carcinogenesis, 3:373-377 .
36.	Santodonato, J., Howard, P., and Basu, 0. (1981): Health
28

-------
POTENTIAL CARCINOGENIC RISKS OF PAHs
and ecolcgica* assessment of polynuclear aromatic hydrocar-
bons . ^pec- aIssje, Envi ron. Pathol. Toxicol., Vol . 5,
No.'l. (Lee, S.D., and Grant, L., eds.).
3?. Se 1 <, J.<., \in explant cultures of human bladder, skin, bronchus
and esoohar:us ""ron eight individuals. Cancer Letters, 18:
11-19.
38.	Tonatis, L. (1979): The predictive value of rodent carci-
nogeneses tests in the evaluation of human risks, Ann.
Rev, Pharmacol. Toxicol., 19:511-530.
39.	Thyssen, J., Alt.noff, J., Kimmerle, G., and Mohr, U.
(1931): Imalation studies with benzo[a]pyrene in Syrian
go'den hamsters. J NCI, Vol. 66, 3:575-577.
4'J. ^.S. -nv1 ronmertal Protection Agency (EPA) (25 May 1976):
Irteri^ orocedures and guidelines for health risks and
econoTic impact assessments of suspected carcinogens.
Feceral Register, 41:21402.
41.	U.S. Env: ronmertal Protection Agency (EPA) (23 Nov. 1984):
Proposed guidelines for carcinogen risk assessment; re-
quest for comments. Federal Regi ster, 49:46294-46301,
?a^t 7.
42.	Uhittemore, A., anc Keller, A. (1978): Quantitative theory
of carcinogenesis. Society for Industrial and Applied
Mathenati cs Kevew, 20:1-30.
43.	W'ni tte-rore, A. (1978): Quantitative theories of carcinoge-
nesis. Acvances in Cancer Research, Vol. 27. Academic
^ress , New York , NY.
4-. Wynrier, E.L., Fritz, L., and Furth, N. (1957): Effect of
concentrat'on of benzopyrene in skin carcinogenesis. JNCI„
19:361-370.
£b. "wynder, E. _., and Hoffman, D. (1959): A study of tobacco
carcinogeresis. VII. The role of higher polycyclic hydro-
carDcrs. Career, 12:1079-1086.
^6. 'Ay n d e r, E.L., ar-d Hoffman, D. ( 1959): The carcinogenicity
of benzyf1uorantnenes. Cancer, 12:1194-1199.
29

-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing;
1 REPORT NO. 2.
F.PA/600/n-fiQ/04q
3 RECIPIENT S ACCESSION NO
PR89-221329
4. TITLE AND SUBTITLE
The Evaluation and Estimation of Potential Carcinogenic
5 REPORT DATE
Jaruarv 1985
Risks of Polynuclear Aromatic Hydrocarbons (PAH)
6. PERFORMING ORGANIZATION COOE
EPA/600/021
7. AUTHOfl(S)
Margaret M.L. Chu and Chao W. Chen
8 PERFORMING ORGANIZATION REPORT NO.
0HEA-C-147
9. PERFORMING ORGANIZATION NAME AND ADORESS
Human Health Assessment Group
Office or Health & Environmental Assessment (RD-689)
U.S. Environmental Protection Agency
Washington, DC 20460
10. PROGRAM fc LEMENT NO.
11 CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT ANO PERIOD COVERED
Symposium Paper
14. SPONSORING AGENCY CODE
EPA/600/21
15. SUPPLEMENTARY NOTES
Presented at the 1984 International Chemical Congress of Pacific Basin Societies,
16. ABSTRACT
The evaluation and estimation of the potential risk of human exposures to a
hazardous substance requires the analysis of all relevant data to answer two
questions (1) does the agent cause the effect; (2) what is the relationship between
dose (exposure) and incidence of the effect in humans? For polynuclear aromatic
hydrocarbons (PAH), carcinogenicity is the effect of concern. Three types of evidence
can be used to evaluate the likelihood that a PAH is carcinogenic to humans. They
are (1) epidemiologic evidence, (2) experimental evidence derived from long-term
animal bioassays, (3) supportive or suggestive evidence from short-term tests,
metabolism, pharmacokinetics and structure-activity correlations. Mathematical
modeling can be used to estinate the potential human risks. The approaches and the
problems associated with these approaches for estimating cancer risk to humans are
addressed with special emphasis on problems related to PAH.
KEY WOHOS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b IDENTIFIE RS/OPEN EN0ED TERMS
c. COSATI Field/Group



18. DISTRIBUTION STATEMENT
Release to the public
19. SECUR:T y CLASS 'Tins Report)
Unclassified
21. no or pages
35
20. SECURITY CLASS (This page
Unclassified
22 PRICE
nc *
EPA Form 2220-1 (R«v. 4-77) previous edition is obso.ete
1	'CE

-------