c/EPA
United States
Environmental Protection
Agency
Environmental Criteria and
Assessment Office
Cincinnati OH 45268
Research and Development
EPA/600/M-86/016 Sept. 1986
ENVIRONMENTAL
RESEARCH BRIEF
The STARA Toxicity Data Base
C. B. Farren1 and R. C. Hertzberg2
Purpose
A toxic chemicals data base has been created by the U.S.
Environmental Protection Agency's (EPAs) Environmental
Criteria and Assessment Office-Cincinnati (ECAO-Cin) to
aid in the development of risk assessment methodology and
to facilitate the evaluation of potential public health dangers
due to uncontrolled hazardous waste site releases and
chemical spills. This data base, "Studies on Toxicity
Applicable to Risk Assessment" (STARA), focuses on
toxicity studies containing quantitative as well as descrip-
tive information on a test animal or human study group,
exposure and type of effects. For each chemical in the data
base a toxicity summary table can be generated. A
discussion of the STARA data base is presented featuring
the types of information available, methods for revision and
expansion, and future uses of the system.
Background
The design and implementation of ECAO's data base has
been an ongoing program since the summer of 1982.
Initially organized as a short-term research project to assist
the implementation of the Comprehensive Environmental
Response, Compensation and Liability Act of 1980 (Super-
fund), the impetus was to: (1) i nvestigate toxic responses to
certain chemicals in animals and humans; and (2) assemble
such data in a methodical fashion so as to be easily
accessed should emergency contaminations arise. Experi-
mental studies cited in the data base are drawn from
searches of peer-reviewed scientific publications similar to
the searches conducted for the Ambient Water Quality
Criteria Documents, Health Assessment and Drinking
'C. Brigitte Farren is now with the Program Evaluation Division; Office of
Policy, Planning and Evaluation, USEPA, Washington, DC 20460.
2Richard C. Hertzberg is with the Environmental Criteria and Assessment
Office, USEPA, Cincinnati, OH 45268.
Water Documents, Reportable Quantities (RQs) and Health
and Environmental Effects Profiles (HEEPs). Currently, the
STARA data base contains animal toxicity data on nearly
200 chemicals and detailed epidemiologic data on 30
chemicals. These chemicals are listed at the end of this
brief.
Requests for situation-specific assessments or other
technical assistance occur irregularly and often involve
repetitive retrieval of toxicity information on a variety of
chemicals. The traditional procedure has been to manually
extract and compile the desired data from various "hard
copy" sources (research articles, review documents) on a
case-by-case basis as the need arose. This approach was
deemed outdated and inappropriate on the basis of
economy, efficiency and even accuracy. The logical solution
was to compile this bulk of information into some form of
computer accessible data base.
After thoroughly investigating the existing data base
management systems, it was concluded that no one specific
system could satisfy the particular requirements unique to
the Superfund mandate under which ECAO-Cin then
operated. Work was initiated to develop a format which
would be structured to allow for reproducibility and access
by varied users, yet flexible enough for expansion and
integration with other data processing systems.
Transposing the large, diverse documents and research
articles used by the EPA into an effective and uniform
source of information without sacrificing the integrity of the
original material wasa major task. Toxicity studies typically
report a large number of variables, many of which are
imprecisely defined or subject to considerable scientific
interpretation. To ensure the best possible evaluations of
such data for risk assessments, as much information as
possible must be retained in the computer files. The data
-------
Chemical Data. Toxicity Tables in the STARA Data Base
Acenaphthene
Acetone
Acetonitrile
Acrolien
Acrylamide
Acrylonitrile
Aldicarb
Aldrin
Ally! alcohol
Aluminum
Ammonia
Antimony
Arsenic
Asbestos
Barium
Benzo(a)pyrene
Benzene
1,2-Benzenedicarboxylic acid,
dibutyl ester
1,2-Benzenedicarboxylic acid,
diethyl ester
Benzldine
Beryllium
Bis(2-chloroisopropyl)ether
Bis(2-chloroethyl)ether
Bis{chloromethyl)etner
Bismuth
Boron
Bromodichloromethane
Bromomethane
1,3-Butadiene
Cadmium
Captan
Carbon disulfide
Carbon tetrachloride
Chlordane
Chlorinated naphthalene
Chlorine
2-Chloro-1,3 butadiene
Chlorobenzene
Chlorodibromomethane
2-Chloroethyl vinyl ether
Chloroform
Chloromethane
Chloromethyl methyl ether
Chloronitrobsnzene
Chlorophenol (m-, p-)
p-Chlorophenol
2-Chlorophenol
Chloropropenes
Chlorotoluenes
Chromium
Chrysene
Copper
Cresols
Creosote
Cyanides
Cyclohexanone
Cyclopentadiene
DDT
Demeton
Dibenzofurans
Dibromochloropropa ne
1,2-Dibromoethane
Dichlorobenzene
Dichlorobenzidine
Dichlorobutenes
Dichlorodifluoromethane
1,1 -Dichloroethane
1,2-Dichloro'ethane
D ich loroethyle nes
Dichloromethane
2,4-Dichloro'phenol
2,4-Dichlorophenoxyacetic acid
Dichloropropane
Dichloropropane/Dichloropropene
Dieldrin |
Diethylamine
Dimethylamjne
2,4-Dimethylphsnol
1,3-Dinitrobenzene
4,6-Dinitro-p-cresol
2,4-Dinitrophenol
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Dioxin (TCDD)
Diphenylhydrazine
Endosulfan
Endrin
Epichlorohydrin
Ethylbenzenje
Ethylene oxide
Fluoranthene
Fluorides
Formaldehyde
Guthion
Haloethers
Heptachlor
Hexachlorofcienzene
Hexachlorobutadiene
Hexachlorocyclohexane
Hexachlorocyclopentadiene
Hexachloroethane
Hexachlorophene
Isophorone
Isoprene
Kepone
Lanthanide Metals
Lead
Malathion
Manganese
Mercury
Methacrylonitrile
Methanol
Methoxychlor
Methyl ethyl ketone
Methyl isobutyl ketone
Methyl methacrylate
Methyl parathion
Mi rex
Monochlorobutanes
n-Propyl alcohol
Naphthalene
Nickel
Nitrites/Nitrates
Nitrobenzene
Parathion
Polybrominated biphenyls
Polychlorinated biphenyls
Penta, hexachlorodibenzo-p-dioxin
Pentachlorobenzene
Pentachloronitrobenzene
Pentachlorophenol
Phenol'
Phosphorus
Phthalate esters
Polynuclear aromatic hydrocarbons
Pyridine
Selenium
Silver
Tetrachlorobenzene
1,1,1,2-Tetrachloroethane
1,1,2,2-Tetrachloroethane
Tetrachloroethylene
2,3,4,6-Tetrachlorophenol
Tetraethyl lead (Plumbane)
Thallium
Toluene
Toxaphene
1,3-Transdichloropropene
Tribromomethane
Trichlorfon
2,4,6-Trichloroanaline
Trichlorobenzenes
1,1,1 -Trichloroethane
1,1,2-Trichloroethane
Trichloroethene(Trichloroethylene)
Trichlorofluoromethane
2,4,5-Trichlorophenol
2,4,5-Trichlorophenoxy acetic acid
Trichloropropanes
Trinitrobenzenes
Uranyl nitrate
Va nad i u m( v)oxide
Vanadyl sulfate
Vinyl chloride
4-Vinyl-1 -cyclohexene
Xylene
Zinc
-------
structure that was selected includes not only all the
measured information (body weight, daily dose, etc.), but
also space for qualitative descriptions of the study.
Equally important was the need to provide data that was
quickly accessible, either in whole (all information available)
or in part (selecting for a certain route or duration of
exposure, species type, etc.). ECAO-Cin has found this last
capability highly beneficial when responding to waste site
assessment questions, e.g., selecting only ingestion studies
for use in assessing groundwater contamination.
Other data bases such asTOXLINE andTDB are structured
for more efficient search strategies, but these are primarily
literature citations with a brief text summarizing the article.
STARA's uniqueness lies in its inclusion of all the available
toxicity data in a format which allows complete statistical
analysis, modeling and graphical presentations.
Database Development
The procedure for a toxicity table begins with the review
and evaluation of all relevant publications including
governmental, industrial and academic documents and
original research articles describing the toxicity of the
specific chemical. All useful dose-effect data are extracted
and encoded into tables according to set guidelines (Tables
1, 2, and 3). The data from these source tables are then
entered into files on the EPA's IBM computer system.
The time required to write and verify a toxicity table may
range from two weeks to several months, depending mostly
on the availability of original journal articles. Actual labor
time spent is less, usually aroung 7-10 working days per
toxicity table. The estimated cost to develop each chemical
table and related graph, including labor and literature
searches, is ~$500.00-$1000.00.
Graphic summaries of each toxicity table are generated by
plotting exposure levels vs. exposure duration and using a
symbol to represent the severity of the effect (Figure 1).
Statistical models to calculate human equivalent dose and
duration have been programmed into STARA so that data
on several species can be displayed on a single graph. In
Figure 1, for example, the equitoxic dose measure is mg per
kg body weight, and the equitoxic duration measure is
fraction of lifespan. Options in the plotting program allow
the user to display all data or to select a specific area of
interest (e.g., inhalation data, all acute oral data, etc.).
These graphs are being used in ECAO-Cin's Rapid Response
toxicity assessments and in evaluating various toxic
equivalence models.
Table 1. Abbreviations for Toxocity Table Categories.
Categories
DBS
CONT
ROUTE
SPECIES
N ANIMALS
BODWGHT
EXPLEVEL
EXPDUR
EXPSCH
STUDY
ORGAN
SEVERITY
REFERENCE
YEAR
COMMENTS
= Observation or record number
= Continuation item, part of the previous record
= Exposure route, or primary route if sequential
or simultaneous multiroute exposure
= Species of test animal
= Number of animals in dose group
= Body weight (kg), estimated average weight
over course of exposure period
= Exposure level in units reported by author
= Exposure duration
= Exposure schedule
= Purpose of study, main effect observed or
sought in the study
= Target organs
= S ubjective category of effect severity based on
EPA definitions in Table 3
First author reference
Year of reference
Comments
Options for Each Category
ROUTE:
D = Dermal, F = Diet, G = Gavage, I = Inhalation, T =
Intratracheal, 0 = Oral (not further specified), W = Water
ingestion, P = Intraperitoneal, V = Intravenous, C =
Subcutaneous, N = Not mentioned.
Options for Each Category (cont'd)
SPECIES:
CT = Cat, DG = Dog, GP = Guinea pig, HA = Hamster,
HU = Human, MD = Monkey, MS = Mouse, PI = Pig,
PR = Primate (unspecified), RB = Rabbit, RT = Rat, N =
Not mentioned.
EXPOSURE DURATION:
DY = Day, HR = Hour, LF = Lifetime, Ml = Minutes,
MO = Month, WK =Week, YR = Year.
EXPOSURE SCHEDULE:
EX = Exposures, HD = Hr/Dy,DW= Day/Week, N = Not
mentioned.
STUDY:
TX = Toxicity, IR = Irritation, CA = Cancer, RP =
Reproductive alteration, CATX = Cancer/toxicity.
TARGET ORGAN:
BL = Blood, BN = Bone, BR = Brain, Gl =
Gastrointestinal, GR = Growth/wt. gain, HT = Heart,
KD = Kidney, LV = Liver, LG = Lung, MT = Metabolism,
MC = Muscle, N = Not mentioned, NL = Nasal passage,
NS = Nervous system including CNS, CV = Nonspecific
cardiovascular, OT = Other organs described in comments,
RP = Reproductive system, SK = Skin, — = No effects
were noted.
SEVERITY:
CTRL = Control group; NOEL = No-observed-effect level;
NOAEL = No-observed-adverse-effect level; EL = Effect
level, not necessarily adverse; AEL = Adverse-effect level;
NOFEL = No-observed-frank-effect level; FEL = Frank-
effect level; NOCEL = No-observed-cancer-effect level;
CEL = Cancer-effect level; N = Not enough information-.
-------
**
CO
1<
CO
Ji
CM «?
CM 2
a
ffl
S
a
a
co
o
I
a
CO
CJ
£
III
REFERENCI
£
OC
LU
S
CO
<
cc
o
STUDY
i
o
CO
0.
ffi
cc
3
Q
Q-
S5
_J
LU
S
D_
X
III
1-
CD
1
O
m
I
z
z
CO
LU
U
LU
0.
CO
LU
1—
5
O
CC
|—
0
CJ
CO
m
o
—
HORIGUCH
_i
LU
U.
is
CO
z
>
3
x
Q
\
CC
CM
•*
S
CM
s
8
CM
00
z
h-
C
CM CM
r- CM
| |
O 2 o
X CO h-
uiujj
< U. LU
> 1—
_.z2
e&
ixx
CM «- T-
>-
Q
§CC OC
XX
i— •* CO
22,-
Q. Q. 2
0.0.0.
§O 0.
O O
o o
CM *— CO
Z co co
2 oc 2
CM
co «* in I I E dm
CO
m ! m m ^] z
(5SSS ff
iiii
r~ r~ r~ CD ^
Q Q O O o
IIII LO
•* •* •* CM T- :
>- >- >-
2 CD CO 2 CO >-
O> CM CM O) CO T— •
2 25
9-_9-o: 2 i
Jrt 2 zr in i Q- i
CO
CO
°8 °
CO
CO
LU
CO
LU
CD
z
o
o
CO
CO
05
LU
o
<
CC
OC
o
ti
O
UJ . S_ u ^-*
^zcotu ...x
1 I
.
— CO
<=
LU O
I 2 O
DC UJ ..
DC OC O-
j ^ c O
i|5
O S Q-
s2s
< CO -1-
>-C3
Q Z
CO
9
0-
< co 2 i
P LU ^ i
L?
Q
<
iHlll
-LU
.9So:
«=Z — C3"-JTT/-1
__ ..— ccccm-JS9
S:zl3BSS§z>-
gOfe<>:u,n, -- •
! Ouj I
X >
LU LU
fgp^i
cslo^pl!
: CO
CO
i O cu ~ i
!°^^;
!o!
• CO DC O5 LL. CO ==
CO
LO
: ii ^ UJ
^'S-^'F^cc'-'Eco^o-
0.irQ.x"J5"Jx3i 1- O CO >
I CO OC u- CO • I-
: 03 ^ o -i bi o
• < CO CO < LU <
CM CM a> cn «- «-
to CD r~ r~
o> co co o a> o
T- CM CM CM T- CO
O) CO CO O) O> O)
r-- co
in r~
co co
-------
c
o
U
N
_g
.0
IS
Ul
o
DC
Ul
UJ
cc
DC
UJ
111
CO
CD
DC
o
W
o
£
EXPDUR
EL
BODWGHT EXPL
ALS
NAN
SPECIES
O
CC
o
o
CO
m
o
I
o
CD — —
UJ O O O O
Q Z Z Z Z
DC -J -I DC
I- UJ Ul I-
o o o o
::>
o
O O
Z Z
-I 3:0:
oo oo
Z CO CD I
111 Ul UJ UJ
u. u- u. u.
DC
o
O O O O O
O O
O
in
CD
CM
CD
2 <3^
f ^CD
Q_ X *r
Si f n ^
s is
T- O «-
00 00
r- r»
CD
^CD
|«
«i
oo 2
CM O
CD
^
CD
CM
CO
Q a
in in
oo oo
r-- r-
^ CD
X H
CD X
1°
oo 2
o co
«- •*
5
lz
00 O
r- T-
oS
^ x
X CD
Sis
2o
co in
r~ co
UJ
CD
^
Ji
2 D-
O CO
in co
CM CM
z g§ §g g
i- woo co,_ H
DC 22 2 cc DC
— CD CD CD CD CD
o o
in in
50
N
z in
CD CD
CD 9: O —
*-cMco^-incor--oocoO'-cMco'*incor~cocoO'
CMCMCMCMCMCMCMCMCMCOCOCOOOCOCOCOCOCOCO'3-'
Z
UJ
O
o
CO
03
o
CM
CO
oo oo
CO CO
00 00
CO CO
00
CO
00 00
CO CO
CMCMCMINCMCMCMCMCMCOCOCOCOCOCOCOCOCOCO^t'*'t
to
3
c
c
o
o
II
I-
o
o _a
o'S
Z a
TE-g
O CO
8 £
DC 3
en en
•»-• o
(0 Q.
O X
" Ul
O II
II
> 0-
Si
m 3
O Q
-------
Tabla 3. Definitions of Effect Levels*
NOEL: No-Observed-Effect Level. That exposure level at w'hich
there are no statistically significant increases in' fre-
quency or severity of effects between the exposed
population and the appropriate control.
NOAEL: No-Observed-Adverse-Effect Level. That exposure level
at which there are no statistically significant increases
in frequency or severity of adverse effects between tne
exposed population and the appropriate control. Effects
are produced at this level, but they are not considered to
be adverse. '
EU
AEL:
NOFEL:
PEL
The exposure level in a study or group of studies which
produces statistically significant increases in frequency
or intensity of effects between the exposed population
and its appropriate control. It has not been decided
whether these effects are adverse.
NOCEL
CEL
CTRL:
Adverse-Effect Level. The exposure level in a study or
group of studies which produces statistically significant
increases in frequency or severity of adverse effects
between the exposed population and the appropriate
control. f
No-Observed-Frank-Effect Level. The study was directed
toward eliciting frank effects, but none were observed of
statistical significance. Other less severe toxic effects
may have been present but were not investigated, i
Frank-Effect Level. That exposure level which produces
unmistakable adverse effects or gross toxicity, such as
irreversible functional impairment or mortality, at a
statistically significant increase in frequency or severity
between an exposed population and its appropriate
control. }
No-Obsorved-Cancer-Effect Level. The study Was
directed toward eliciting carcinogenic response. No
such responses of statistical significance were observed
at this exposure level. Other toxic effects may have been
present but were not investigated. i
Cancer-Effect Level. Statistically significant cancer
responses were observed at this level. Significance
could be based on comparison with the control group or
on a significant dose-response trend using several dose
groups. :
i
Control group. No experimental exposure although a
background exposure may exist. I
•These designations only note the effects actually observed
and
reported by the research scientist. Levels where no effects were
observed (NOEL, NOAEL, NOFEL, NOCEL) do not ensure safety or
freedom from risk and may only reflect the limitations of the
study. ;
Applications of the STARA Data Base
The STARA data base is specifically designed for easy
access by statistical routines and mathematical modeling
programs. Thus, it is especially suitable for development
and testing of risk assessment algorithms and extrapolation
models. Because STARA is organized first by chemical, it is
also useful for rapid evaluation of a chemical's toxicity. The
graphical output in particular provides a ready tool for
determining how well an existing standard or criterion is
supported by the toxicity data.
Species Extrapolation of Dose. The frequent lack of
adequate human data forces the risk assessor to rely on
animal studies and use some type of extrapolation from
animal to man. The development of standard procedures for
dose extrapolation has been dramatically enhanced by the
STARA data base. An extrapolation model can be pro-
grammed and then automatically applied to hundreds of
chemicals with minimal effort, since the programs can
access the needed data directly from the computer files.
The behavior of the model can then be evaluated regarding
its general applicability to any chemical. Other issues that
can be similarly tested are the extrapolation from one route
of exposure to another, and the influence of aging on
toxicity.
Rapid Response Preliminary Health Hazard Assessments.
The Rapid Response toxicity assessment project at ECAO-
Cin was the first application of the STARA data base. This
project provides EPA Regional or Program offices with a
preliminary prediction of health hazards attributable to
contamination from spills or hazardous waste site releases.
These assessments are telephoned to the requestor within
two working days of the request and are often followed by a
longer written report within two to four weeks. Rapid
Response assessments address only toxic potential. No
judgments of the safety of a site nor recommendations for a
course of action are included in either the preliminary
assessment or the follow-up report.
The STARA data base has made projects such as the Rapid
Response preliminary site assessments not only possible
but practical as well. Before STARA was implemented, site
assessments, whether emergency or routine, were per-
formed in a similar and time-consuming fashion—sifting
through quantities of literature before finding pertinent
information. Response could take as long as several weeks,
which is not very useful in emergency cases but was the
best effort then available.
Now, however, specific data can be accessed for any
chemical within minutes. Comparisons between chemicals
may be made in any number of areas: target organs
attacked, type of length of exposure, reactions of different
species tested, and so on. Graphs are used to pinpoint
studies in relation to dosages, effect levels and other
distinctive characteristics. Human equivalent exposures
can be calculated in the STARA system and displayed
allowing direct comparison between monitored levels and
estimated toxic levels. All these features allow the risk
assessor to make several quantitative and judgmental
comparisons so that the assessment is based on as much
information as possible.
Conclusions
A practical solution for condensing large volumes of toxicity
data was found through the creation of the STARA data
base by the Environmental Criteria and Assessment Office
of the USEPA. The data base is designed for quantitative
investigations and has features not available in other
toxicity data systems. The system was planned in such a
way that modifications or expansions may be accomplished
without difficulty.
Efforts are now underway to incorporate STARA data into a
public access system. The National Library of Medicine and
NTIS are two such options being considered.
-------
Figure 1. Graphical display of alltoxicitydatafor1,1,2,2-tetrachloroethane. Equivalence: DOSE = mg/kg.DURATION = day, see
text. For severity categories, see Tables 2 and 3. Symbols: <^ NOEL, ^ IMOAEL, ^ AEL, H PEL
100000.0 ^
-3 1000.0
en
01
CO
O
a
a
10.0 -
0.1 -I
A
A
A
0
6hr
2.5 dy
3wk
8.5 mo
7yr
HU EQ DURATION
Equivalence: DOSE = mg/kg
DURATION = day
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268 i
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
------- |