United States
kS^laMIjk Environmental Protection
^J^iniiil m11 Agency
EPA/690/R-07/006F
Final
5-21-2007
Provisional Peer Reviewed Toxicity Values for
4-Chlorobenzotrifluoride
(CASRN 98-56-6)
Superfund Health Risk Technical Support Center
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
Acronyms and Abbreviations
bw body weight
cc cubic centimeters
CD Caesarean Delivered
CERCLA Comprehensive Environmental Response, Compensation and
Liability Act of 1980
CNS central nervous system
cu.m cubic meter
DWEL Drinking Water Equivalent Level
FEL frank-effect level
FIFRA Federal Insecticide, Fungicide, and Rodenticide Act
g grams
GI gastrointestinal
HEC human equivalent concentration
Hgb hemoglobin
i.m. intramuscular
i.p. intraperitoneal
IRIS Integrated Risk Information System
IUR inhalation unit risk
i.v. intravenous
kg kilogram
L liter
LEL lowest-effect level
LOAEL lowest-observed-adverse-effect level
LOAEL(ADJ) LOAEL adjusted to continuous exposure duration
LOAEL(HEC) LOAEL adjusted for dosimetric differences across species to a human
m meter
MCL maximum contaminant level
MCLG maximum contaminant level goal
MF modifying factor
mg milligram
mg/kg milligrams per kilogram
mg/L milligrams per liter
MRL minimal risk level
MTD maximum tolerated dose
MTL median threshold limit
NAAQS National Ambient Air Quality Standards
NOAEL no-ob served-adverse-effect level
NOAEL(ADJ) NOAEL adjusted to continuous exposure duration
NOAEL(HEC) NOAEL adjusted for dosimetric differences across species to a human
NOEL no-ob served-effect level
OSF oral slope factor
p-IUR provisional inhalation unit risk
p-OSF provisional oral slope factor
p-RfC provisional inhalation reference concentration
1
-------
p-RfD
provisional oral reference dose
PBPK
physiologically based pharmacokinetic
ppb
parts per billion
ppm
parts per million
PPRTV
Provisional Peer Reviewed Toxicity Value
RBC
red blood cell(s)
RCRA
Resource Conservation and Recovery Act
RDDR
Regional deposited dose ratio (for the indicated lung region)
REL
relative exposure level
RfC
inhalation reference concentration
RfD
oral reference dose
RGDR
Regional gas dose ratio (for the indicated lung region)
s.c.
subcutaneous
SCE
sister chromatid exchange
SDWA
Safe Drinking Water Act
sq.cm.
square centimeters
TSCA
Toxic Substances Control Act
UF
uncertainty factor
l^g
microgram
[j,mol
micromoles
voc
volatile organic compound
11
-------
5-21-2007
PROVISIONAL PEER REVIEWED TOXICITY VALUES FOR
4-CHLOROBENZOTRIFLUORIDE (CASRN 98-56-6)
Background
On December 5, 2003, the U.S. Environmental Protection Agency's (EPA's) Office of
Superfund Remediation and Technology Innovation (OSRTI) revised its hierarchy of human
health toxicity values for Superfund risk assessments, establishing the following three tiers as the
new hierarchy:
1. EPA's Integrated Risk Information System (IRIS).
2. Provisional Peer-Reviewed Toxicity Values (PPRTV) used in EPA's Superfund
Program.
3. Other (peer-reviewed) toxicity values, including:
~ Minimal Risk Levels produced by the Agency for Toxic Substances and Disease
Registry (ATSDR),
~ California Environmental Protection Agency (CalEPA) values, and
~ EPA Health Effects Assessment Summary Table (HEAST) values.
A PPRTV is defined as a toxicity value derived for use in the Superfund Program when
such a value is not available in EPA's Integrated Risk Information System (IRIS). PPRTVs are
developed according to a Standard Operating Procedure (SOP) and are derived after a review of
the relevant scientific literature using the same methods, sources of data and Agency guidance
for value derivation generally used by the EPA IRIS Program. All provisional toxicity values
receive internal review by two EPA scientists and external peer review by three independently
selected scientific experts. PPRTVs differ from IRIS values in that PPRTVs do not receive the
multi-program consensus review provided for IRIS values. This is because IRIS values are
generally intended to be used in all EPA programs, while PPRTVs are developed specifically for
the Superfund Program.
Because new information becomes available and scientific methods improve over time,
PPRTVs are reviewed on a five-year basis and updated into the active database. Once an IRIS
value for a specific chemical becomes available for Agency review, the analogous PPRTV for
that same chemical is retired. It should also be noted that some PPRTV manuscripts conclude
that a PPRTV cannot be derived based on inadequate data.
Disclaimers
Users of this document should first check to see if any IRIS values exist for the chemical
of concern before proceeding to use a PPRTV. If no IRIS value is available, staff in the regional
Superfund and RCRA program offices are advised to carefully review the information provided
1
-------
5-21-2007
in this document to ensure that the PPRTVs used are appropriate for the types of exposures and
circumstances at the Superfund site or RCRA facility in question. PPRTVs are periodically
updated; therefore, users should ensure that the values contained in the PPRTV are current at the
time of use.
It is important to remember that a provisional value alone tells very little about the
adverse effects of a chemical or the quality of evidence on which the value is based. Therefore,
users are strongly encouraged to read the entire PPRTV manuscript and understand the strengths
and limitations of the derived provisional values. PPRTVs are developed by the EPA Office of
Research and Development's National Center for Environmental Assessment, Superfund Health
Risk Technical Support Center for OSRTI. Other EPA programs or external parties who may
choose of their own initiative to use these PPRTVs are advised that Superfund resources will not
generally be used to respond to challenges of PPRTVs used in a context outside of the Superfund
Program.
Questions Regarding PPRTVs
Questions regarding the contents of the PPRTVs and their appropriate use (e.g., on
chemicals not covered, or whether chemicals have pending IRIS toxicity values) may be directed
to the EPA Office of Research and Development's National Center for Environmental
Assessment, Superfund Health Risk Technical Support Center (513-569-7300), or OSRTI.
INTRODUCTION
No RfD assessment for 4-chlorobenzotrifluoride is available on IRIS (U.S. EPA, 2007) or
in the Drinking Water Standards and Health Advisories list (U.S. EPA, 2004). The HEAST
(U.S. EPA, 1997) lists subchronic and chronic RfD values for 4-chlorobenzotrifluoride of 2E-1
and 2E-2 mg/kg-day, respectively. The source document for these assessments was a Health and
Environmental Effects Document (HEED) (U.S. EPA, 1988). Both reference dose (RfD) values
were based on a no-observed-adverse-effect level (NOAEL) of 15 mg/kg-day in a study that
exposed rats to 4-chlorobenzotrifluoride for 90 days after exposure in utero and during lactation
(EBL, 1981). The critical effect was tubular degeneration in the kidneys in rats treated with 40
mg/kg-day or more in a different subchronic study (Arthur and Probst, 1983). Uncertainty
factors of 100 and 1000 were used to derive the subchronic and chronic RfDs, respectively.
Other than the HEED discussed above, the Chemical Assessments and Related Activities
(CARA) list (U.S. EPA, 1991a, 1994a) does not include any relevant documents. The Agency
for Toxic Substances and Disease Registry (ATSDR, 2006) and the World Health Organization
(WHO, 2006) have not assessed the toxicity of 4-chlorobenzotrifluoride.
An RfC for 4-chlorobenzotrifluoride is not available on IRIS (U.S. EPA, 2007) or in the
HEAST (U.S. EPA, 1997). The HEED (U.S. EPA, 1988) reports that there were no pertinent
available inhalation toxicity data in humans or animals at the time of publication. Occupational
exposure limits for 4-chlorobenzotrifluoride have not been derived by the American Conference
for Governmental Industrial Hygienists (ACGIH, 2005), the National Institute for Occupational
2
-------
5-21-2007
Safety and Health (NIOSH, 2006) or the Occupational Safety and Health Administration
(OSHA, 2006).
A cancer assessment for 4-chlorobenzotrifluoride is not available on IRIS (U.S. EPA,
2007) or in the Drinking Water Standards and Health Advisories list (U.S. EPA, 2004). A
cancer weight-of-evidence classification for 4-chlorobenzotrifluoride is not listed in the HEAST.
The HEED (U.S. EPA, 1988) assigned 4-chlorobenzotrifluoride to U.S. EPA (1986) Group D
(not classifiable as to human carcinogenicity) based on lack of carcinogenicity data via any
exposure route. The carcinogenicity of 4-chlorobenzotrifluoride has not been assessed by NTP
(2006) or IARC (2006).
Literature searches were conducted from 1988 through October, 2005 for studies relevant
to the derivation of provisional toxicity values for 4-chlorobenzotrifluoride. Databases searched
included: TOXLINE/TOXCENTER (including BIOSIS and NTIS subfiles), MEDLINE,
CANCERLIT, TSCATS, RTECS, CCRIS, DART/ETIC, EMIC/ EMICBACK, HSDB,
GENETOX and Current Contents.
REVIEW OF PERTINENT DATA
Human Studies
Relevant data regarding the toxicity of 4-chlorobenzotrifluoride in humans were not
located.
Animal Studies
The toxicity of 4-chlorobenzotrifluoride following repeated oral exposure has been
studied by NTP (1992), Macri et al. (1987), Arthur and Probst (1983) and EBL (1981), and the
subchronic toxicity of 4-chlorobenzotrifluoride following inhalation exposure has been studied
by Newton et al. (1998). EBL (1981) also included a reproductive toxicity phase. All of these
studies are described below. No chronic toxicity studies via any exposure route were located,
and no studies that evaluated the developmental toxicity of 4-chlorobenzotrifluoride via any
exposure route were located.
Oral Exposure. 4-Chlorobenzotrifluoride was the subject of a 14-day study conducted by the
National Toxicology Program (NTP, 1992; Yuan et al., 1992). Male and female F344/N rats and
B6C3Fi mice (five per sex per dose per vehicle of each species) were given 4-
chlorobenzotrifluoride (97% pure) either suspended in a-cyclodextrin (a-CD) or dissolved in
corn oil via oral gavage daily for 14 days. The test substance was administered to rats and mice
as daily doses of 0, 10, 50 or 400 mg/kg suspended in a-cyclodextrin and as daily doses of 0, 10
(mice only), 50, 400 or 1000 mg/kg dissolved in corn oil. Additional groups of five male rats per
dose and vehicle were treated as described above, but used only for determination of a2U-globulin
and total protein levels in the kidneys. During the exposure period, all animals were observed
twice daily for signs of toxicity, and body weight was recorded every 2 days. All animals were
3
-------
5-21-2007
sacrificed after the 14th treatment. Prior to sacrifice, blood was obtained from all treated and
control animals. The following hematology parameters were evaluated (rats only): erythrocyte,
leukocyte and platelet counts, hematocrit, hemoglobin concentration, differential leukocyte count
and reticulocyte count. The following clinical chemistry parameters were evaluated in both rats
and mice: alanine aminotransferase (ALT), total protein, albumin, glucose, triglycerides,
cholesterol, urea nitrogen, creatinine, creatine kinase, alkaline phosphatase, 5' nucleotidase,
sorbitol dehydrogenase and total bile acids. At sacrifice, a complete necropsy examination was
performed; the liver, kidney (right), heart, lungs, thymus and testis (right) were weighed from all
treatment and control animals. A comprehensive set of tissues was microscopically examined
from all control and high-dose animals and animals that died prior to scheduled sacrifice by two
independent pathologists. Only gross lesions, liver, kidney and adrenal glands were
microscopically examined in the other rat dose groups, and only gross lesions and the liver were
microscopically examined in the other mouse dose groups. 4-Chlorobenzotrifluoride residues
were determined in the blood, kidney and liver in rats and mice from all treatment groups (both
vehicles). The kidneys were stained to detect the presence of hyaline droplets in the tubular
epithelium and lumen. Sections of the liver, kidney and adrenal glands of high dose and control
male and female rats were examined by electron microscopy for ultrastructural changes.
Rats showed minimal clinical signs indicative of irritation immediately following dosing
(burrowing in bedding, rubbing face with forepaws) with either vehicle (NTP, 1992). One
female rat treated with 1000 mg/kg-day in corn oil died after eight treatments. The cause of
death was not determined and no other animals died prior to scheduled sacrifice. Final mean
body weight was statistically significantly reduced in male rats treated with 1000 mg/kg-day in
corn oil; the difference from controls was 6%. Final body weight was similar to controls in the
other treated groups. Hematology analyses revealed mild anemia, characterized by slight,
statistically significant decreases in erythrocyte counts, hemoglobin and hematocrit, in male and
female rats treated with 1000 mg/kg-day in corn oil. The males also showed mild leukocytosis
due to increases in segmented neutrophils and monocytes. Minimal changes in some of these
parameters were present in the 400 mg/kg-day corn oil groups as well. Serum chemistry changes
occurred mainly in the 1000 mg/kg-day group (only corn oil tested) and occasionally in the 400
mg/kg-day groups (both solvents) and included increases in cholesterol, 5'-nucleotidase, bile
acids, triglycerides and total protein. Serum cholesterol levels were statistically significantly
increased in male rats (a-CD vehicle) at 50 and 400 mg/kg-day (17 and 37% higher than
controls), female rats (a-CD vehicle) at 400 mg/kg-day (35% higher), male rats (corn oil vehicle)
at 400 and 1000 mg/kg-day (51 and 134% higher) and female rats (corn oil vehicle) at 400 and
1000 mg/kg-day (58 and 119% higher). Serum 5'-nucleotidase activities were increased in male
rats (a-CD vehicle) at 50 and 400 mg/kg-day (10 and 15% higher than controls), male rats (corn
oil vehicle) at 400 and 1000 mg/kg-day (13 and 62% higher) and female rats (corn oil vehicle) at
400 and 1000 mg/kg-day (21 and 46% higher). Serum levels of bile acids were increased in
male rats (corn oil vehicle) at 1000 mg/kg-day (123% higher than controls) and female rats (a-
CD vehicle) at 400 mg/kg-day (13% higher). Serum triglyceride levels were increased in female
rats (corn oil vehicle) at 1000 mg/kg-day (71% higher than controls). Serum total protein levels
were increased in male rats (corn oil vehicle) at 400 and 1000 mg/kg-day (7 and 12% higher than
controls) and female rats (corn oil vehicle) at 400 and 1000 mg/kg-day (13 and 17% higher).
4
-------
5-21-2007
Serum alkaline phosphatase and sorbitol dehydrogenase were increased in male rats (corn oil
vehicle) at 1000 mg/kg-day (17 and 7% higher than controls).
The only organ weight changes of note were statistically significant increases in absolute
and relative liver weights at >50 mg/kg-day in males and >400 mg/kg-day in females, and
absolute and/or relative kidney weights at >400 mg/kg-day in males and 1000 mg/kg-day in
females (NTP, 1992). Organ weights at a given dose were similar for the different vehicles.
Gross pathology examinations revealed enlarged livers in males and females and enlarged
kidneys in males at 1000 mg/kg-day. Dose-related microscopic changes in the liver, kidneys and
adrenal glands occurred in groups administered the test substance in both vehicles (Table 1).
Incidence and severity of these changes increased with increasing dose. In the liver, increased
incidences of hepatocyte hypertrophy occurred in male rats at >50 mg/kg-day and in female rats
at >400 mg/kg-day. This effect was described as minimal to mild, consisting of enlargement of
hepatocytes in the centrilobular region (involving most hepatocytes within the lobule at the
highest dose). Fatty changes were noted in the liver of the 1000 mg/kg-day female rat that died.
Nephropathy was observed only in male rats. Severity progressed from minimal to mild at 50
mg/kg-day (primarily accumulation of protein droplets) to moderate at >400 mg/kg-day
(including degeneration, necrosis and regeneration of tubular cells). Electron microscopic
examination revealed focal mineralization along the basal lamina of renal tubules. Kidney levels
of a2u-globulin were significantly increased in a dose-related fashion in males treated with >50
mg/kg-day in either vehicle. Cytoplasmic vacuolation of the adrenal cortex occurred at >400
mg/kg-day in both males and females, although the cause and biological significance of the
adrenal lesion were not clear to the investigators.
The hepatocellular hypertrophy occurred at 50 mg/kg-day and higher dose levels, but is
not considered to be adverse because it is likely to reflect the proliferation of smooth
endoplasmic reticulum and the presumed induction of hepatic microsomal enzymes, which is an
adaptive response to chemical exposure. The increases in serum cholesterol, triglycerides, 5'-
nucleotidase and bile acids are indicators of cholestasis, generally appeared at 400 mg/kg-day,
and were most prevalent at 1000 mg/kg-day. Based on the serum chemistry changes suggestive
of cholestatic liver disease, the EPA identified a LOAEL of 400 mg/kg-day and NOAEL of 50
mg/kg-day for rats. The occurrence of adrenal cortex cytoplasmic vacuolation, a possible
adverse effect observed at >400 mg/kg-day, supports classifying 400 mg/kg-day as the study
LOAEL. The renal lesions observed in male rats at >50 mg/kg-day are not considered relevant
for human health risk assessment. The occurrence only in male rats, nature of the lesions, and
correlation with measured levels of a2u-globulin all indicate that the lesions result from
accumulation of a2U-globulin in the kidney, which is a toxic response peculiar to male rats and
not predictive of renal toxicity in humans (U.S. EPA, 1991b).
In mice, clinical signs of irritation similar to those seen in rats were observed (NTP,
1992). There were no deaths during the study and no effects on body weight. Serum chemistry
findings were generally similar to those in rats. The most notable changes were increases in
serum cholesterol, 5'-nucleotidase and triglycerides in both males and females. The changes
occurred primarily in the 400 mg/kg-day groups (more marked in the a-CD groups than the corn
5
-------
5-21-2007
Table 1. Summary of Microscopic Lesions in Rats Administered 4-Chlorobenzotrifluoride
by Oral Gavage for 14 Consecutive Days (NTP, 1992)
Tissue
Lesion
Vehicle
Dose (mg/kg-day)
0
10
50
400
1000
Males
Liver
Hepatocyte
hypertrophy
a-CD
0/5a'
0/5
2/5 (1.0)b
5/5 (1.0)
"
Corn oil
0/5
--
0/5
5/5 (2.0)
5/5 (2.0)
Kidney
Nephropathy
a-CD
0/5
0/5
5/5 (1.2)
5/5 (3.0)
--
Corn oil
0/5
--
5/5 (1.0)
5/5 (2.8)
5/5 (3.0)
Adrenal
cortex
Cytoplasmic
vacuolation
a-CD
0/5
0/5
0/5
5/5 (1.0)
--
Corn oil
0/5
--
0/5
4/5 (1.0)
5/5 (2.0)
Females
Liver
Hepatocyte
hypertrophy
a-CD
0/5
0/5
0/5
3/5 (1.0)
--
Corn oil
0/5
--
0/5
5/5 (1.0)
5/5 (2.0)
Adrenal
cortex
Cytoplasmic
vacuolation
a-CD
0/5
0/5
0/5
2/5 (1.0)
--
Corn oil
0/5
--
0/5
2/5 (1.0)
4/5 (1.0)
incidence (number with lesion/number examined)
bAverage severity is reported in parenthesis: 1.0 = minimal; 2.0 = mild; 3.0 = moderate; 4.0 = marked
oil groups) and 1000 mg/kg-day group (only corn oil tested). Serum cholesterol levels were
statistically significantly increased in male and female mice (a-CD vehicle) at 400 mg/kg-day
(26 and 36% higher than controls), and male and female mice (corn oil vehicle) at 1000 mg/kg-
day (84 and 65% higher). Serum 5'-nucleotidase activities were increased in female mice (a-CD
vehicle) at 400 mg/kg-day (30% higher than controls) and male and female mice (corn oil
vehicle) at 1000 mg/kg-day (166 and 159% higher than controls). Serum triglyceride levels
were increased in female mice (a-CD vehicle) at 50 and 400 mg/kg-day (47 and 53% higher than
controls), and female mice (corn oil vehicle) at 1000 mg/kg-day (78% higher than controls). The
only organ weight change of note was a 71% increase in liver weight in males and 55% increase
in females in the 1000 mg/kg-day corn oil group. No treatment-related gross lesions were
observed in males or females of any treatment group. The only microscopic change observed in
mice was hepatocellular hypertrophy, characterized by enlargement of the hepatocytes in the
centrilobular area, which was observed at >400 mg/kg-day in males and females (both vehicles).
All treated males at >400 mg/kg-day had this lesion. The incidence in females was 1/5 (a-CD)
and 4/5 (corn oil) at 400 mg/kg-day and 5/5 (corn oil) at 1000 mg/kg-day. Severity increased
from minimal to mild as the dose increased from 400 mg/kg-day to 1000 mg/kg-day, and was
generally greater in males than in females. The hepatocellular hypertrophy is not considered to
be adverse because it is likely an adaptive response to chemical exposure. On the basis of serum
chemistry changes suggestive of cholestatic liver disease, this study identified a LOAEL of 400
mg/kg-day and NOAEL of 50 mg/kg-day for mice.
6
-------
5-21-2007
Macri et al. (1987) administered 4-chlorobenzotrifluoride (99.3% pure) in olive oil to
Sprague-Dawley rats (6 per sex and dose) via oral gavage daily for 28 days at 0, 10, 100 or 1000
mg/kg. Clinical signs of toxicity were recorded at unspecified intervals. Body weight and food
and water consumption were recorded daily throughout the treatment period. Clinical chemistry
parameters evaluated included glucose, blood urea nitrogen, total proteins, triglycerides,
cholesterol, creatinine, bilirubin, serum electrolytes, alanine aminotransferase, aspartate
aminotransferase, creatine phosphokinase, lactate dehydrogenase and y-glutamyl transpeptidase.
Hematology parameters evaluated included total erythrocyte and leucocyte counts, hemoglobin
and differential leucocyte counts. All rats were sacrificed after 28 days of treatment, the liver,
kidneys, spleen, heart and adrenals were weighed. These organs and the brain, thymus, lungs,
thyroid, stomach, duodenum, pancreas, testes, ovaries and uterus were microscopically
examined.
No animals died prior to scheduled sacrifice (Macri et al., 1987). The only clinical sign
noted was an increase in the incidence of salivation during the last week of treatment in both
males and females treated with 1000 mg/kg-day. Body weights were similar to controls, except
for males treated with 1000 mg/kg-day. In this group, the rate of body weight gain was
significantly reduced and the final mean body weight was approximately 10% lower than
controls. No effects on food or water consumption were observed. Serum chemistry changes
included statistically significant, dose-related increases in serum cholesterol and triglyceride
levels in males but not in females at >100 mg/kg-day. Serum cholesterol levels at 10, 100 and
1000 mg/kg-day were 24, 62 and 75% higher than controls, respectively. Serum triglyceride
levels at 10, 100 and 1000 mg/kg-day were 19, 92 and 137% higher than controls, respectively.
Relative liver weight was significantly increased (p<0.01; not otherwise quantified) in males at
>100 mg/kg-day and females at 1000 mg/kg-day. Relative kidney weight was significantly
increased only in males (>100 mg/kg-day). Other organ weights were similar to controls. No
treatment-related gross lesions were observed at any dose. Microscopic lesions were observed in
the liver, kidney and adrenal cortex. In the liver, the only reported effects were slight to
moderate fatty changes in several treated and control animals (incidences not reported). Due to
the lack of relation to dose, the liver lesions were not attributed to 4-chlorobenzotrifluoride, but
reportedly could have been related to use of the oil vehicle. Severe hyaline droplet nephrosis,
often accompanied by dilated tubules and hypercellular foci, was observed in the proximal
convoluted tubules of the kidneys in all (6/6) male rats at 1000 mg/kg-day. Slight hyaline
droplet nephrosis with granular material was observed in 4/6 males at 100 mg/kg-day. Females
at 1000 mg/kg-day showed occasional, slight-to-moderate presence of intraluminal granular
material. In the adrenal cortex, marked fatty change (vacuolation of the cells of the zona
fasciculata) was observed in 5/6 males at 1000 mg/kg-day, but was not observed in females at
this dose or any animals in the lower dose groups. The authors noted that the adrenal cortex
vacuolation could indicate decreased adrenal activity and be consistent with the
hypercholesterolemia in the males. As previously discussed, the hyaline droplet nephrosis of the
kidney, and other effects induced by this compound in male rats, are consistent with a2U-globulin
accumulation and are not considered to be relevant to human health risk assessment. Based on
the observed hepatic changes (increases in serum cholesterol and triglyceride levels), this study
identified a LOAEL of 100 mg/kg-day and a NOAEL of 10 mg/kg-day in rats.
7
-------
5-21-2007
In a 3-month study (Arthur and Probst, 1983), groups of 15 F344 rats/sex were treated
daily by gavage with 4-chlorobenzotrifluoride (97.8% pure) in corn oil at doses of 0, 10, 40, 150
or 500 mg/kg-day. The following parameters were evaluated during the 90-day treatment
period: ophthalmology (at study initiation and near study termination); clinical signs of toxicity
and mortality (daily); detailed physical examination (weekly); and body weight, food
consumption and food utilization (weekly). After the treatment period, selected clinical
chemistry (serum glucose, urea nitrogen, creatinine, total bilirubin, alkaline phosphatase and
alanine transaminase), hematology (erythrocyte count, hemoglobin, PCV, total and differential
leukocyte counts, erythrocyte morphology, mean cell volume, mean cell hemoglobin and mean
cell hemoglobin concentration) and urinalysis (color, clarity, specific gravity, pH, protein,
glucose and blood) indices were evaluated. Liver p-nitroanisole o-demethylase activity was
determined from five rats per sex. After sacrifice, the adrenals, heart, kidneys, liver, ovaries,
prostate, spleen, testes, thyroids and uterus were weighed and organ:body weight ratios were
calculated. A complete gross examination was conduced at sacrifice, and a comprehensive set of
tissues was microscopically examined.
One male rat at 10 mg/kg-day and two male rats at 500 mg/kg-day died (Arthur and
Probst, 1983). No significant dose-related physical or behavioral signs or ophthalmic changes
were noted. Body weight gain was slightly, but statistically significantly depressed throughout
the study in males treated with 500 mg/kg-day. However, terminal body weight in this group
was only 5% lower than controls, and not statistically different. Body weights in other treated
groups were similar to controls throughout the study. Food consumption and efficiency of
utilization were generally similar to controls. Slight, statistically significant decreases in red
blood cell count, hemoglobin and hematocrit were noted in males at 500 mg/kg-day. A minimal
decrease in hematocrit was also present in males at 150 mg/kg-day. These effects were not seen
in females. Serum chemistry changes of note were significant increases in total bilirubin in both
males and females at 500 mg/kg-day (46 and 75% higher than controls, respectively) and slightly
elevated levels of serum urea nitrogen in male rats at 150 and 500 mg/kg-day. Alkaline
phosphatase was significantly elevated in males at all doses, but without a clear dose-response
relationship (41, 46, 46 and 40% higher than controls in low to high dose groups), and in females
at 500 mg/kg-day (43% higher than controls). Urinalysis indicated mild proteinuria in males at
500 mg/kg-day and females at >150 mg/kg-day. Hepatic p-nitroanisole o-demethylase activity
was significantly increased at >40 mg/kg-day in males and at >150 mg/kg-day in females.
Significant, dose-related increases in absolute and relative organ weight were observed for the
liver (>10 mg/kg-day in males and >40 mg/kg-day in females), kidney (>150 mg/kg-day in
males and females) and adrenals (>150 mg/kg-day in males and 500 mg/kg-day in females).
Relative liver weights in males were 8, 13, 36 and 90% higher than controls at 10, 40, 150 and
500 mg/kg-day, respectively. Relative adrenal weights in males were 3, 8, 17 and 50% higher
than controls at 10, 40, 150 and 500 mg/kg-day, respectively.
Treatment-related microscopic changes were observed in the liver and kidneys, but not in
the adrenals or other tissues (Arthur and Probst, 1983). In the liver, centrilobular hypertrophy
was observed in 15/15 males and 1/15 females at 150 mg/kg-day and in 13/15 males and 15/15
females at 500 mg/kg-day. This change was not observed at <40 mg/kg-day in either sex.
Severity increased from slight to moderate in males and from minimal to slight in females as the
8
-------
5-21-2007
dose increased from 150 to 500 mg/kg-day. Kidney lesions included renal tubular degeneration
in one low-dose male rat (10 mg/kg-day) and in all male rats dosed at >40 mg/kg-day. The
severity of the renal lesions was dose-related, ranging from minimal (decreased cellular height,
increased cytoplasmic basophilia, increased hyaline droplet formation) to moderate (increased
number of necrotic cortical epithelial cells and prominent hyaline casts in tubules of the outer
zone medulla and occasionally in the cortex). Tubular degeneration was not observed in female
rats at any dose or in male control rats. A general increase in the amount of colloid present in
the thyroid occurred with increasing dose; however, the toxicological significance of this
observation is not clear because no morphological changes in the thyroid were observed at any
dose. As previously discussed, kidney effects induced by 4-chlorobenzotrifluoride are consistent
with a2u-globulin accumulation and are not considered to be relevant to human health risk
assessment (U.S. EPA, 1991b). The increases in liver weight (>10 mg/kg-day), hepatic p-
nitroanisole o-demethylase activity (>40 mg/kg-day) and hepatocellular hypertrophy (>150
mg/kg-day), in the absence of liver histopathology, are indicative of enzyme induction and are
not considered by the EPA to be adverse (U.S. EPA, 2002). The increases in serum alkaline
phosphatase (>10 mg/kg-day) are not considered to be adverse because there was no clear dose-
response relationship and the magnitude of the changes is unlikely to be toxicologically
significant. Based on increased serum bilirubin in both sexes, this study, the EPA identified a
LOAEL of 500 mg/kg-day and NOAEL of 150 mg/kg-day in rats.
EBL (1981) reported results of a reproductive/90-day toxicity study with an in utero
exposure phase. Groups of 20 Sprague-Dawley rats/sex were treated daily by gavage with
4-chlorobenzotrifluoride (97% pure) in corn oil at doses of 0, 5, 15 or 45 mg/kg. The parental
generation was treated for 4 weeks before mating, throughout reproduction and through the
weaning of the Fi generation (total treatment duration of 76 to 83 days). The Fi generation was
culled tolO pups/litter on postnatal day (PND) 14. Twenty Fi rats were randomly selected from
each dose group on PND 21 and given the same doses as the F0 animals for 90 days post-
weaning. Mortality, clinical signs of toxicity, body weight and food consumption were recorded
over the treatment period. Fi animals were sacrificed 90 days after the start of treatment. Prior
to sacrifice, blood and urine were collected for hematology, clinical chemistry and urinalysis
parameters. Hematology and clinical chemistry parameters were also determined for F0 animals
2 weeks after study initiation. The clinical chemistry evaluations included the following liver
enzymes: serum glutamic-oxaloacetic transaminase (i.e., aspartate aminotransferase), serum
glutamic-pyruvic transaminase (i.e., alanine aminotransferase) and serum alkaline phosphatase.
After sacrifice, the heart, spleen, kidneys, liver, ovaries and testes were weighed, and organ to
body weight ratios were calculated. A complete gross pathology examination was conducted,
and a comprehensive set of tissues (including liver and adrenals) was microscopically examined
from control and high-dose Fi animals. Tissues from other dose groups were only
microscopically examined if gross lesions were present, and tissues from F0 animals were not
evaluated unless gross lesions were observed. Total number of pups born dead and alive per
litter, pup sex, weight, litter weight and pup survival were recorded.
No treatment-related mortalities, clinical signs or effects on weight gain or food
consumption were observed in the F0 or Fi rats (EBL, 1981). Treatment of rats with
4-chlorobenzotrifluoride had no effect on the number of pups/litter, pup survival or length of
9
-------
5-21-2007
gestation period. No clear treatment-related effects on pup body weight were observed. No
significant treatment-related changes in hematology or clinical chemistry were noted in F0 or Fi
rats. At necropsy, no dose-related gross lesions were observed in the F0 rats; therefore,
histological examinations were not performed on these animals. In the Fi generation,
determination of organ weights revealed a nonsignificant dose-related increase in the mean liver
weights and mean liver-to-body weight ratios in both sexes. Histological examinations of major
tissues and organs completed on Fi controls and on rats treated at 45 mg/kg-day did not reveal
any treatment-related effects. The NOAEL in this study was 45 mg/kg-day, the highest dose
tested. A LOAEL was not achieved.
Inhalation Exposure. Newton et al. (1998) exposed Charles River CD rats (25 per sex and
dose) via whole body inhalation to 4-chlorobenzotrifluoride (99% pure) vapor at 0, 10, 51 or 252
ppm (0, 74, 377 or 1860 mg/m3, respectively) 6 hours/day, 5 days/week, for 13 weeks. Selection
of the exposure concentrations was based on results of a 4-week preliminary inhalation study in
which 10 rats/sex/concentration were exposed to the test substance via whole body inhalation at
0, 100, 262, 494 or 1044 ppm (0, 740, 1940, 3650 or 7710 mg/m3, respectively) for 6 hours/day
and 5 days/week. Although few details regarding the preliminary study were specified in the
publication, the following parameters appear to have been evaluated: clinical signs of toxicity,
mortality, hematology, clinical chemistry and macroscopic and microscopic pathology. In the
13-week study, the following parameters were recorded during the exposure period: clinical
signs of toxicity (daily), body weight (weekly) and food consumption (weekly). All animals
were subjected to weekly detailed physical examinations over the 13-week exposure period.
Ophthalmic examinations were conducted prior to study initiation and 1 day prior to sacrifice.
Motor activity and a functional observational battery were evaluated in 10 animals per group
before the start of exposure, at weeks 4, 8 and 13 of exposure, and after a 13-week recovery
period (five animals per group). Hematology (blood count) and clinical chemistry (aspartate
aminotransferase, alanine aminotransferase, alkaline phosphatase, blood urea nitrogen, fasting
glucose, creatine kinase, total protein, albumin, globulin, albumin/globulin ratio, total bilirubin,
sodium, potassium, chloride, calcium and inorganic phosphorus) parameters were evaluated after
13 weeks of treatment from 10 animals per group. The brain, adrenals, kidneys, liver, lungs,
ovaries and testes (with epididymides) were weighed, a complete gross pathology evaluation was
performed and a comprehensive set of tissues was microscopically examined from 10 animals
per sex and dose at the end of the 13-week exposure and from five animals per sex and dose after
the 13-week recovery period. A comprehensive neuropathology examination was conducted on
an additional five animals per sex and dose sacrificed at the end of exposure and on five animals
per sex and dose after the 13-week recovery period; tissues other than the brain and spinal cord
were not examined from these animals. 4-Chlorobenzotrifluoride levels were determined in the
blood and tissues of three female animals at each sacrifice period.
No deaths occurred in the 4-week study (Newton et al., 1998). Dose-related increases in
activity were noted in both male and female rats at >262 ppm that seemed to become more
prominent as the daily exposure sessions progressed. In the detailed weekly observations,
hyperactivity and tremors were reported, but only at 1044 ppm. Hematology findings were
unremarkable. The serum chemistry analysis showed small but statistically significant increases
in serum protein in both sexes at 492 and 1044 ppm (5-6 and 11% higher than controls,
10
-------
5-21-2007
respectively). Similar increases in serum calcium and phosphorous occurred at >492 ppm, but
the effects could be related to the increase in serum protein because these chemicals are bound to
proteins in serum. Absolute and relative liver weights were significantly increased in a dose-
related fashion in males and females at >492 ppm. Absolute and/or relative kidney weights were
increased in males and females at >262 ppm, but the changes were slight and not directly related
to dose. Pathological examinations found dose-related increases in incidence and severity of
centrilobular hepatocellular hypertrophy in the liver of male and female rats at >100 ppm.
Incidence was 7/10 males and 0/10 females, 10/10 males and 5/10 females, 10/10 males and
10/10 females and 10/10 males and 10/10 females in the 100, 262, 492 and 1044 ppm groups,
respectively. From the write-up, it appears that the lesion was not found in controls, although
control incidence was not specifically reported. Severity increased from minimal at 100 ppm to
moderate at the higher doses. Male rats exposed to 262 ppm and above had intracytoplasmic
eosinophilic granules in the epithelium lining of the proximal convoluted tubules that were
shown to contain a2U-globulin. The LOAEL in this study is 1044 ppm based on clinical signs of
neurotoxicity. The NOAEL is 492 ppm because the liver effects are not considered to be
adverse; the hepatocellular hypertrophy is likely an adaptive effect reflecting the induction of
hepatic cytochrome P-450 (P450) enzymes (see discussion of the 13-week study below), and the
increases in serum proteins were small and unlikely to be clinically significant.
In the 13-week study, there were no deaths and no clinical signs were found during the
exposures or during the detailed weekly observations (Newton et al., 1998). Body weights were
unaffected and there was only a minimal effect on food consumption (6% decrease in the 252
ppm group during the first 2 weeks of the study). The neurobehavioral, neuropathology and
hematology evaluations were unremarkable. Clinical chemistry changes at the end of the
exposure period included small but statistically significant increases in serum total protein level,
albumin level and serum alanine aminotransferase (ALT) activity in females at 252 ppm (15, 19
and 40% higher than controls, respectively). No serum chemistry changes occurred in males at
the end of the exposure period or in either sex after the 13-week recovery period. Statistically
significant organ weight changes were increased relative liver weights in male and female rats at
252 ppm and increased absolute kidney weight in males at 252 ppm. No gross lesions were
observed at any concentration. Histopathology examinations revealed hepatocellular
hypertrophy in 3/10 males and 3/10 females at 252 ppm. The highest tested concentration, 252
ppm, is classified as a NOAEL based on liver effects. As discussed below, the increases in
hepatocellular hypertrophy and liver weight are likely to be adaptive responses to
4-chlorobenzotrifluoride exposure, and the serum chemistry findings are insufficient evidence of
liver damage.
Experimental data support the speculation that the hepatocellular hypertrophy and
increased liver weight are due to induction of hepatic P450. Pelosi et al. (1998) found that the
rats exposed to 252 ppm 4-chlorobenzotrifluoride for 13 weeks in the Newton et al. (1998) study
had slight, but statistically significant, increased activities of various P450 enzymes. Although
hepatic P450 isozymes were not evaluated after 4 weeks of exposure, the finding that the 13-
week exposure did not dramatically increase P450 activities is consistent with the diminished
incidences of hepatocellular hypertrophy after 13 weeks. Incidences of hepatocellular
hypertrophy decreased with continued exposure; for example, in the male rats, the incidences
11
-------
5-21-2007
were high after 4 weeks (7/10-10/10 at 100-1044 ppm) and only minimally increased after 13
weeks (3/10 at 252 ppm). No hepatocellular hypertrophy was observed after the 13-week
recovery period. Other evidence provides no indication that the hepatocellular hypertrophy and
increased liver weight were part of a spectrum of liver toxicity. The increases in relative liver
weight were small and did not increase with duration of exposure (10-11% in both sexes at 262
ppm after 4 weeks and at 252 ppm after 13 weeks). The increases in serum ALT, total protein
and albumin at 252 ppm after 13 weeks also were small, of questionable toxicological
significance, only occurred in females (the less sensitive sex based on the hypertrophy findings
in the 4-week study), and were not accompanied by increases in other liver enzymes (serum AST
and AP) in either sex. Additionally, the only histological change in the liver was hepatocellular
hypertrophy; no degenerative or other kinds of adverse lesions were induced.
A PBPK model for inhaled 4-chlorobenzotrifluoride was developed to facilitate
extrapolation of the results of the Newton et al. (1998) rat inhalation toxicity studies to humans
(Knaak et al., 1998). Concentrations of 4-chlorobenzotrifluoride in human tissues were
estimated using a model that was derived from human and rat partition coefficients, scaled
metabolic rate constants and default human physiological characteristics. The blood:air partition
coefficient was found to be a primary determinant of the fate of inhaled 4-chlorobenzotrifluoride;
percentage of body fat and muscle were also important. Simulations using the model predicted
similar tissue doses (area under the curve or AUC) in the livers of rats and humans exposed to
4-chlorobenzotrifluoride under the same conditions (10, 50 or 250 ppm for 1 day or 13 weeks).
A study from the Russian literature (Rapoport et al., 1986) included exposure of male
nonpurebred albino rats (numbers not specified) to 4-chlorobenzotrifluoride at 5.5, 20.5, 71.6 or
440 mg/m3 continuously for 120 hours, with subsequent observation for at least 115 days.
Examinations included body weight, serum chemistry and hematology and neurobehavior
("summation-threshold index," motor activity, grasping reflex). Exposure to 440 or 71.6 mg/m3
resulted in a change in "practically all of the parameters studied." The 20.5 mg/m3 concentration
was considered the "minimally effective" concentration, while the 5.5 mg/m3 was considered to
be "subthreshold." The results are presented as time of onset of significant changes in
parameters, which generally increased with decreasing exposure concentration. The study does
not relate exposure concentration to incidence or severity of effects. Therefore, a reliable
NOAEL or LOAEL cannot be derived from this study.
Other Studies
Genotoxicity testing results for 4-chlorobenzotrifluoride are mostly negative.
4-Chlorobenzotrifluoride did not induce reverse mutations in Salmonella typhimurium TA1535,
TA1537, TA1538, TA98 and TA100 with or without addition of exogenous metabolic activation
(S9) (NTP, 1992; Haworth et al., 1983; Litton Bionetics, 1978a; Mazza et al., 1986) or gene
conversions in Saccharomyces cerevisiae strain D4, +/- S9 (Litton Bionetics, 1978a). The
substance was also negative in a DNA repair deficiency assay with Escherichia coli W3110
polA+ and P3478 polA- tester strains, +/- S9 (Litton Bionetics, 1978b). Urine from mice treated
with the substance also tested negative in S. typhimurium TA1535, TA1537, TA98 and TA100
(Litton Bionetics, 1979a). The substance did not induce forward mutations at the thymidine
12
-------
5-21-2007
kinase (TK) locus in mouse lymphoma L5178Y cells, +/- S9 (Litton Bionetics, 1978c).
4-Chlorobenzotrifluoride induced sister chromatid exchanges (SCEs) in mouse lymphoma
L5178Y cells, +/- S9 (Litton Bionetics, 1979b), but did not induce chromosomal aberrations in
Chinese hamster ovary cells, +/- S9 (Lilly Research Laboratories, 1983a). Negative results were
also reported for 4-chlorobenzotrifluoride in an in vivo chromosomal aberration assay in rats
(Lilly Research Laboratories, 1983b). The substance was positive in an assay for unscheduled
DNA synthesis in EUE cells (not further described) (Benigni and Dogliotti, 1980). Litton
Bionetics (1980) reported that 4-chlorobenzotrifluoride tested negative for cell transformation in
Balb/C3T3 cells; the study did not report whether an activating system was used. Lilly Research
Laboratories (1983c) also found that 4-chlorobenzotrifluoride (97% pure) tested negative for cell
transformation in Balb/C3T3 cells in the presence of S9.
DERIVATION OF PROVISIONAL SUBCHRONIC AND CHRONIC
ORAL RfD VALUES FOR 4-CHLOROBENZOTRIFLUORIDE
The available animal data indicate that the liver, kidneys and adrenals are the most
sensitive targets of response for 4-chlorobenzotrifluoride after oral exposure. Liver effects
included increases in organ weight, hepatocellular hypertrophy and serum chemistry changes
(increases in cholesterol, triglycerides, 5'-nucleotidase, bile acids and bilirubin) in rats and mice
(NTP, 1992; Macri et al., 1987; Arthur and Probst, 1983). Adrenal effects included increased
organ weight and cytoplasmic vacuolation in the cortex in rats (NTP, 1992; Macri et al., 1987).
Renal effects included increased organ weight, hyaline droplet accumulation, tubule dilation,
inflammation of the interstitial cells and regeneration of the epithelium in male rats (NTP, 1992;
Macri et al., 1987; Arthur and Probst, 1983). The effects on the kidneys appear to be related to
a2u-globulin accumulation, a male rat-specific effect that is not predictive for health effects in
humans (U.S. EPA, 1991b). Support for this relationship comes from the observation that
kidney lesions were observed only in male rats in these studies, and not female rats or mice of
either sex; the types of lesions reported are consistent with those typical of a2U-globulin
accumulation, as described in U.S. EPA (1991b); and a2u-globulin levels in the kidney were
significantly increased and correlated with the presence of renal lesions in male rats in the NTP
(1992) study. Other effects of 4-chlorobenzotrifluoride included minor hematological alterations
at doses in the range of those causing the hepatic serum chemistry and adrenal effects (NTP,
1992; Arthur and Probst, 1983). A limited one-generation reproductive toxicity assessment
found no effects on reproductive function in rats (EBL, 1981), suggesting that the reproductive
system is not a sensitive target for 4-chlorobenzotrifluoride. No developmental toxicity studies
on 4-chlorobenzotrifluoride were located; therefore, it is not possible to determine whether the
chemical is a developmental toxicant.
Hepatocellular hypertrophy occurred at doses as low as 50 mg/kg-day (NTP, 1992) and
150 mg/kg-day (Arthur and Probst, 1983) in rats and 400 mg/kg-day in mice (NTP, 1992).
However, this effect alone is not considered by the EPA to be adverse because, in the absence of
other indicators of liver toxicity, such as histopathology or changes in serum chemistry, it likely
reflects adaptive enzyme induction. The serum chemistry changes observed in rats and mice are
indicative of cholestasis (an impaired production or flow of bile) and therefore are considered
13
-------
5-21-2007
indicators of liver toxicity. Increases in serum bilirubin, alkaline phosphatase, gamma glutamyl
transpeptidase, 5'-nucleotidase, cholesterol, triglycerides and total bile acids are associated with
cholestasis (U.S. EPA, 2002; NTP, 1992) and serum 5'-nucleotidase activity is a particularly
sensitive and specific indicator of biliary duct damage in rats, mice and humans (Carakostas et
al., 1986, 1990). NTP (1992) suggested a cellular rather than physical mechanism for 4-
chlorobenzotrifluoride cholestasis because the serum chemistry changes were not accompanied
by histopathologic evidence of hepatocellular toxicity. Cellular mechanisms that could account
for the cholestatic effects include 4-chlorobenzotrifluoride-induced alterations in plasma
membranes (function and composition), disruptions in microfilaments and microtubules, loss of
tight junction integrity and/or changes in the concentration and composition of bile acids (NTP,
1992). U.S. EPA (2002) similarly indicates that cholestasis can be due to hepatocellular
dysfunction that is not evident histologically. The lowest dose of 4-chlorobenzotrifluoride that
induced a definitive serum chemistry change was 100 mg/kg-day for increased cholesterol and
triglycerides in male rats (62 and 92% higher than controls, respectively) (Macri et al., 1987).
Changes in these indices were generally dose-related, also occurring at 400 and 1000 mg/kg-day
in one or both sexes; at 1000 mg/kg-day, serum cholesterol was increased 75-134% in rats and
65-84%) in mice (NTP, 1992; Macri et al., 1987) and serum triglyceride was increased 71-137%)
in rats and 78%> in mice (NTP, 1992; Macri et al., 1987). Other serum chemistry changes at
doses above 100 mg/kg-day included increases in 5'-nucleotidase at 400 and 1000 mg/kg-day in
rats (13-21%o and 46-62%>) and mice (30%> and 159-166%>) in one or both sexes (NTP, 1992),
serum bilirubin at 500 mg/kg-day in male and female rats (46-75%>) (Arthur and Probst, 1983)
and total bile acids at 1000 mg/kg-day in male rats (123%>) (NTP, 1992). Increases in serum
alkaline phosphatase were observed at doses ranging from 10-500 mg/kg-day in rats (Arthur and
Probst, 1983), but the response was not dose-related or of magnitudes likely to be toxicologically
significant; additionally, serum alkaline phosphatase was not increased in rats exposed to <45
mg/kg-day (EBL, 1981).
Adrenal cortex cytoplasmic vacuolation occurred in male and female rats exposed to 400
and 1000 mg/kg-day for 2 weeks (NTP, 1992) and males rats exposed to 1000 mg/kg-day for 4
weeks (Macri et al., 1987). No histological changes occurred in the adrenals of male or female
rats exposed to <500 mg/kg-day for 3 months (Arthur and Probst, 1983) or <45 mg/kg-day for 3
months (EBL, 1981). NTP (1992) concluded that the cause and biological significance of the
adrenal lesion are not clear because vacuolation and other microscopic changes are often
observed in the adrenal cortex in toxicity studies, and the study was not designed for determining
if it was due to stress, an increase in ACTH stimulation due to impaired adrenal steroid
production or other mechanisms. Macri et al. (1987) observed that the effect could indicate
decreased adrenal activity and be consistent with the observed hypercholesterolemia.
The primary liver and adrenal findings (NTP, 1992; Macri et al., 1987), summarized in
Table 2, are that male rats are the species and sex that appear to be most sensitive to 4-
chlorobenzotrifluoride. The lowest dose of 4-chlorobenzotrifluoride that induced serum
chemistry changes likely to be toxicologically significant was 100 mg/kg-day for increased
cholesterol and triglycerides in rats exposed for 28 days (Macri et al., 1987), indicating that this
dose is the LOAEL for subchronic oral exposure based on liver effects; the NOAEL is 50 mg/kg-
day (NTP, 1992). The lowest dose inducing adrenal cortex cytoplasmic vacuolation, a possible
14
-------
5-21-2007
Table 2. Effects of Oral Exposure to 4-Chlorobenzotrifluorideon Selected Liver and Adrenal Endpoints in Male Rats
Effect
Dose (mg/kg-day)
Reference
0
10
40
50
100
150
400
500
1000
Hepatocellular
hypertrophyf
0/5
0/5
5/5
5/5
NTP, 1992ac
0/5
0/5
2/5
5/5
NTP, 1992b'°
0/6
0/6
0/6
0/6
Macri et al., 1987d
0/15
0/15
0/15
15/15
13/15
Arthur and Probst, 1983®
Serum cholesterol
(% increase)
-
8
51s
134s
NTP, 1992a
-
3
17s
37s
NTP, 1992b
-
24
62s
75s
Macri et al., 1987
not evaluated
Arthur and Probst, 1983
Serum triglycerides
(% increase)
-
10
17
54
NTP, 1992a
-
3
(-8)
(-18)
NTP, 1992b
-
19
92s
137s
Macri et al., 1987
not evaluated
Arthur and Probst, 1983
Serum
5'-nucleotidase
(% increase)
-
2
13s
62s
NTP, 1992a
-
4.3
10g
15s
NTP, 1992b
not evaluated
Macri et al., 1987
not evaluated
Arthur and Probst, 1983
Serum total bile acids
(% increase)
-
0
9
123s
NTP, 1992a
-
3.4
(-20)
(-16)
NTP, 1992b
not evaluated
Macri et al., 1987
not evaluated
Arthur and Probst, 1983
Serum bilirubin
(% increase)
-
0f
0f
0f
Macri et al., 1987
-
4
(-6)
23
46s
Arthur and Probst, 1983
not evaluated
NTP, 1992
Serum alkaline phosphatase
(% increase)
-
41s
46s
46s
40s
Arthur and Probst, 1983
not evaluated
NTP, 1992
not evaluated
Macri et al., 1987
Adrenal cortex
cytoplasmic vacuolationf
0/5
0/5
4/5
5/5
NTP, 1992a
0/5
0/5
5/5
NTP, 1992b
0/6
0/6
0/6
5/6
Macri et al., 1987
0/15h
0/15h
0/15h
0/15h
0/15h
Arthur and Probst, 1983
e90-day study, n=15 per group
incidence data not statistically evaluated in any study
Statistically different from control group
hNo histological effects in adrenals (incidence data not specifically reported)
Corn oil vehicle
ba-Cyclodextrin vehicle
c 14-Day study, n=5 per group
d28-Day study, n=6 per group
15
-------
5-21-2007
adverse effect, is 400 mg/kg-day (NTP, 1992), which is higher than the 100 mg/kg-day LOAEL
based on liver effects. Key effect levels in the available oral studies of 4-chlorobenzotrifluoride
are summarized in Table 3.
The Macri et al. (1987) study was selected as the key study for derivation of the RfD.
The critical endpoint(s) chosen for analysis from this study were changes in serum cholesterol
and serum triglycerides observed in rats. Based on the database available, this study identifies
the lowest subchronic LOAEL for 4-chlorobenzotrifluoride, 100 mg/kg-day for hepatic effects,
and provides the most appropriate basis for benchmark dose modeling for derivation of a
reference value. The study was conducted over a range of exposure concentrations, included a
control group, and demonstrated a dose-related effect. These changes in serum chemistry are
indicators of liver toxicity associated hepatocellular hypertrophy. Available continuous-variable
models in the EPA Benchmark Dose Software (linear, polynomial, power and Hill models;
BMDS version 1.3.2) were fit to the dose-response data for changes in serum cholesterol and
triglycerides in male rats. The BMDs and 95% lower confidence limits (BMDLs) calculated for
these endpoints are estimates of the doses associated with a change of 1 standard deviation from
the control (U.S. EPA, 2000). The predicted BMDs and BMDLs for the two endpoints are
summarized in Table 4 and detailed in Appendix A.
Table 3. Summary of Effect Levels from Oral Studies of 4-Chlorobenzotrifluoride
Study
Exposure
Duration
Species
(Strain)
NOAEL
(mg/kg-day)
LOAEL
(mg/kg-day)
Main Endpoints
Macri et al. (1987)
4 weeks
rat
(SD)
10
100
increased serum cholesterol and
triglycerides
NTP (1992)
2 weeks
rat (F344)
50
400
increased serum cholesterol,
adrenal cortex vacuolation
NTP (1992)
2 weeks
mouse
(B6C3F1)
50
400
increased serum cholesterol and
triglycerides
Arthur and Probst
(1983)
3 months
rat
(F344)
150
500
increased serum bilirubin
EBL (1981)
3 months
rat
(SD)
45
ND
none
SD= Sprague-Dawley; ND = not determined
Table 4. Summary of BMD Model Predictions for Male Rats Exposed to
4-Chlorobenzotrifluoride by Daily Gavage for 28 Days3
Endpoint
bmd1sd
(mg/kg-day)
BMDLi SD
(mg/kg-day)
Serum cholesterol level
NA
NA
Serum triglycerides level
15.1
8.8
"Doses associated with a 1 standard deviation change from the control
NA = not available: no adequate fits were obtained
16
-------
5-21-2007
The BMDL i sd of 8.8 mg/kg-day derived by benchmark dose modeling of the Macri et al.
(1987) serum triglycerides data is similar to the NOAEL of 10 mg/kg-day for increases in serum
triglycerides and cholesterol in this study. Because the BMDLi Sd incorporates more
information about the shape of the dose-response curve than the NOAEL, the BMDLi Sd is used
as the point of departure for derivation of the RfD. The BMDLi sd of 8.8 mg/kg-day is divided
by a composite uncertainty factor (UF) of 300 [10 for interspecies extrapolation, 10 for human
variability and 3* for database deficiencies (discussed below)] to derive a provisional
subchronic RfD of 3E-2 mg/kg-day, as follows:
sRfD = BMDLi sd/UF
= 8.8 mg/kg-day / 300
= 0.03 or 3E-2 mg/kg-day
*Half-log of 10 rounded to 3
Because no chronic oral toxicity studies were located in the literature, an additional UF
of 10 is applied to the provisional subchronic RfD to derive the provisional chronic RfD of 3E-3
mg/kg-day, as follows:
RfD = sRfD - UF
= 0.03 mg/kg-day ^-10
= 0.003 or 3E-3 mg/kg-day
Confidence in the key study is medium. The study evaluated a sufficient number of
doses and an adequate set of toxicological parameters in rats, but is limited by small numbers of
animals (6/sex/dose) and a duration of 28 days. The key study is supported by 14-day studies in
rats and mice that produced similar changes in serum chemistry indicative of cholestatic liver
damage (NTP, 1992). Confidence in the database is medium to low. The database lacks a
comprehensive study longer than 28 days in duration; a 90-day study in rats provides limited
support for the key study because its serum chemistry evaluations only included a few of the
endpoints assessed in the key study. The database is also limited by the absence of
developmental toxicity studies. Reproductive toxicity was examined but only in a single-
generation study that included assessment of few reproductive endpoints and did not include
doses high enough to produce adverse effects on any endpoint (EBL, 1981). Confidence in the
provisional subchronic RfD is medium. Confidence in the chronic RfD is low because no
chronic toxicity studies were located.
DERIVATION OF PROVISIONAL SUBCHRONIC AND CHRONIC
INHALATION RfC VALUES FOR 4-CHLOROBENZOTRIFLUORIDE
Only one adequately conducted inhalation toxicity study was located (Newton et al.,
1998). This study included 4- and 13-week experiments in rats, each of which included
evaluations of a comprehensive set of standard toxicological parameters. Both experiments
identified the liver as the most sensitive target of response for 4-chlorobenzotrifluoride. Liver
17
-------
5-21-2007
effects in the two experiments included hepatocellular hypertrophy, increased liver weight and
minor changes in serum chemistry (small increases in serum protein and, in the 13-week
experiment, a small increase in serum ALT). As discussed previously, the liver effects are not
considered to be adverse because (1) the hepatocellular hypertrophy and increased liver weight
are likely adaptive responses reflecting the induction of hepatic P450 enzymes, (2) the serum
chemistry changes were small and of questionable toxicological significance and (3) there is no
histological evidence (adverse lesions) to suggest that the hypertrophy, increased liver weight
and serum chemistry changes are part of a spectrum of liver toxicity. Other effects included
clinical signs of neurotoxicity (hyperactivity and tremors) at the highest exposure level in the 4-
week study. The kidney also was identified as a target in male rats in the 4-week study;
however, these effects were associated with a2U-globulin accumulation and are not considered
relevant to human health risk assessment (U.S. EPA, 1991b). The 4-week study identified a
NOAEL of 492 ppm based on the liver effects and a LOAEL of 1044 ppm based on the clinical
signs of neurotoxicity. The EPA identified a NOAEL of 252 ppm for liver effects from the 13-
week study; this was the highest tested concentration, precluding identification of a possible
LOAEL. The 13-week NOAEL of 252 ppm was chosen, based on length of the study, as the
most appropriate point of departure for derivation of the RfC.
Calculation of an RfC using the 252 ppm NOAEL in rats first involves determination of a
human equivalent concentration (HEC). The U.S. EPA (1994b) default procedure for calculating
the HEC for an extrarespiratory effect from a vapor is to adjust for intermittent exposure and
multiply the duration-adjusted NOAEL by the ratio of animal to human blood:air partition
coefficients, as follows:
NOAELmg/m3 = ppm x MW / 24.45
252x 180.55/24.45
1860.9
NOAELadj = 1860.9 mg/m3 x 6 hours/24 hours x 5 days/7 days
= 332 mg/m3
NOAELrec - NOAEL adj x (Hb/g)A / (Hb/g)H
= 332 mg/m3 x 1
= 332 mg/m3
where,
(Hb/g)A / (Hb/g)H= rat to human blood:air partition coefficient ratio
= default ratio of 1, because LR (47.9; Knaak et al., 1998) is
greater than LH (16.7; Knaak et al., 1998)
The PBPK model developed by Knaak et al. (1998) supports this derivation of human
equivalent concentration. The blood:air partition coefficient was found to be a primary
determinant of 4-chlorobenzotrifluoride disposition in the model, and that is the factor used to
make the adjustment to human equivalent concentration in the default EPA procedure. The
18
-------
5-21-2007
results of the PBPK modeling simulations performed by Knaak et al. (1998) show delivery of
similar doses to the liver in humans and rats with intermittent subchronic exposure to 250 ppm, a
concentration virtually the same as the NOAEL of 252 ppm, indicating that a multiplier of 1 is
appropriate to calculate NOAELHec from NOAELAdj
The NOAELrec of 332 mg/m3 is divided by a composite UF of 100 (10 for human
variability, 3* for toxicodynamic considerations of interspecies extrapolation after making the
toxicokinetic adjustments described above and 3* for database deficiencies, including lack of a
subchronic LOAEL in rats exposed for longer than 4 weeks, lack of subchronic study in a second
species and lack of reproductive and developmental toxicity studies) to derive a provisional
subchronic RfC of 3E-0 mg/m3, as follows:
sRfC = NOAELrec / UF
332 mg/m3/ 100
= 3 or 3E-0 mg/m3
*Half-log of 10 rounded to 3
Because no chronic inhalation toxicity studies were located in the literature, an additional
UF of 10 is applied to the provisional subchronic RfC to derive a provisional chronic RfC of
3E-1 mg/m3, as follows:
RfC = sRfC - UF
= 3 mg/m3 + 10
= 0.3 or 3E-1 mg/m3
Confidence in the key study is high. The study evaluated a sufficient number of doses
and an adequate set of toxicological parameters. The study design also included a
comprehensive neurotoxicology evaluation. Confidence in the database is low. No other
inhalation toxicity studies are available that support the key study. In addition, the database is
limited by the absence of adequate reproductive and developmental toxicity studies. The results
of a one-generation oral study (EBL, 1981) suggest that the reproductive system is not likely to
be a sensitive target for 4-chlorobenzotrifluoride, but only a limited evaluation of reproductive
toxicity endpoints was performed. Based on the limited inhalation toxicity database, low
confidence in the provisional subchronic and chronic RfC follows.
DERIVATION OF A PROVISIONAL CARCINOGENICITY ASSESSMENT
FOR 4-CHLOROBENZOTRIFLUORIDE
Relevant studies regarding the carcinogenicity of 4-chlorobenzotrifluoride in humans or
animals following oral or inhalation exposure were not located. Available genotoxicity data
indicate that 4-chlorobenzotrifluoride is not a potent genotoxic agent. In accordance with
current EPA cancer guidelines (U.S. EPA, 2005), the available data are inadequate for an
assessment of human carcinogenic potential. Therefore, derivation of quantitative estimates
19
-------
5-21-2007
(oral slope factor or inhalation unit risk) of cancer risk for 4-chlorobenzotrifluoride is precluded
by the absence of carcinogenicity data.
REFERENCES
ACGIH (American Conference of Governmental Industrial Hygienists). 2005. Threshold limit
values for chemical substances and physical agents and biological exposure indices. 2003 TLVs
and BEIs. Cincinnati, OH.
Arthur, B.H. and K.S. Probst. 1983. A Subchronic (Three-Month) Toxicity Study in Fischer
344 Rats Given Daily Gavage Doses of 4-Chlorobenzotrifluoride (PCBTF). Toxicology
Division, Lilly Research Laboratories, Division of Eli Lilly and Company, Greenfield, IN. U.S.
EPA/OPTS Public Files. Microfiche #0TS0507306.
ATSDR (Agency for Toxic Substances and Disease Registry). 2006. Toxicological Profile
Information Sheet. U.S. Department of Health and Human Services, Public Health Service.
Online, http://www.atsdr.cdc.gov/toxpro2.html
Benigni R. and R. Dogliotti. 1980. UDS studies on selected environmental chemicals
(Abstract). Mutat. Res. 74:217.
Carakostas, M.C., K.A. Gossett, G.E. Church and B.L. Cleghorn. 1986. Evaluating toxin-
induced hepatic injury in rats by laboratory results and discriminant analysis. Vet. Pathol.
23:264-269.
Carakostas, M.C., R.J. Power and A.K. Baneijee. 1990. Serum 5'nucleotidase activity in rats: A
method for automated analysis and criteria for interpretation. Vet. Clin. Pathol. 19:109-113.
EBL (Elars Bioresearch Laboratories). 1981. Modified 90-Day Gavage and Reproduction Study
in Rats: PCBTF. Fort Collins, CO. U.S. EPA/OPTS Public Files. Microfiche #0TS0508148.
Haworth, S T., T. Lawlor and K. Mortlemans et al. 1983. Salmonella mutagenicity test results
for 250 chemicals. Environ. Mutagen. Suppl. 1:3-5, 8-9, 16-21, 41, 51, 78.
IARC (International Agency for Research on Cancer). 2006. Search IARC Monographs.
Online, http://monographs.iarc.fr/
Knaak, J.B., L.W. Smith, R.D. Fitzpatrick et al. 1998. In vitro hepatic metabolism of PCBTF:
development of vmax and km values and partition coefficients and their use in an inhalation
PBPK model. Inhal. Toxicol. 10(l):65-85.
Lilly Research Laboratories. 1983a. Chromosome Aberrations in Chinese Hamster Ovary Cells:
Test Article Compound 38502, Lot No. 633F02. Prepared by Microbiological Associates,
Bethesda, MD. U.S. EPA/OPTS Public Files. Microfiche #0TS0507306.
20
-------
5-21-2007
Lilly Research Laboratories. 1983b. Activity of Compound 38502 (T2025) in the Acute In Vivo
Cytogenetic Assay in Male and Female Rats. Prepared by Microbiologjcal Associates,
Bethesda, MD. U.S. EPA/OPTS Public Files. Microfiche #0TS0507306.
Lilly Research Laboratories. 1983c. Evaluation of Compound #38502 in the BALB/C3T3
Neoplastic Transformation Assay with an Aroclor-Induced Rat Liver Microsomal (S9)
Metabolic Activating System. Submitted by Arthur D. Little, Incorporated, Cambridge, MA.
U.S. EPA/OPTS Public Files. Microfiche #0TS0507306.
Litton Bionetics. 1978a. Mutagenicity Evaluation of Parachlorobenzo Trifluoride (PCBTF) in
the Ames Salmonella/Microsome Plate Test. Final Report. Submitted to Hooker Chemical
Corp. U.S. EPA/OPTS Public Files. Microfiche #0TS0508133.
Litton Bionetics. 1978b. Mutagenicity Evaluation of Parachlorobenzo Trifluoride (PCBTF) in
the Mouse Lymphoma Forward Mutation Assay. Final Report. Submitted to Hooker Chemical
Corp. U.S. EPA/OPTS Public Files. Microfiche #OTS0508135.
Litton Bionetics. 1978c. Mutagenicity Evaluation of Parachlorobenzotrifluoride in the Mouse
Lymphoma Forward Mutation Assay. Final Report. Submitted to Hooker Chemical Corp. U.S.
EPA/OPTS Public Files. Microfiche #OTS0508135.
Litton Bionetics. 1979a. Mutagenicity Evaluation of Parachlorobenzo Trifluoride in a In
Vivo/In Vitro Urine Assay. Final Report. Submitted to Hooker Chemical Corp. U.S.
EPA/OPTS Public Files. Microfiche #0TS05O8139.
Litton Bionetics. 1979b. Mutagenicity Evaluation of Parachlorobenzo Trifluoride (PCBTF) in
the Sister Chromatid Exchange Assay in L5178Y Mouse Lymphoma Cells. Final Report.
Submitted by Litton Bionetics, Inc., Kensington, MD. U.S. EPA/OPTS Public Files. Microfiche
#OTS0508136.
Litton Bionetics. 1980. Evaluation of p-Chlorobenzotrifluoride in the In Vitro Transformation
of BALB/3T3 Cells Assay. Submitted by Litton Bionetics, Inc., Kensington, MD. U.S.
EPA/OPTS Public Files. Microfiche #0TS0508144.
Macri, A., C. Ricciardi, A.V. Stazi et al. 1987. Subchronic oral toxicity of 4-chloro-a-a-a-
trifluorotoluene in Sprague-Dawley rats. Food Chem. Toxicol. 25(10):781-786.
Mazza, G., C. Dacarro, C. Bonferoni and B. Bonferoni. 1986. Studies on the mutagenic activity
of benzotrifluoride and twelve derivatives in microbial short-term assays. Farmaco. Ed. Prat.
41(7): 215-225. (Cited in CCRIS database.) Online, http://toxnet.nlm.nih.gov/
NIOSH (National Institute for Occupational Safety and Health). 2006. NIOSH Pocket Guide to
Chemical Hazards. Online, http://www.cdc.gov/niosh/npg/
21
-------
5-21-2007
Newton P.E., H.F. Bolte, W.R. Richter and M.B. Akinsanya. 1998. Inhalation toxicity,
neurotoxicity, and toxicokinetic studies of p-chlorobenzotrifluoride. Inhal. Toxicol. 10:33-48.
NTP (National Toxicology Program). 1992. NTP Technical report on toxicity studies of
p-chloro-a,a,a-trifluorotoluene (CAS NO: 98-56-6): Administration in corn oil and
a-cyclodextrin to F344/N rats and B6C3Fi mice in 14-day comparative gavage studies. NIH
Publication 92-3133.
NTP. 2006. Management Status Report. Online.
http://ntp.niehs.nih.gov/index.cfm?obiectid=78CC7E4C-FlF6-975E-72940974DE301C3F
OSHA (Occupational Safety and Health Administration). 2006. OSHA Standard 1910.1000
Table Z-l. Part Z, Toxic and Hazardous Substances. Online.
http://www.osha.gov/pls/oshaweb/owadisp.show document?p table=STANDARDS&p id=999
2.
Pelosi, G.D., J. Oberdoerster, J.R. Olson et al. 1998. Characterization of rat hepatic cytochrome
P-450 activities following inhalation exposure to p-chlorobenzotrifluoride. Inhal. Toxicol.
10:49-63.
Rapoport, K.A., L.A. Tepikina, Y.G. Fel'dman et al. 1986. [Setting the limits for
parachlorobenzotrifluoride in atmospheric air.] Gig. Sanit. 10:82-83. (Russian with English
translation)
U.S. EPA. 1988. Health and Environmental Effects Document (HEED) for 4-
Chlorobenzotrifluoride. Prepared by the Office of Health and Environmental Assessment,
Cincinnati, OH for the Office of Solid Waste and Emergency Response, Washington, DC.
U.S. EPA. 1991a. Chemical Assessments and Related Activities (CARA). Office of Health and
Environmental Assessment, Washington, DC.
U.S. EPA. 1991b. Alpha-2u-globulin: Association with Chemically Induced Renal Toxicity and
Neoplasia in the Male Rat. Risk Assessment Forum, Washington, DC. EPA/625/3-91/019F.
U.S. EPA. 1994a. Chemical Assessments and Related Activities (CARA). Office of Health and
Environmental Assessment, Washington, DC. December.
U.S. EPA. 1994b. Methods for Derivation of Inhalation Reference Concentrations and
Application of Inhalation Dosimetry. Office of Research and Development, Washington, DC.
EPA/600/8-90/066F.
U.S. EPA. 1997. Health Effects Assessment Summary Tables. FY-1997 Update. Prepared by
the Office of Research and Development, National Center for Environmental Assessment,
22
-------
5-21-2007
Cincinnati OH for the Office of Emergency and Remedial Response, Washington, DC. July.
EPA/540/R-97/036. NTIS PB97-921199.
U.S. EPA. 2000. Benchmark Dose Technical Guidance Document. Risk Assessment Forum,
Washington, DC. External Review Draft. EPA/630/R-00/001.
U.S. EPA. 2002. Hepatocellular Hypertrophy. HED Guidance Document # G2002.01. HED
Toxicology Science Advisory Council, Health Effects Division, Office of Pesticide Programs.
October 21, 2002.
U.S. EPA. 2004. 2004 Edition of the drinking water standards and health advisories. Office of
Water, Washington, DC. EPA 822-R-02-038. Washington, DC.
http://www.epa.gov/waterscience/drinking/standards/dwstandards.pdf
U.S. EPA. 2005. Guidelines for Carcinogen Risk Assessment and Supplemental Guidance for
Assessing Susceptibility from Early-Life Exposure to Carcinogens. Risk Assessment Forum,
Washington, DC. EPA/630/P-03/001F. Online.
http://cfpub. epa.gov/ncea/cfm/recordisplav. cfm?deid=l 16283
U.S. EPA. 2007. Integrated Risk Information System (IRIS). Office of Research and
Development, National Center for Environmental Assessment, Washington, DC. Online.
http ://www. epa. gov/iri s/
WHO (World Health Organization). 2006. Online catalogs for the Environmental Health
Criteria Series. Online, http://www.inchem.org/pages/ehc.html
Yuan J., C.W. Jameson, T.J. Goehl et al. 1992. Application of molecular encapsulation for
toxicology studies: comparative toxicity of p-chloro-a,a,a-trifluorotoluene in a-cyclodextrin
vehicle versus corn oil vehicle in male and female Fischer 344 rats and B6C3Fi mice. Fundam.
Appl. Toxicol. 18:460-470.
23
-------
5-21-2007
APPENDIX A. DETAILS OF BMD ANALYSIS FOR THE SUBCHRONIC ORAL RfD
FOR 4-CHLOROBENZOTRIFLUORIDE
Hepatic effects data from the Macri et al. (1987) 28-day study in rats were used for
benchmark dose modeling. Available continuous-variable models in the EPA Benchmark Dose
Software (linear, polynomial, power and Hill models; BMDS version 1.3.2) were fit to the data
for changes in serum cholesterol and triglycerides in male rats shown in Table A-l. The BMDs
and 95% lower confidence limits (BMDLs) calculated for these endpoints are estimates of the
doses associated with a change of 1 standard deviation from the control (U.S. EPA, 2000).
Table A-l. Serum Levels of Cholesterol and Triglycerides in Male Rats Exposed to
4-Chlorobenzotrifluoride by Daily Gavage for 28 Days (Macri et al., 1987)
Dose (mg/kg-day)
0
10
100
1000
Group size
6
6
6
6
Serum cholesterol
(mg/100 mL)a
56.50 ±7.50
70.17 ±8.97
91.33 ±23.69b
98.83 ± 34.69°
Serum triglycerides
(mg/100 mLf
33.50 ±6.05
39.83 ±7.82
64.33 ± 30.30b
79.33 ±25.28c
"Values are means ± standard deviation
Statistically significant increase from control value (p<0.05)
Statistically significant increase from control value (p<0.01)
The simplest model (linear) was applied to the data first while assuming constant
variance. If the data were consistent with the assumption of constant variance (/>>0.05), then the
fit of the linear model to the means was evaluated. If the linear model adequately fit the means
(P>0.1), then it was selected as the model for BMD derivation. If the linear model did not
adequately fit the means, then the more complex models were fit to the data while assuming
constant variance. Among the models providing adequate fit to the means (p>0.1), the one with
the lowest AIC for the fitted model was selected for BMD derivation. If the test for constant
variance was negative, the linear model was run again while applying the power model
integrated into the BMDS to account for nonhomogenous variance. If the nonhomogenous
variance model provided an adequate fit (p>0.05) to the variance data, then the fit of the linear
model to the means was evaluated. If the linear model did not provide adequate fit to the means
while the variance model was applied, then the polynomial, power, and Hill models were fit to
the data and evaluated while the variance model was applied. Among those providing adequate
fit to the means (p>0.1), the one with the lowest AIC for the fitted model was selected for BMD
derivation. If the test for constant variance was negative and the nonhomogenous variance
model did not provide an adequate fit to the variance data, then the data set was considered not
to be suitable for BMD modeling. The predicted BMDs and BMDLs for the two endpoints are
detailed below.
24
-------
5-21-2007
SERUM CHOLESTEROL DATA
For the serum cholesterol data, the assumption of constant variance did not hold. The
nonhomogeneous variance model was applied and provided adequate fit to the variance. With
the nonhomogeneous variance model applied, the linear model did not provide an adequate fit to
the means, and neither did any of the other available models (insufficient degrees of freedom for
Hill model). Results are shown in Table A-2. Given that the most pertinent part of the dose-
response curve is that which lies at the lower doses, the high-dose data point (1000 mg/kg-day)
was removed from the dataset and the model fitting was conducted, as described before, in an
attempt to achieve an adequate fit of the models to the data. As with the complete data set, the
assumption of constant variance did not hold, but the nonhomogeneous variance model provided
adequate fit to the variance. Sufficient degrees of freedom were available for fitting only the
linear model. However, the fit of the linear model to the means was not adequate. Results are
shown in Table A-3.
25
-------
5-21-2007
Table A-2. Model Predictions for Changes in Serum Cholesterol Levels in Male Rats
Exposed to 4-Chlorobenzotrifluoride by Daily Gavage for 28 Days
Model
Variance
p-valuea
Means
p-valuea
AIC
BMDisd
(mg/kg-day)
BMDLlsd
(mg/kg-day)
Linear (constant variance)
0.0007
0.0339
178.33
775.61
465.39
Linear (modeled variance)
0.5294
<0.0001
178.71
494.25
202.50
Polynomial15'0 (modeled variance)
0.5294
<0.0001
178.71
494.47
202.50
Powerd (modeled variance)
<0.0001
<0.0001
180.71
495.23
202.50
Hilld (modeled variance)
0.5294
NA
161.64
4.64
NA
"Values <0.1 fail to meet conventional goodness-of-fit criteria.
b2-degree polynomial; no adequate fit with any polydegree.
°betas restricted to > 0.
dpower restricted to >1.
AIC = Akaike's Information Criteria; p = p value from the Chi-squared test; BMD = benchmark dose; BMDL =
lower confidence limit (95%) on the benchmark dose; NA = not available (BMD software could not generate a model
output).
Table A-3. Model Predictions for Changes in Serum Cholesterol Levels in Male Rats
Exposed to 4-Chlorobenzotrifluoride by Daily Gavage for 28 Days (High Dose Dropped)
Model
Variance
p-valuea
Means
p-valuea
AIC
BMDlsd
(mg/kg-day)
BMDLiSd
(mg/kg-day)
Linear (constant variance)
0.0082
0.1945
118.49
48.07
32.03
Linear (modeled variance)
0.3584
0.04357
116.12
25.12
12.90
Polynomial13'0 (modeled variance)
0.3584
NA
116.04
26.34
13.05
Powerd (modeled variance)
0.0071
NA
117.98
25.32
13.14
Hilld (modeled variance)
NA
aValues <0.1 fail to meet conventional goodness-of-fit criteria.
b2-degree polynomial; no adequate fit with any polydegree.
°betas restricted to > 0.
dpower restricted to >1.
AIC = Akaike's Information Criteria; p = p value from the Chi-squared test; BMD = benchmark dose; BMDL =
lower confidence limit (95%) on the benchmark dose; NA = not available (BMD software could not generate a
model output).
26
-------
5-21-2007
SERUM TRIGLYCERIDE DATA
For the serum triglyceride level data, the assumption of constant variance did not hold.
The nonhomogeneous variance model was applied and provided adequate fit to the variance.
With the nonhomogeneous variance model applied, the linear model did not provide an adequate
fit to the means, and neither did any of the other available models (insufficient degrees of
freedom for Hill model). Results are shown in Table A-4. The results after dropping the high
dose are shown in Table A-5. As with the complete data set, the assumption of constant variance
did not hold, but the nonhomogeneous variance model provided adequate fit to the variance.
Sufficient degrees of freedom were available for fitting only the linear model. The linear model
provided adequate fit to the means while the variance model was applied, resulting in a predicted
BMD i sd and BMDLi sd of 15.1 and 8.8 mg/kg-day, respectively (Table A-5, Figure A-l).
Table A-4. Model Predictions for Changes in Serum Triglyceride Levels in Male Rats
Exposed to 4-Chlorobenzotrifluoride by Daily Gavage for 28 Days
Model
Variance
p-valuea
Means
p-valuea
AIC
BMDisd
(mg/kg-day)
BMDLlsd
(mg/kg-day)
Linear (constant variance)
0.0003
0.0440
174.48
572.29
376.90
Linear (modeled variance)
0.5716
<0001
177.82
486.21
237.10
Polynomial13'0 (modeled variance)
0.5716
<0001
177.82
486.21
237.10
Powerd (modeled variance)
0.3041
<.00001
179.82
486.21
237.10
Hilld (modeled variance)
0.5719
NA
158.71
9.36
NA
aValues <0.1 fail to meet conventional goodness-of-fit criteria.
b2-degree polynomial; no adequate fit with any polydegree.
°betas restricted to > 0.
dpower restricted to >1.
AIC = Akaike's Information Criteria; p = p value from the Chi-squared test; BMD = benchmark dose; BMDL =
lower confidence limit (95%) on the benchmark dose; NA = not available (BMD software could not generate a
model output).
27
-------
5-21-2007
Table A-5. Model Predictions for Changes in Serum Triglyceride Levels in Male Rats
Exposed to 4-Chlorobenzotrifluoride by Daily Gavage for 28 Days (High Dose Dropped)
Model
Variance
p-valuea
Means
p-valuea
AIC
BMDisd
(mg/kg-day)
BMDLlsd
(mg/kg-day)
Linear (constant variance)
0.0001
0.7261
123.69
57.33
36.56
Linear (modeled variance)
0.6451
0.3826
110.91
15.11
8.78
Polynomial13'0 (modeled variance)
0.6451
NA
113.20
17.37
8.49
Powerd (modeled variance)
0.4062
NA
112.69
17.72
9.00
Hilld (modeled variance)
NA
aValues <0.1 fail to meet conventional goodness-of-fit criteria.
b2-degree polynomial; no adequate fit with any polydegree.
°betas restricted to > 0.
dpower restricted to >1.
AIC = Akaike's Information Criteria; p = p value from the Chi-squared test; BMD = benchmark dose; BMDL =
lower confidence limit (95%) on the benchmark dose; NA = not available (BMD software could not generate a
model output).
28
-------
5-21-2007
Linear Model with 0.95 Confidence Level
100
90
80
-7 n
70
c
o
CL
a>
a- 60
c
ra
------- |