United States
Environmental Protection
1=1 m m Agency
EPA/690/R-09/016F
Final
9-28-2009
Provisional Peer-Reviewed Toxicity Values for
Diethylene Glycol Monobutyl Ether
(DGBE, CASRN 112-34-5)
Superfund Health Risk Technical Support Center
National Center for Environmental Assessment
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268

-------
COMMONLY USED ABBREVIATIONS
BMD
Benchmark Dose
IRIS
Integrated Risk Information System
IUR
inhalation unit risk
LOAEL
lowest-observed-adverse-effect level
LOAELadj
LOAEL adjusted to continuous exposure duration
LOAELhec
LOAEL adjusted for dosimetric differences across species to a human
NOAEL
no-ob served-adverse-effect level
NOAELadj
NOAEL adjusted to continuous exposure duration
NOAELhec
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
p-RfD
provisional oral reference dose
RfC
inhalation reference concentration
RfD
oral reference dose
UF
uncertainty factor
UFa
animal to human uncertainty factor
UFC
composite uncertainty factor
UFd
incomplete to complete database uncertainty factor
UFh
interhuman uncertainty factor
UFl
LOAEL to NOAEL uncertainty factor
UFS
subchronic to chronic uncertainty factor
1

-------
FINAL
9-28-2009
PROVISIONAL PEER-REVIEWED TOXICITY VALUES FOR
DIETHYLENE GLYCOL MONOBUTYL ETHER (DGBE, CASRN 112-34-5)
Background
On December 5, 2003, the U.S. Environmental Protection Agency's (U.S. EPA) 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)	U.S. EPA's Integrated Risk Information System (IRIS)
2)	Provisional Peer-Reviewed Toxicity Values (PPRTVs) used in U.S. 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
~	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 U.S. EPA's 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 U.S. EPA IRIS Program. All provisional toxicity values receive internal
review by two U.S. EPA scientists and external peer review by three independently selected
scientific experts. PPRTVs differ from IRIS values in that PPRTVs do not receive the
multiprogram consensus review provided for IRIS values. This is because IRIS values are
generally intended to be used in all U.S. 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 5-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 documents 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 Resource Conservation and Recovery Act (RCRA) program offices are advised to
carefully review the information provided 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 document and understand the strengths
1

-------
FINAL
9-28-2009
and limitations of the derived provisional values. PPRTVs are developed by the U.S. EPA
Office of Research and Development's National Center for Environmental Assessment,
Superfund Health Risk Technical Support Center for OSRTI. Other U.S. 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 U.S. EPA Office of Research and Development's National Center for Environmental
Assessment, Superfund Health Risk Technical Support Center (513-569-7300), or OSRTI.
Diethylene glycol monobutyl ether (DGBE) is a colorless liquid. Its synonyms include
2-(2-butoxyethoxy)-ethanol, butoxydiglycol, butadigol, and butyl carbitol. The formula for
DGBE is C8H18O3 (Figure 1) with a molecular weight of 162.23, boiling temperature of
230.4 C°, and melting point of -68.IC0.
There is no RfD assessment for DGBE on IRIS (U.S. EPA, 2008) or in the HEAST
(U.S. EPA, 1997) or Drinking Water Standards and Health Advisories list (U.S. EPA, 2006).
Subchronic and chronic RfDs of 0.04 and 0.004 mg/kg-day, respectively, were derived for
DGBE in a draft Health and Environmental Effects Document (HEED) on glycol ethers
(SRC, 1992). The RfDs in the draft HEED were based on a subchronic LOAEL of
36 mg/kg-day for hematological effects in rats (Hobson et al., 1987) and an UF of 1,000
(subchronic RfD) or 10,000 (chronic RfD). An earlier Health Effects Assessment (HEA) on
glycol ethers (U.S. EPA, 1984) did not derive acceptable subchronic or chronic oral intake
values for DGBE. This HEA is the only document relevant to DGBE in the Chemical
Assessments and Related Activities (CARA) list (U.S. EPA, 1991, 1994a).
There is no RfC assessment for DGBE on IRIS (U.S. EPA, 2008) or in the HEAST
(U.S. EPA, 1997). No occupational exposure limits have been recommended or promulgated by
the American Conference of Governmental Industrial Hygienists (ACGIH, 2001, 2007), the
National Institute of Occupational Safety and Health (NIOSH, 2005), or the Occupational Safety
and Health Administration (OSHA, 2008).
INTRODUCTION
Figure 1. Diethylene Glycol Monobutyl Ether Structure
2

-------
FINAL
9-28-2009
DGBE is not listed in the HEAST cancer table (U.S. EPA, 1997), and no carcinogenicity
assessment is available on IRIS (U.S. EPA, 2008) or indicated in the Drinking Water Standards
and Health Advisories list (U.S. EPA, 2006). The carcinogenicity of DGBE has not been
assessed by the International Agency for Research on Cancer (IARC, 2008) or the National
Toxicology Program (NTP, 2005, 2008).
There are no Toxicological Profiles available from ATSDR (2008) or World Health
Organization Environmental Health Criteria documents (WHO, 2008) for DGBE. CalEPA
(2002, 2005a,b) has not derived oral or inhalation recommended exposure limits (RELs) or a
cancer potency factor for DGBE.
Literature searches were conducted from the 1960s through August 2009 for studies
relevant to the derivation of provisional toxicity values for DGBE. Databases searched included:
MEDLINE (including cancer subset), TOXLINE (Special), BIOSIS, TSCATS 1/TSCATS 2,
CCRIS, DART/ETIC, GENETOX, HSDB, RTECS, and Current Contents (February-August
2009).
REVIEW OF PERTINENT DATA
Human Studies
No pertinent data were located regarding health effects of DGBE in humans following
oral or inhalation exposure.
Animal Studies
Animal studies are summarized in Table 1.
Oral Exposure
Short-term Studies—As a preliminary study to a 13-week oral subchronic toxicity
investigation discussed above, Johnson et al. (2005) administered DGBE (99.2% pure) in
drinking water to Fischer 344 (F344) rats (five/sex/group) at doses of 0, 1,000, 1,500, or
2,000 mg/kg-day for 2 weeks. Animals were evaluated for clinical signs, body weight, food
consumption, water consumption, urinalysis variables, serum chemistry, and hematological
variables, and were given a complete gross necropsy at termination. There were no effects on
mortality and no clinical signs other than urine soiling (all doses). Body weight, food
consumption, and water consumption were reduced at all doses (but not statistically significant),
with effects greatest in high-dose males (14% lower body weight than controls by study
termination). Red blood cell count (RBC), hemoglobin (Hgb), and hematocrit (Hct) were
decreased (not statistically significant) relative to controls at all doses, with the largest decreases
observed in the 2,000 mg/kg-day dose group (all variables: 10—13% decrease for males, 5—7%
decrease for females). The data showed no statistically significant treatment-related adverse
effects on serum chemistry, urinalysis, or organ weights. The only effect observed at gross
necropsy was urine staining of the perineal fur (all doses).
3

-------
FINAL
9-28-2009
Table 1. Summary Table for Diethylene Glycol Monobutyl Ether
Species and
Study Type
Exposure
Critical Effects
NOAEL
LOAEL
Comments
Reference
Oral Studies
Rat
Short-term
0,51,94,210, 650,
970, or
1,830 mg/kg-day in
drinking water for
30 days.
Reduced water intake at
>94 mg/kg-day and
histopathological changes
in the liver (congestion
and slight cloudy
swelling) and kidney
tubules (cloudy swelling
and increased secretion) at
>650 mg/kg-day.
Cannot be
determined
Cannot be
determined
Endpoints included mortality, water
intake, body weight and histology of the
liver, kidneys, spleen, and testis. Small
groups of rats (five/sex/dose). This is a
range-finding study that is limited in
scope and inadequately reported.
Smyth, 1940.
Smyth and Carpenter,
1948.
Rat
Subchronic
891, 1,782, or
3,564 mg/kg-day by
gavage on
5	days/week, for
6	weeks.
(average daily dose:
636, 1,273, or
2,546 mg/kg-day)
Hyperkeratosis of the
stomach at
>891 mg/kg-day and
hematologic effects
(reduced erythrocyte
count, hemoglobin
concentration, and MCHC
and increased MCV) at
>1,782 mg/kg-day.
891
mg/kg-day
(average
daily dose of
636
mg/kg-day)
1,782
mg/kg-day
(average daily
dose of 1,273
mg/kg-day)
Hematological
effects
Endpoints included clinical signs, food
consumption, body weight, hematology,
clinical chemistry, organ weights, gross
pathology, and histopathology. Human
relevance of gastric hyperkeratosis is
questionable due to bolus exposure and
likely direct irritant properties of
DGBE.
Kodak, 1984.
4

-------
FINAL
9-28-2009
Table 1. Summary Table for Diethylene Glycol Monobutyl Ether
Species and
Study Type
Exposure
Critical Effects
NOAEL
LOAEL
Comments
Reference
Rat
Subchronic
0, 70, 330, or
1,630 mg/kg-day
(males) or 0, 50, 250,
or 1,270 mg/kg-day
(females) by gavage
on 5 days/week, for
13 weeks.
(Average daily dose:
0, 50, 236, or
1,164 mg/kg-day
[males]; 0, 36, 179,
or 907 mg/kg-day
[females].)
Decreased total WBC and
lymphocyte counts and
MCHC in females at
>50 mg/kg-day. Mortality
in both sexes at >250/330
mg/kg-day (some of the
deaths were due to gavage
error).
N/A
50 mg/kg-day
(average daily
dose of
36 mg/kg-day)
Lymphopenia
Endpoints included clinical signs, food
consumption, body weight, hematology,
clinical chemistry, urinalysis, organ
weights, gross pathology, and
histopathology. The histological
examinations did not include the
stomach.
Hobsonetal, 1987.
Rat
Subchronic
0, 50, 250, or
1,000 mg/kg-day in
drinking water for
13 weeks.
Decreased RBC count,
hemoglobin, and
hematocrit at
>250 mg/kg-day. Other
effects only occurred at
1,000 mg/kg-day and
mainly involved the liver;
these included increases
in organ weight and
hepatic cytochrome P450s
and UGT levels, decreases
in serum total protein,
cholesterol, and serum
AST, and hepatocyte
hypertrophy and
individual hepatocyte
degeneration.
50 mg/kg-day
250mg/kg-day
Reduced RBC
count and Hgb
in both sexes.
Endpoints included clinical signs, food
and water consumption, body weight,
hematology, clinical chemistry,
urinalysis, functional observational
battery, sperm analysis, liver metabolic
enzymes, organ weights, gross
pathology, and histopathology. Study
possibly useful for subchronic and
chronic p-RfD derivation.
Johnson etal, 2005.
5

-------
FINAL
9-28-2009
Table 1. Summary Table for Diethylene Glycol Monobutyl Ether
Species and
Study Type
Exposure
Critical Effects
NOAEL
LOAEL
Comments
Reference
Rat
Reproductive
0, 250, 500, or
1,000 mg/kg-day by
gavage in water.
Untreated males were
bred to treated
females and vice
versa. The males
were treated for
60 days prior to
mating, and the
females were treated
from 14 days prior to
mating to GD 13 or
until weaning of the
offspring.
Slight reduction in pup
weight during the last
week of lactation in the
offspring of the females
dosed with 1,000 mg/kg-
day. No maternal toxicity
or reproductive effects
were observed.
500
mg/kg-day
1,000
mg/kg-day
Decreased pup
weight during
the last week
of lactation.
Endpoints included body weight in both
sexes; numbers of corpora lutea,
implants, resorptions, viable embryos
and live or dead pups; and clinical
condition, body weight, and external
development in pups.
Nolenetal, 1985.
Rat
Reproductive
and
Developmental
0, 25, 115, or
633 mg/kg-day in
diet on GD 0-20.
Rats were killed on
GD 20 for uterine
and fetal
examinations or
allowed to deliver
pups that were reared
until 10 weeks of age.
Reduced maternal body
weight gain during
pregnancy at >25
mg/kg-day (the effect was
variable and not
dose-related). No prenatal
or postnatal
developmental toxicity.
633
mg/kg-day

Maternal endpoints included clinical
signs, food consumption, and body
weight. Developmental endpoints
included numbers of corpora lutea,
implantations, litters, and live fetuses
per litter, pre and postimplantation
losses, fetal and placental weights, sex
ratio, external and oral cavity anomalies,
skeletal and internal anomalies,
gestation length, numbers of live
newborns, and pup body weight and
survival.
Emaetal., 1988.
Mouse
Developmental
0 or 500 mg/kg-day
gavage in water on
GD 7-14.
No maternal or
developmental toxicity.
500
mg/kg-day

Developmental toxicity screening assay.
Study endpoints included maternal body
weight, fetal survival, pup perinatal and
postnatal survival, and pup body weight.
Teratogenicity not evaluated.
Schuleretal., 1984.
6

-------
FINAL
9-28-2009
Table 1. Summary Table for Diethylene Glycol Monobutyl Ether
Species and
Study Type
Exposure
Critical Effects
NOAEL
LOAEL
Comments
Reference
Inhalation Study
Rat
Subchronic
0, 2, 6, or 18 ppm (0,
13, 40, or
119 mg/m3) for
6 hours/day,
5 days/week, for up
to 22 exposures in
5 weeks.
(average daily
concentration: 0, 2.3,
7.1, or 21.3 mg/m3)
Liver histopathology
(slight hepatocyte
vacuolization consistent
with fatty change) in
females at >6 ppm and
gross paleness of the liver
in 3/10 females at 18 ppm.
2 ppm
(average
concentration
of 13 mg/m3)
6 ppm (average
concentration
of 40 mg/m3)
Endpoints included clinical signs, body
weight, hematology, clinical chemistry,
urinalysis, organ weights, gross
pathology, and histopathology.
Histological exams were comprehensive
at 0 and 18 ppm but limited to the liver
at 2 and 6 ppm. Study used to derive
subchronic and chronic RfCs in
previous PPRTV document.
Gushowetal., 1984.
7

-------
FINAL
9-28-2009
In an earlier drinking water study, groups of Sherman rats (5/sex/group) were exposed to
DGBE (reported as butyl carbitol) in drinking water at reported doses of 0, 51, 94, 210, 650, 970,
or 1,830 mg/kg-day for 30 days (Smyth, 1940; Smyth and Carpenter, 1948). The study was
summarized only briefly. Endpoints included mortality, water intake, body weight, and
histology of the liver, kidney, spleen, and testis. No effects were observed at 51 mg/kg-day.
Deaths did not occur at any dose. Water intake was reduced at >94 mg/kg-day, but growth was
not statistically significantly affected at any dose level. Histopathological changes occurred at
>650 mg/kg-day that included cloudy swelling and increased secretion in the kidney tubules and
congestion and slight cloudy swelling of the liver. The limited scope and inadequate reporting of
this study precludes identification of a NOAEL or LOAEL.
Subchronic Studies—Groups of 10 male albino rats (Charles River COBS, CD, BR)
were administered 0, 891, 1,782, or 3,564 mg/kg-day doses of undiluted DGBE (>99.5% pure)
by gavage on 5 days/week, for 6 weeks (Kodak, 1984). The group of control rats was treated
with distilled water equal in volume to that of the highest-dose group. Endpoints included
clinical condition, food consumption and body weight, measured hematology values (Hgb, Hct,
RBC, and total and differential white blood cell counts [WBC]) and calculated red blood cell
indices (mean corpuscular volume [MCV], mean corpuscular hemoglobin [MCH], and mean
corpuscular hemoglobin concentration [MCHC]), several serum chemistry indicators of liver
damage and kidney function (alanine aminotransferase [ALT or SGPT], aspartate
aminotransferase [AST or SGOT], alkaline phosphatase [ALP], lactate dehydrogenase [LDH],
blood urea nitrogen [BUN], creatinine, and glucose), organ weights (liver, kidneys, heart, testes,
brain, and spleen), and comprehensive gross and histopathology (-30 tissues). Deaths attributed
to gavage error occurred in the high- (4/10), mid- (2/10), and low- (1/10) dose groups. An
additional two rats in the high-dose group were killed in moribund condition. Effects observed
in the high-dose group included clinical signs of toxicity (dyspnea, prostration, and unkempt hair
coat, as well as bloody urine and blood around the nose and mouth in one animal), reduced
body-weight gain and food consumption, and histopathological changes in the spleen
(congestion, red pulp hypocellularity, and hemosiderin-like pigmentation). Statistically
significant effects occurring in the mid- and high-dose groups included hematologic indications
of RBC damage (clear dose-related reductions in RBC [14% and 28%], Hgb [11% and 15%],
and MCHC [9% and 18%] and increases in MCV [14% and 45%] and MCH [18% at high dose
only], with no effect on total or differential WBC), decreased serum glucose concentration,
increased absolute and relative liver and spleen weights, and histopathological changes in the
kidneys (hyaline droplet degeneration, proteinaceous casts, and hemosiderin in the proximal
tubules). The investigators noted that the renal proteinaceous casts and hemosiderin appeared to
be compound-related but may have been secondary to the hematological effects, and that the
significance of the hyaline droplet degeneration is uncertain because it was also seen in all
10 control rats. The only clear treatment-related effect at the low dose was hyperkeratosis of the
stomach (10/10), which was also observed in the mid- and high-dose groups (10/10 and
8/10 incidences, respectively) compared to 0/10 in the controls. The human relevance of the
gastric hyperkeratosis is questionable due to the bolus (undiluted gavage) type of exposures and
likely direct irritant properties of DGBE, as indicated by aspiration-related respiratory tract
lesions after gavage exposure in the Hobson et al. (1987) study summarized below. Based on the
hematological effects, the NOAEL and LOAEL of 891 and 1,782 mg/kg-day were identified, and
the corresponding average daily doses were 636 and 1,273 mg/kg-day.
8

-------
FINAL
9-28-2009
A 13-week study was conducted, in which groups of 16 male and 16 female F344 rats
were treated with DGBE (99% pure) by gavage in water (0.2% of body weight) 5 days/week
(Hobson et al., 1987). Reported dose levels were 0, 70, 330, and 1,630 mg/kg-day in the males
and 0, 50, 250, and 1,270 mg/kg-day in the females. An interim sacrifice was conducted at
6 weeks, after which all groups except the high dose consisted of 10 rats/sex; the high-dose
group had 4 rats/sex due to early mortality, as discussed below. Endpoints included clinical
condition, food consumption and body weight, hematology (RBC and platelet counts, Hgb, Hct,
MCV, MCHC, and total and differential WBC), serum chemistry (11 indices, including BUN),
urine chemistry (14 indices, including blood in urine), organ weights, and comprehensive gross
and histopathology (-30 tissues, but not including stomach). A dose-related increase in
mortality occurred in mid- and high-dose rats of both sexes. Deaths occurred in mid-dose males
starting in Week 8, mid-dose females starting in Week 4, and high-dose rats of both sexes
starting in Week 1. Surviving to termination were 10, 10, 4, and 2 males and 10, 9, 8, and
1 females in the control, low-, mid-, and high-dose groups, respectively. The small numbers of
survivors in the high-dose group precluded statistical analysis of endpoints evaluated at
termination for this group.
Other effects in the mid- and high-dose rats of both sexes included inflammatory lesions
in the respiratory tract (Hobson et al., 1987). Respiratory tract lesions that were commonly
observed in the females at >250 mg/kg-day and males at >330 mg/kg-day included acute rhinitis,
laryngitis, tracheitis, pulmonary congestion, and edema. Mild squamous metaplasia of the nasal
epithelium was found in two mid-dose and five high-dose males and one high-dose female. The
study authors suggested that these effects, as well as sporadic observations of foreign body
pneumonia and acute pleuritis, indicated that gavage-related aspiration of DGBE may have
occurred, contributing to mortality in some of the rats. Study authors reported that over the
entire study, respiratory tract lesions in three males and six females in the mid-dose groups and
four males and five females in the high-dose groups were consistent with gavage accident.
Although a number of the deaths were clearly attributable to gavage error, the dose-related
distribution of the mortality suggests that some deaths were related to systemic toxicity.
Histological findings in the other mid- and high-dose rats (i.e., unclear pathogenesis of
respiratory lesions) and experimental observations by the dosing technicians also did not
completely support a definitive relationship between gavage procedure and mortality. The lack
of histological examinations of the stomach precludes possibly corroborating the gastric
hyperkeratosis observed in the Kodak (1984) 6-week study. Gastric hyperkeratosis is not
necessarily expected to have been induced in the Hobson et al. (1987) study due to testing of
lower doses diluted in water compared to higher undiluted doses in the Kodak (1984) study
discussed above.
Hematological effects after 13 weeks of DGBE exposure (Hobson et al., 1987) included
dose-related decreases in total WBC and lymphocyte counts in surviving females at 50 (by
32.1% in WBC and 34.9% in lymphocyte counts) and 250 mg/kg-day (by 42.6% in WBC and
47.9% in lymphocyte counts) (see Table 2). No decrease in these parameters was seen in the
lone surviving female at 1,270 mg/kg-day or in males at any dose at 13 weeks. RBC and Hgb
were not significantly reduced at any dose in either sex at 13 weeks. However, the researchers
reported that at 6 weeks (i.e., before all the animals died), RBC and Hgb were significantly
decreased (data not reported) in high-dose males and females, and there were dose-dependent
hemoglobinuria in males and females exposed to DGBE. In addition, lymphocyte counts were
significantly decreased (data not reported) in high-dose DGBE males; nevertheless, neutrophil
9

-------
FINAL
9-28-2009
counts were significantly increased in mid- and high-dose males at the interim sacrifice. At
13 weeks, nonhematological effects of DGBE that were statistically significant (p < 0.05)
included increased absolute liver weight in males at >70 mg/kg-day, increased relative liver
weight in males at 330 mg/kg-day, increased BUN and serum alkaline phosphatase values in
males at >330 mg/kg-day, and increased hyaline droplet formation in the renal tubular epithelium
in females at 1,270 mg/kg-day. The hyaline droplets in females were considered by the
researchers to be indicative of hemoglobinuria. Renal hyaline droplets were commonly observed
in males of all dose and control groups and thereby regarded as being within normal physiologic
limits. The available data indicate that hematological changes are the most sensitive effects of
DGBE and that the 50 mg/kg-day-dose (average daily dose of 36 mg/kg-day) is most
appropriately classified as a LOAEL based on the induction of lymphopenia in only females, and
no lymphopenia in the high-dose female rat. MCHC was also reduced in females at
>50 mg/kg-day, but is not clearly reflective of an adverse effect in the absence of significant
decreases in RBC count, Hgb, or Hct.
Table 2. Data Sets for Leukopenia/Lymphopenia in Female Rats
Exposed to DGBE by Gavage for 13 Weeks"
Endpoint (Mean ± SD)
Duration-Adjusted Doseb (mg/kg-day)
0
35.7
178.6
Number Examined
7
9
7
WBC count (x 103/mm3)
4.46 ±0.73
3.03 ±0.59c
(67.9%)d
2.56 ± 0.31°
(57.4%)
Lymphocyte count (x 103/mm3)
3.55 ±0.6
2.31 ±0.42c
(65.1%)
1.85 ± 0.31°
(52.1%)
aHobson et al. (1987)
bDuration-Adjusted Dose = Dose x 5/7 days
Statistically significantly different from controls (p < 0.05)
Percentage relative to the control
Note: the results from the high-dose group are not presented because this dose group was dropped from further
analysis due to high mortality.
Johnson et al. (2005) administered DGBE (99.2% pure) in drinking water daily to groups
(10 rats/sex) of F344 rats at doses of 0, 50, 250, or 1,000 mg/kg-day for 13 weeks. Animals were
observed twice daily for general appearance and were assessed weekly for detailed clinical
evaluation, body weight, and food and water consumption. Ophthalmological examinations and
a behavioral functional observational battery (FOB) were administered before initiation of
exposure and during the last week of the study. Clinical pathology, including hematology, serum
chemistry, and urinalysis, was assessed prior to sacrifice and necropsy. Evaluations of sperm
and liver metabolic enzymes were also made prior to necropsy. A comprehensive necropsy,
including organ-weight determinations, was conducted for all rats. Comprehensive
histopathological evaluations were made for control and high-dose rats, with subsequent
evaluations of selected organs from the low- and mid-dose groups as indicated based on findings
at the high dose.
10

-------
FINAL
9-28-2009
All rats survived to study termination, and other than some urine soiling in the high-dose
group, there were no treatment-related clinical signs (Johnson et al., 2005). Decreases in water
and food consumption and concomitant reductions in body weight were noted in both sexes at
the high dose throughout the study, with the greatest reductions noted in males; body weight at
the end of the study was reduced by 10% and 6% in males and females, respectively, in
comparison with controls. Water consumption among high-dose rats was 7-8% less than
controls throughout the study, and weekly food consumption among high-dose rats was
5-11% less than control values. No treatment-related effects on body weight or food or water
consumption were noted in the mid- and low-dose groups. No treatment-related effects on
urinalysis, ophthalmology or the FOB (sensory evaluation, rectal temperature, grip performance,
motor activity) were observed for any treatment group (data not shown). There were small, but
statistically significant, dose-related decreases in RBC hematological variables relative to
controls in mid- (RBC, Hgb) and high-dose (RBC, Hgb, Hct) rats of both sexes. The decreases
ranged from 2-4% at 250 mg/kg-day and 5—9% at 1,000 mg/kg-day. RBC counts in both sexes
and Hgb in males were outside of the historical control range for the similar study in the same
strain rats (see Table 3). There were no other treatment-related effects on hematological
variables, including total and differential WBC (data not shown).
Table 3. Data Sets for Anemia in Male and Female Rats
Exposed to DGBE in Drinking Water for 13 Weeksa
Endpoint
Dose (mg/kg-day)
0
50
250
1000
Males
Number Examined
10
10
10
10
RBC Count (x 106/(iL)
9.27 ±0.35
9.13 ±0.22
(98.5%)c
8.94 ± 0.34b
(96.4%)
8.53 ±0.31b
(92.0%)
Hgb (g/dL)
15.8 ±0.5
15.7 ±0.4
(99.3%)
15.3 ±0.4b
(96.8%)
14.8 ± 0.4b
(93.6%)
Females
Number Examined
10
10
10
10
RBC Count (x lOV^L)
8.26 ±0.22
8.06 ±0.31
(97.6%)
8.07 ± 0.24b
(97.7%)
7.54 ± 0.17b
(91.3%)
Hgb (g/dL)
15.6 ±0.3
15.2 ±0.4
(97.4%)
15.2 ± 0.4b
(97.4%)
14.8 ± 0.3b
(94.9%)
aJohnson et al. (2005)
Statistically significantly different from controls (p < 0.05)
Percentage relative to the control
Note: historical control ranges: RBC in males 9.11-9.22 (106/|jL), RBC in females 7.81-8.20 (106/|jL), Hgb in
males 15.0-15.6 (g/dL), and Hgb in females 14.3-14.9 (g/dL).
11

-------
FINAL
9-28-2009
Clinical chemistry findings were generally unremarkable, although there were slight
decreases in serum protein, cholesterol, and AST in high-dose males and females (Johnson et al.,
2005). Liver metabolic enzymes were slightly increased relative to controls only for high-dose
animals as follows: EROD1 (24% both sexes); PROD2 (38% males, 24%, females); and UGT3
(17%) males, 16% females). Organ weight results included increased absolute and relative liver
and kidney weights in the high-dose groups of both sexes, despite decreased body weight in
these groups. Absolute and/or relative spleen weights were increased in all male and female
dose groups, but the changes from controls were small (absolute weight changes <6% in males
and <3% in females) and did not increase with dose. The only treatment-related
histopathological changes were noted in the livers of high-dose females. These changes were
considered by the study authors to be treatment-related and consisted of hepatocellular
hypertrophy ("very slight" in 6/10 high-dose females) and foci of necrotic cells in the
centrolobular region. No treatment-related histopathologic changes were noted in any other
tissues, including kidneys, spleen, or male livers (incidence data not shown). No
treatment-related effects on sperm count, sperm morphology or motility, or male reproductive
tissues were observed (data not shown). The NOAEL for this study is 50 mg/kg-day, and the
LOAEL is 250 mg/kg-day on the basis of significantly reduced RBC count and Hgb in both
sexes.
Chronic Studies—No pertinent data were located regarding health effects of DGBE in
animals following chronic oral exposure.
Reproductive/Developmental Studies—Groups of 25 male and 25 female Charles River
CD rats were treated with 0, 250, 500, or 1,000 mg/kg-day doses of DGBE (95% pure) by
gavage in distilled water in a fertility study (Nolen et al., 1985). Untreated males were bred to
treated females and vice versa. The males were treated for 60 days prior to mating, and the
females were treated from 14 days prior to mating to gestation day (GD) 13 or until weaning of
the offspring. Half of the females in each group were killed on GD 13 for examination of uterine
contents. The remaining females delivered their young, and the offspring were followed to
weaning. Endpoints included weekly body weight in both sexes; numbers of corpora lutea,
implants, resorptions, viable embryos, and live or dead pups; and clinical condition, body weight,
and external development in pups. The only effect attributable to treatment was a slight (-8%)
reduction in pup weight during the last week of lactation among the offspring of the females
dosed with 1,000 mg/kg-day relative to the control group. The NOAEL and LOAEL were 500
and 1,000 mg/kg-day, respectively, for mild toxicity to the neonate, while no maternal or
reproductive effects were found even at 1,000 mg/kg-day.
These findings are supported by another study in rats. Groups of 19-21 pregnant Wistar
rats were fed DGBE (purity not reported) in the diet at reported intake levels of 0, 25, 115, or
633 mg/kg-day from GD 0-20 (Ema et al., 1988). Some 14 or 15 rats in each group were killed
on Day 20 for uterine and fetal examinations, and the remaining 5 or 6 rats/group were allowed
to deliver spontaneously. From each litter, eight pups (four rats/sex) were reared until 10 weeks
1 Ethoxyresorufin-O-dealky lase.
2Pentoxyresorufin-0-dealkylase.
3UDP-glucuronosyltransferase.
12

-------
FINAL
9-28-2009
of age. Maternal endpoints included daily evaluations for clinical signs of toxicity and effects on
food consumption and body weight. Prenatal developmental endpoints included numbers of
litters and corpora lutea, implantations and live fetuses per litter, pre and postimplantation losses,
fetal and placental weights, sex ratio, external and oral cavity anomalies, skeletal anomalies
(approximately half the fetuses in each litter), and internal anomalies (remaining half of the
fetuses). Postnatal endpoints included gestation length, numbers of live newborns, body weight,
and survival rate. The only effect noticed was reduced maternal body-weight gain during
pregnancy at >25 mg/kg-day (12-18% lower than controls,/? < 0.05). The maternal body-weight
gain is not only affected by changes in the maternal weight, but also changes in litter size and
fetal body weight. Thus, the absolute changes in maternal weight during the gestation in any of
these dose groups were relatively small (<10%); therefore, this endpoint is not considered to be
an adverse response to DGBE. The lack of any prenatal or postnatal developmental effects
indicates that 633 mg/kg-day is a NOAEL for developmental as well as maternal toxicity.
Similar results were found in mice. In a screening assay, groups of 50 pregnant CD-I
mice were treated with 0 or 500 mg/kg doses of DGBE (>99% pure) by gavage in aqueous
solution on GD 7-14 (Schuler et al., 1984). Study endpoints included pup survival in utero
(percent of live litters/pregnant survivors), pup perinatal and postnatal survival (numbers of live
and dead pups per litter and pup survival to age 2.5 days), and pup body weights (at birth and age
2.5 days). Maternal indices included body weight on GD 7, GD 18, and Day 3 postpartum. No
treatment-related maternal, litter, or pup effects were observed, indicating that 500 mg/kg-day
was a NOAEL for maternal and developmental toxicity.
Inhalation Exposure
Chronic Studies—No pertinent data were located regarding health effects of DGBE in
animals following chronic inhalation exposure.
Subchronic Studies—In a subchronic inhalation study, groups of 15 male and 15 female
F344 rats were exposed to 0, 2, 6, or 18 ppm of DGBE (-98.6% pure) (0, 13, 40, or 119 mg/m3)
for 6 hours/day, 5 days/week, for up to 22 exposures in 5 weeks (Gushow et al., 1984). The
119 mg/m3 concentration was the highest sustainable vapor concentration of DGBE due to its
low vapor pressure. RBC fragility was tested in five rats/sex/group following the 15th exposure.
Endpoints evaluated in the remaining 10 rats/sex/group included clinical condition and body
weight throughout the study; hematology (RBC, total and differential WBC, platelet, Hgb, Hct,
MCV, MCH, and MCHC), clinical chemistry (BUN, ALT, ALP, glucose, albumin, total protein,
and total bilirubin), and urine chemistry (pH, glucose, ketones, bilirubin, urobilinogen, occult
blood, protein, and specific gravity) after 16-22 exposures; and organ weights, gross pathology,
and histology after 22 exposures. The histological examinations were comprehensive (including
nasal turbinates and lungs) in the control and 119-mg/m3 groups but limited to the liver in the
"3
13 and 40 mg/m groups because this was the only tissue identified as a possible target at
119 mg/m3. Statistically significant (p < 0.05) decreases in serum glucose levels were observed
3	3
in females at >13 mg/m and males at >40 mg/m , but the reductions were minimal
(9-16%) lower than controls). Other effects included slightly increased WBC in males (within
the normal range of variation and not considered to be toxicologically significant by the
researchers) and hepatic changes in both sexes at >40 mg/m3. The hepatic effects included
inconsistent changes in relative liver weight (decreased in males at >40 mg/m and increased in
13

-------
FINAL
9-28-2009
females at 119 mg/m ) with no effects on absolute liver weight, histological changes (slight
hepatocyte vacuolization consistent with fatty change) in females at >40 mg/m3, and gross
3	3
paleness of the liver in 3/10 females at 119 mg/m . Hepatocytes in the control and 13-mg/m
females were vacuolated to a somewhat lesser degree (i.e., very slight), indicating that the effect
-3
was only minimally increased at >40 mg/m . No other biologically significant effects were
found in clinical chemistry, hematology, and urine chemistry. The low concentration of DGBE
(i.e., 13 mg/m ) is identified as aNOAEL based on significant incidences of hepatocyte
vacuolization consistent with fatty change that occurred in female rats at >40 mg/m3 and is
supported by increased relative liver weight at 119 mg/m .
Other Studies
There are several short-term and subchronic toxicity studies (Proctor and Gamble, 1982;
Bio/dynamics, Inc., 1989; Auletta et al., 1993), reproductive toxicity studies (Bio/dynamics, Inc.,
1989; Auletta et al., 1993), and a subchronic neurotoxicity study (Bio-Research Labs, 1989) with
dermal exposure to DGBE. The only changes reported by Proctor and Gamble (1982)
(hematological parameters or hematuria) and those (occult blood in the urine) reported by
Bio/dynamics, Inc. (1989) and Auletta et al. (1993) are consistent with the effects that were seen
following oral exposure to DGBE.
Genotoxicity
The genotoxicity of DGBE has been tested in several assay systems. In in vitro assays,
DGBE was negative for reverse mutation in Salmonella typhimurium with or without metabolic
activation (Thompson et al., 1984; Zeiger et al., 1992), unscheduled DNA synthesis in primary
rat hepatocytes (Thompson et al., 1984), forward mutation (HGPRT locus) in Chinese hamster
ovary (CHO) cells (Gollapudi et al., 1993) and sister chromatid exchanges (SCEs) in CHO cells
(Thompson et al., 1984), and weakly positive for forward mutations in mouse lymphoma
L5178Y cells (Thompson et al., 1984). In vivo testing of DGBE was negative for sex-linked
recessive mutations in Drosophila melanogaster (Thompson et al., 1984) and induction of
micronuclei in bone marrow cells of mice (Gollapudi et al., 1993).
DERIVATION OF PROVISIONAL SUBCHRONIC AND CHRONIC ORAL RfD
VALUES FOR DIETHYLENE GLYCOL MONOBUTYL ETHER (DGBE)
Information on systemic effects of repeated oral exposure to DGBE is available from
several studies in rats as summarized in Table 1. A range-finding study was conducted in which
groups of five Sherman rats/sex that were exposed to DGBE in drinking water at doses ranging
from 51-1,830 mg/kg-day for 30 days had no effects at 51 mg/kg-day, reduced water intake at
>94 mg/kg-day, and histopathological changes in the kidneys and liver at >650 mg/kg-day
(Smyth, 1940; Smyth and Carpenter, 1948). However, a NOAEL or LOAEL cannot be
identified from this study due to the limited number of endpoints investigated and inadequate
reporting. A more recent range-finding study (Johnson et al., 2005) in which groups of five
F344 rats/sex were exposed to DGBE in drinking water at doses of 1,000-2,000 mg/kg-day for
2 weeks noted decreases in body weight, food consumption, water consumption, RBC, Hgb, and
Hct; however, these changes were not statistically significant in comparison with controls.
14

-------
FINAL
9-28-2009
In a 6-week gavage study, DGBE was administered to groups of 10 male rats (Charles
River COBS, CD, BR) at doses of 891, 1,782, or 3,564 mg/kg-day on 5 days/week, for 6 weeks
(Kodak, 1984). This study identified a hematological NOAEL and LOAEL of 891 and
1,782 mg/kg-day (average daily doses of 636 and 1,273 mg/kg-day), respectively, based on
reduced RBC counts, Hgb, and MCHC, with related effects, including bloody urine, occurring at
3,564 mg/kg-day. In a 13-week study, DGBE was administered to groups of 16 male/16 female
F344 rats at doses of 0, 50/70, 250/330, or 1,270/1,630 mg/kg-day by gavage on 5 days/week, for
13 weeks (Hobson et al., 1987). This study identified a LOAEL of 50 mg/kg-day (36 mg/kg-day
average daily dose) based on leukopenia and lymphopenia in female rats (Hobson et al., 1987).
In a subsequent comprehensive drinking water study, groups of 10 F344 rats/sex were exposed to
DGBE in drinking water at doses of 0, 50, 250, and 1,000 mg/kg-day for 13 weeks
(Johnson et al., 2005). This study identified a NOAEL of 50 mg/kg-day based on the critical
effects of decreased RBC and Hgb noted at doses >250 mg/kg-day in both sexes.
The database for DGBE consistently indicates that hematological effects are the most
sensitive response after exposure to DGBE; however, inconsistent findings of leukopenia and
lymphopenia were reported. The most sensitive response was reported from Hobson et al.
(1987), noting leukopenia and lymphopenia in female rats at doses >50 mg/kg-day. However,
confidence in the findings from this study (Hobson et al., 1987) is limited by gavage accidents
and associated deaths and respiratory tract pathology in the higher-dose groups
(>250 mg/kg-day), inconsistent changes in neutrophil counts (increased) in male rats in the mid-
and high-dose groups (>330 mg/kg-day) at 6 weeks, lack of leukopenia and lymphopenia in male
rats at the end of experiment, and no similar responses in the lone surviving female in the
highest-dose group. In addition, neither leukopenia nor lymphopenia was reported in the same
strain of rats in a 13-week drinking water study (Johnson et al., 2005) at dose levels as high as
1,000 mg/kg-day, and no similar effects were found in male rats in the Kodak (1984) 6-week
gavage study (females were not tested) at doses at high as 3,564 mg/kg-day. In contrast to the
leukopenia and lymphopenia responses, consistent findings of decreased RBC counts and Hgb
levelswere reported in male rats treated with DGBE >1,782 mg/kg-day for 6-weeks (Kodak,
1984), in male rats at a dose level of 1,630 mg/kg-day and female rats at a dose level of
1,270 mg/kg-day at 6-week interim sacrifice (Hobson et al., 1987), and in both sexes at a dose
level of >250 mg/kg-day at the end of 13 weeks of exposure (Johnson et al., 2005). Therefore,
decreased RBC counts and Hgb levels after 13-weeks of treatment (Johnson et al., 2005) are
considered the critical effects for the DGBE-induced hematological response.
There are no indications that reproductive or developmental toxicity are effects of
concern for DGBE, as multiple oral studies in rats and mice found no effects at doses of 500 and
633 mg/kg-day and only a mild, transitory effect on neonate weight at 1,000 mg/kg-day
(Ema et al., 1988; Nolen et al., 1985; Schuler et al., 1984). Dermal studies also showed no
effects on reproduction or development (Bio/dynamics, Inc., 1989; Nolen et al., 1985).
There are no indications that neurotoxicities are effects of concern for DGBE, as an oral
study in rats found no behavioral effects at doses of 50, 250, or 1,000 mg/kg-day (Johnson et al.,
2005). A dermal study in rats also showed no effects on a number of neurotoxicity endpoints at
doses corresponding to 200, 600, or 2,000 mg/kg (Bio-Research Labs, 1989).
15

-------
FINAL
9-28-2009
Based on the available studies, a potential point of departure (POD) for the derivation of
p-RfD values is the NOAEL of 50 mg/kg-day for anemia (reduced RBC counts, Hgb levels) in
rats (Johnson et al., 2005). The relevant data sets for these endpoints are shown in Table 3
(Johnson et al., 2005). No duration adjustment for the doses for the Johnson et al. (2005) study
is needed because exposure was continuous for the duration of the study. Attempts to apply
BMD modeling to the data sets (RBC and Hgb) were successful for RBC counts in male (BMDL
of 81 mg/kg-day) and female rats (BMDL of 280 mg/kg-day), and for Hgb levels in males
(BMDL of 328 mg/kg-day) but not in females (Appendix A). Because the changes in Hgb levels
in females were comparable to those in males (see Table 3), a POD for this endpoint is expected
to be at the range close to 328 mg/kg-day based on male data. Thus, the lowest BMDL of
81 mg/kg-day (based on changes in RBC in male rats) among these four endpoints is the most
sensitive POD and is used for the derivation of both subchronic and chronic p-RfDs.
Subchronic p-RfD
A subchronic p-RfD is derived by applying an UF of 300 to the BMDL of 81 mg/kg-day
as follows:
Subchronic p-RfD = BMDL UF
= 81 mg/kg-day -^300
= 0.3 mg/kg-day or 3 x 10"1 mg/kg-day
The composite UF is composed of the following factors:
•	An UF of 10 is applied for interspecies extrapolation to account for potential
pharmacokinetic and pharmacodynamic differences between rats and humans.
•	An UF of 10 for intraspecies differences is used to account for potentially susceptible
individuals in the absence of information on the variability of response in humans.
•	An UF for LOAEL to NOAEL extrapolation is not needed because the POD is a
BMDL.
•	An UF of 3 (10°5) is applied to account for deficiencies in the database. The database
includes three subchronic studies in rats, two developmental studies in rats and mice,
and a one-generation reproductive study in rats; however, the database lacks a
multigeneration reproduction study and a subchronic study in a second species.
Confidence in the principal study (Johnson et al., 2005) is medium. The principal study
was well designed, and the critical effect is consistent with that found from other short-term
studies (Kodak, 1984, Hobson et al., 1987). However, a 13-week study (Hobson et al., 1987)
exposed F344 rats to DGBE by gavage for the same duration but did not identify a similar
response at the end of treatment at doses of >1,270 mg/kg-day. Instead, this study
(Hobson et al., 1987) identified a LOAEL of 50 mg/kg-day for decreased WBC and lymphocyte
counts, which were not shown by Johnson et al. (2005). Confidence in the database is medium
due to some inconsistent findings from subchronic studies (Johnson et al., 2005; Hobson et al.,
1987), a lack of a multigeneration reproduction study, and a subchronic study conducted in a
second species. Overall, confidence in the subchronic p-RfD is medium.
16

-------
FINAL
9-28-2009
Chronic p-RfD
A chronic p-RfD is similarly derived by applying an UF of 3,000 to the BMDL of
81 mg/kg-day as follows:
Chronic p-RfD = BMDL UF
= 81 mg/kg-day3,000
= 0.03 mg/kg-day or 3 x 10"2 mg/kg-day
The composite UF is composed of the following factors:
•	An UF of 10 is applied for interspecies extrapolation to account for potential
pharmacokinetic and pharmacodynamic differences between rats and humans.
•	An UF of 10 for intraspecies differences is used to account for potentially susceptible
individuals in the absence of information on the variability of response in humans.
•	An UF for LOAEL to NOAEL extrapolation is not needed because the POD is a
BMDL.
•	An UF of 10 was applied for using a study with a subchronic duration of exposure to
approximate chronic exposure.
•	An UF of 3 (10°5) is applied to account for deficiencies in the database. The database
includes critical three subchronic studies in rats, two developmental studies in rats
and mice, and a one-generation reproductive study in rats; however, the database
lacks a multigeneration reproduction study and a subchronic study in a second
species.
Confidence in the principal study (Hobson et al., 1987) is medium, as discussed above for
the subchronic p-RfD. Confidence in the database is low due to some inconsistent findings from
subchronic studies (Johnson et al., 2005; Hobson et al., 1987), a lack of chronic oral toxicity
studies, and a lack of a multigeneration reproduction study. Overall, confidence in the chronic
p-RfD is low.
DERIVATION OF PROVISIONAL SUBCHRONIC AND CHRONIC INHALATION RfC
VALUES FOR DIETHYLENE GLYCOL MONOBUTYL ETHER (DGBE)
Information on the inhalation toxicity of DGBE is essentially limited to the results of a
-3
comprehensive 5-week study in which rats were exposed to 0, 13, 40, or 119 mg/m for
6 hours/day, 5 days/week (Gushow et al., 1984). Evaluation of clinical condition, body weight,
hematology (including RBC fragility), blood and urine chemistry, organ weights, gross
pathology, and histopathology (including nasal turbinates and lungs) showed slight hepatocyte
-3
vacuolization consistent with fatty change in female rats at concentrations of >40 mg/m and
increased relative liver weight and gross paleness of the liver at 119 mg/m3. Although there
were inconsistent changes (decreases) in relative liver weight in male rats at >40 mg/m with no
effects on absolute liver weight, the hepatic changes in female rats were considered biologically
significant. The potential for hepatic effects is supported by findings of increased liver weight
and congestion/cloudy swelling in subchronic oral studies of DGBE (Johnson et al., 2005;
Hobson et al., 1987; Kodak, 1984; Smyth, 1940; Smyth and Carpenter, 1948). Exposure to
concentrations higher than 119 mg/m could not be tested because this level was the maximum
sustainable vapor concentration of DGBE.
17

-------
FINAL
9-28-2009
Subchronic and chronic p-RfCs were derived for DGBE using the data for hepatocelluar
vacuolization in female rats (Gushow et al., 1984) based on U.S. EPA (1994b) RfC
methodology. BMD modeling was conducted for the incidence data on hepatocyte vacuolization
(Table 4), and detailed results are presented in Appendix A. Because the exposure was not
continuous in the principal study (Gushow et al., 1984), duration-adjusted concentrations and
human equivalent concentration (HEC) (see Table 4) were calculated before BMD modeling.
An example of the duration adjustment calculation is presented below:
NOAELadj = NOAEL mg/m3 x 6/24 hr x 5/7 d
= 13 mg/m3 x 6/24 hr x 5/7 d
= 2.3 mg/m3
For purposes of deriving a p-RfC based on extrarespiratory effects, DGBE was treated as a
Category 3 gas. The HEC was calculated assuming periodicity was attained using the following
equation:
NOAELhec = NOAELadj x (Hb/g)A/(Hb/g)H
where (Hb/g)A/(Hb/g)H is the ratio of blood:gas (air) partition coefficients of the chemical in the
test animals and humans. Since a blood:gas partition coefficient is not available for DGBE in
humans or rats, a unity value is assumed for the ratio (U.S. EPA, 1994b). The BMD modeling
resulted in an estimated BMCioadj of 1.7 mg/m3 and BMCLioadj of 0.32 mg/m3, and the
BMCLioadj is equivalent to BMCLiohec-
Table 4. Data Sets for Hepatic Changes in Female Rats
Exposed to DGBE Through Inhalation for 5 Weeksa
Endpoint
Concentration (mg/m3)
0
13
40
119
Females
Duration-adjusted concentration
(mg/m3)b
0
2.3
7.1
21.3
Human equivalent concentration
(mg/m3)c
0
2.3
7.1
21.3
# Examined
3
4
10
10
Slight vacuolization consistent
with fatty change
3
4
9d
10d
aGushow et al. (1984)
bThe duration adjustment was calculated as follows:
ConcADj =119 mg/m3 x 6/24 hr x 5/7 d = 21.3 mg/m3
°HEC = ConcADj x (Hb/g)A/(Hb/g)H
Statistically significantly different from controls (p < 0.05)
18

-------
FINAL
9-28-2009
Subchronic p-RfC
A subchronic p-RfC is derived by applying an UF of 300 to the BMCLiohec of
"3
0.32 mg/m as follows:
Subchronic p-RfC = BMCLiohec UF
= 0.32 mg/m3 - 300
= 0.001 or 1 x 10 3 mg/m3
The composite UF of 300 is composed of the following factors:
•	An UF of 3 (10°5) is applied for interspecies extrapolation to account for potential
pharmacodynamic differences between rats and humans. Converting the rat data to
human equivalent concentrations by the dosimetric equations, accounts for
pharmacokinetic differences between rats and humans; thus, an UF of 10 for
interspecies extrapolation was not used.
•	An UF of 10 for intraspecies differences is used to account for potentially susceptible
individuals in the absence of information on the variability of response in humans.
•	An UF of 10 is applied to account for deficiencies in the database. The database
includes only one 5-week study; it lacks developmental toxicity studies and a
multigeneration reproduction study.
•	An UFl is not applied, as a BMCL was used as the POD.
•	An UFs is not applied because the principal study is a subchronic study.
The principal study was well conducted with respect to scope of examinations, numbers
of animals, and exposure levels but is given low-to-medium confidence because the duration was
short (5 weeks). Confidence in the database is low due to the lack of a 90-day inhalation study,
supporting data in a second species, a two-generation reproduction study, and developmental
toxicity studies. With regard to the latter, however, it is noteworthy that studies by oral and
dermal exposure did not identify reproductive and developmental endpoints as sensitive for this
chemical. Low confidence in the subchronic p-RfC follows.
Chronic p-RfC
A chronic p-RfC is derived by applying an UF of 3,000 to the BMCLiohec of
0.32 mg/m as follows:
Chronic p-RfC = BMCLiohec UF
= 0.32 mg/m3 - 3,000
= 0.0001 mg/m3 or 1 x 10"4 mg/m3
The composite UF of 3,000 is composed of the following factors:
•	An UF of 3 (10°5) is applied for interspecies extrapolation to account for potential
pharmacodynamic differences between rats and humans. Converting the rat data to
human equivalent concentrations by the dosimetric equations accounts for
pharmacokinetic differences between rats and humans; thus, an UF of 10 for
interspecies extrapolation was not used.
•	An UF of 10 for intraspecies differences is used to account for potentially susceptible
individuals in the absence of information on the variability of response in humans.
•	An UF of 10 is applied for using a study with a subchronic duration of exposure to
approximate chronic exposure.
19

-------
FINAL
9-28-2009
•	An UFl was not applied, as a BMCL was used as the POD.
•	An UF of 10 is applied to account for deficiencies in the database. The database
includes only one 5-week study; it lacks chronic toxicity studies, developmental
toxicity studies, and a multigeneration reproduction study.
Confidence in the chronic RfC is low, for the same reasons as described above for the subchronic
RfC, with the additional shortcoming that chronic studies are also lacking from the database.
PROVISIONAL CARCINOGENICITY ASSESSMENT FOR
DIETHYLENE GLYCOL MONOBUTYL ETHER (DGBE)
Weight-of-Evidence Descriptor
There is no adequate information on the carcinogenicity of DGBE due to a lack of oral,
inhalation, or dermal studies in animals longer than 13 weeks in duration (Johnson et al., 2005;
Bio/dynamics, Inc., 1989; Gushow et al., 1984; Hobson et al., 1987; Kodak, 1984). Genotoxicity
testing of DGBE in vitro with bacteria and mammalian cells and in vivo with Drosophila and
mice have yielded negative results in all but one study (a weakly positive mutagenic response in
mouse lymphoma cells in vitro) (Thompson et al., 1984; Zeiger et al., 1992;
Gollapudi et al., 1993). Under current U.S. EPA (2005) cancer guidelines, "there is inadequate
information to assess the carcinogenic potentiaF of DGBE.
Quantitative Estimates of Carcinogenic Risk
Derivation of quantitative estimates of cancer risk for DGBE is precluded by the lack of
carcinogenicity data.
REFERENCES
ACGIH (American Conference of Governmental Industrial Hygienists). 2001. Documentation
of the threshold limit values for chemical substances. 7th Edition. Cincinnati, OH.
ACGIH (American Conference of Governmental Industrial Hygienists). 2007. Threshold limit
values for chemical substances and physical agents and biological exposure indices. Cincinnati,
OH.
ATSDR (Agency for Toxic Substances and Disease Registry). 2008. Toxicological Profile
Information Sheet. U.S. Department of Health and Human Services, Public Health Service.
Online, http://www.atsdr.cdc.eov/toxprofiles/index.asp.
Auletta, C.S, R.E. Schroeder, W.J. Krasavage and C.R. Stack. 1993. Toxicology of diethylene
glycol butyl ether. 4. Dermal subchronic/reproduction study in rats. J. Am. Coll. Toxicol.
12:61-168.
Bio/dynamics, Inc. 1989. A 90-day dermal toxicity/fertility study in rats with diethylene glycol
butyl ether (DGBE) in rats. Submitted to U.S. EPA under TSCA Section 4, Fiche No.
OTS0521735.
20

-------
FINAL
9-28-2009
Bio-Research Labs. 1989. A 3-month study of potential effects of diethylene glycol butyl ether
on behavior and neuromorphology in rats. Submitted to U.S. EPA under TSCA Section 4, Fiche
No. OTS0521736.
CalEPA (California Environmental Protection Agency). 2002. Hot Spots Unit Risk and Cancer
Potency Values. Online, http://www.oehha.ca.gov/air/hot spots/pdf/TSDlookup2002.pdf.
CalEPA (California Environmental Protection Agency). 2005a. OEHHA/ARB Approved
Chronic Reference Exposure Levels and Target Organs. Online, http://www.arb.ca.gov/toxics/
healthy at/chroni c. pdf.
CalEPA (California Environmental Protection Agency). 2005b. Air Chronic Reference
Exposure Levels Adopted by OEHHA as of February 2005.
Ema, M., T. Itami and H. Kawasaki. 1988. Teratology study of diethylene glycol mono-n-butyl
ether in rats. Drug Chem. Toxicol. 11:97-111.
Gollapudi, B.B., V.A. Linscombe, M.L. McClintock et al. 1993. Toxicology of diethylene
glycol butyl ether. 3. Genotoxicity evaluation of an in vitro gene mutation assay and an in vivo
cytogenetic test. J. Am. Coll. Toxicol. 12(2): 155—159.
Gushow, T.S., R.R. Miller and B.L. Yano. 1984. Dowanol DB: A 5-week repeated vapor
inhalation study in rats. Dow Chemical USA, Midland, MI. Submitted to U.S. EPA under
TSCA Section 4, Fiche No. OTS0512379.
Hobson, D.W., J.F. Wyman, L.H. Lee et al.
monobutyl ether administered orally to rats.
TSCA Section 4, Fiche No. OTS05211652.
1987. The subchronic toxicity of diethylene glycol
U.S. Navy Report. Submitted to U.S. EPA under
IARC (International Agency for Research on Cancer). 2008. Search IARC Monographs.
Johnson, K.A., P.C. Baker, H.L. Kan et al. 2005. Diethylene glycol monobutyl ether (DGBE):
Two- and thirteen-week oral toxicity studies in Fischer 344 rats. Food Chem. Toxicol.
43:467-481.
Kodak. 1984. Toxicity Studies with Diethylene Glycol Monobutyl Ether. III. Six-Weeks
Repeated Dose Study. Eastman Kodak Co., Rochester, NY. Submitted to U.S. EPA under
TSCA Section 4, Fiche No. OTS0512376.
NIOSH (National Institute for Occupational Safety and Health). 2005. NIOSH Pocket Guide to
Chemical Hazards. Online, http://www.cdc.gov/niosh/npg/.
Nolen, G.A., W.B. Gibson, J.H. Benedict et al. 1985. Fertility and teratogenic studies of
diethylene glycol monobutyl ether in rats and rabbits. Fund. Appl. Toxicol. 5:1,137-1,143.
NTP (National Toxicology Program). 2005. 11th Report on Carcinogens. Online.
http://ntp.niehs.nih.gov/ntp/roc/tocl 1 .htm.
21

-------
FINAL
9-28-2009
NTP (National Toxicology Program). 2008. Management Status Report. Online.
http://ntp. niehs.nih.gov/index.cfm? obi ectid=78CC7E4C-F 1F6-975E-72940974DE301C3F.
OSHA (Occupational Safety and Health Administration). 2008. OSHA Standard 1910.1000
Table Z-l. Part Z, Toxic and Hazardous Substances. Online, http://www.osha.eov/pls/
oshaweb/owadisp.show document?p tabie=STANDARDS&p id=9992.
Proctor and Gamble Co. 1982. 28-Day subchronic percutaneous study of DGBE in rabbits.
Huntington Research Center. (Cited in SRC, 1992).
Schuler, R.L., B.D. Hardin, R.W. Niemeier et al. 1984. Results of testing fifteen glycol ethers in
a short-term in vivo reproductive toxicity assay. Environ. Health Perspect. 57:141-146.
Smyth, H.F. 1940. Special report on the toxicity of the glycols and their derivatives. 61.
Repeated doses of butyl "carbitol" by mouth. Mellon Institute of Industrial Research. Submitted
to U.S. EPA under TSCA Section 4, Fiche No. OTS0512492.
Smyth, H.F., Jr. and C.P. Carpenter. 1948. Further experience with the range finding test in the
industrial toxicology laboratory. J. Ind. Hyg. Toxicol. 30(l):63-68.
SRC (Syracuse Research Corporation). 1992. Health and Environmental Effects Document on
Glycol Ethers. Prepared by Syracuse Research Corporation, Syracuse, NY for the Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office,
Cincinnati, OH. SRC TR-92-007.
Thompson, E.D., W.J. Coppinger, R. Valencia et al. 1984. Mutagenicity testing of diethylene
glycol monobutyl ether. Environ. Health Perspect. 57:105-112.
U.S. EPA. 1984. Health Effects Assessment for Glycol Ethers. Prepared by the Office of
Health and Environmental Assessment, Environmental Criteria and Assessment Office,
Cincinnati, OH, for the Office of Emergency and Remedial Response, Washington, D.C.
EPA/540/1-86-052.
U.S. EPA. 1991. Chemical Assessments and Related Activities (CARA). Office of Health and
Environmental Assessment, Washington, DC. April.
U.S. EPA. 1994a. Chemical Assessments and Related Activities (CARA). Office of Health and
Environmental Assessment, Washington, DC. December.
U.S. EPA. 1994b. Methods of Derivation of Inhalation Reference Concentrations and
Application of Inhalation Dosimetry. Office of Research and Development, Washington, DC.
EPA/600/8-90/066F. October.
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,
Cincinnati OH for the Office of Emergency and Remedial Response, Washington, DC. July.
EPA/540/R-97/036. NTIS PB97-921199.
22

-------
FINAL
9-28-2009
U.S. EPA. 2000. Benchmark Dose Technical Guidance Document [external review draft],
EPA/63 0/R-00/001.
U.S. EPA. 2005. Guidelines for Carcinogen Risk Assessment. U.S. Environmental Protection
Agency, Risk Assessment Forum, Washington, DC. EPA/630/P-03/001B. Online.
http://www.thecre.com/pdf/200504Q4 cancer.pdf.
U.S. EPA. 2006. 2006 Edition of the Drinking Water Standards and Health Advisories. Office
of Water, Washington, DC. EPA 822-R-06-013. Washington, DC. Online.
http://www.epa.eov/waterscience/drinkine/standards/dwstandards.pdf.
U.S. EPA. 2008. Integrated Risk Information System (IRIS). Online. Office of Research and
Development, National Center for Environmental Assessment, Washington, DC.
http ://www. epa. gov/iris/.
WHO (World Health Organization). 2008. Online catalogs for the Environmental Health
Criteria Series. Online, http://www.who.int/ipcs/publications/ehc/ehc alphabetical/en/
index.html.
Zeiger, E., B. Anderson, S. Haworth et al. 1992. Salmonella mutagenicity tests: V. Results
from the testing of 311 chemicals. Environ. Mol. Mutagen. 19:2-141.
23

-------
FINAL
9-28-2009
APPENDIX A. DETAILS OF BENCHMARK DOSE MODELING
FOR SUBCHRONIC/CHRONIC p-RfD
Model Fitting Procedure for Continuous Data
The model fitting procedure for continuous data is as follows. The simplest model
(linear) is first applied to the data while assuming constant variance. If the data are consistent
with the assumption of constant variance (p> 0.1), then the fit of the linear model to the means is
evaluated, and the polynomial, power, and Hill models are 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 Akaike Information Criterion (AIC) for the fitted model is selected for BMD
derivation. If the test for constant variance is negative, the linear model is run again while
modeling the variance as a power function of the mean to account for nonhomogenous variance.
If the nonhomogenous variance model provides an adequate fit (p> 0.1) to the variance data,
then the fit of the linear model to the means is evaluated, and the polynomial, power, and Hill
models are fit to the data and evaluated while the variance model is applied. Among those
providing adequate fit to the means (p> 0.1), the one with the lowest AIC for the fitted model is
selected for BMD derivation. If the test for constant variance is negative and the
nonhomogenous variance model does not provide an adequate fit to the variance data, then the
data set is considered unsuitable for modeling. If after these attempts, no model provides an
adequate fit to the data, the highest dose is dropped, if appropriate, and the entire procedure is
repeated. If no fit is obtained after dropping the highest dose, the next highest dose is dropped, if
appropriate, and the procedure is repeated. Dose-dropping continues until (1) adequate fit is
obtained; (2) there are only controls and two dose groups remaining; or (3) it is inappropriate to
continue dropping doses due to a lack of statistical significance or biologically important
differences between controls and the remaining treatment groups. If no fit is obtained following
application of this procedure, then the data set is not considered to be amenable to BMD
modeling.
24

-------
FINAL
9-28-2009
Model Fitting Results for Decreased RBCs in Male Rats (Johnson et al., 2005)
Following the above procedure, continuous-variable models in the EPA Benchmark Dose
Software (BMDS) (version 2.1) were fit to the data shown in Table 3 (main text) for decreased
RBCs in male rats. In the absence of a biologically relevant response level, the benchmark
response (BMR) was chosen to be 1 standard deviation (SD) from the control mean, as
recommended by U.S. EPA (2000). As shown in Table A-l, all models provide adequate fit to
the data. The Hill model (Figure A-l) estimated a significantly lower BMDL than other models;
therefore, the BMDL for this endpoint is 81 mg/kg-day.
Table A-l. Model Predictions for Decreased RBCs in Male Rats3
Data Set/Model
Variance
7?-Valueb
Means
p-\alueb
AIC
bmd1sd
(mg/kg-day)
BMDL1sd
(mg/kg-day)
Linear (constant variance)0
0.4943
0.454
-50.5176
440.732
328.449
2-Degree Polynomial (constant variance)0
0.4943
0.454
-50.5176
440.732
328.449
3-Degree Polynomial (constant variance)0
0.4943
0.454
-50.5176
440.732
328.449
Power (constant variance)d
0.4943
0.454
-50.5176
440.732
328.449
Hill (constant variance/
0.4943
0.6312
-49.8664
222.31
81.408
aJohnson et al. (2005)
bValues <0.10 fail to meet conventional goodness-of-fit criteria
Coefficients restricted to be negative
dPower restricted to >1
25

-------
FINAL
9-28-2009
Hill Model with 0.95 Confidence Level
9.6
Hill
9.4
9.2
9
8.8
8.6
8.4
BMDL
BMD
8.2
0
200
400
600
800
1000
Dose
16:27 07/10 2008
Figure A-l. Fit of Hill Model to Data on Decreased RBCs
in Male Rats (Johnson et al., 2005)
Hill Model. (Version: 2.14; Date: 06/26/2008)
Input Data File: C:\USEPA\BMDS2l\Data\hilRBCmaleHill.(d)
Gnuplot Plotting File: C:\USEPA\BMDS21\Data\hilRBCmaleHill.plt
Fri Aug 28 16:22:41 2009
BMDS Model Run
The form of the response function is:
Y[dose] = intercept + v*dose^n/(k^n + dose^n)
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
Power parameter restricted to be greater than 1
A constant variance model is fit
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
26

-------
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
FINAL
9-28-2009
Default Initial Parameter Values
alpha =
rho =
intercept =
v =
n =
k =
0.09565
0
9.27
-0.74
0.812681
323.171
Specified
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho -n
have been estimated at a boundary point, or have been specified by
the user,
alpha
intercept
and do not appear in the correlation matrix )
alpha
1
1.8e-009
2.2e-008
-9.9e-009
intercept
1. 8e-009
1
0.33
-0.57
v
2 . 2e-008
0.33
1
-0. 94
-9.9e-009
-0.57
-0.94
1
Interval
Variable
Limit
alpha
0.124528
intercept
9.40457
v
0.0729792
n
k
2226.33
Estimate
0.0865823
9.24677
-1.20681
1
689.459
Parameter Estimates
Std. Err.
0.0193604
0. 080516
0.578497
NA
784.134
NA - Indicates that this parameter has hit a bound
implied by some ineguality constraint and thus
has no standard error.
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
0.0486366
9.08896
-2.34065
-847.415
Table of Data and Estimated Values of Interest
Dose
Obs Mean
Est Mean Obs Std Dev Est Std Dev Scaled Res.
0
50
250
1000
10
10
10
10
9.27
9.13
8.94
8.53
9.25
9.17
8.93
8.53
0.35
0.22
0.34
0.31
0.294
0.294
0.294
0.294
0.25
-0.378
0.155
-0.0263
27

-------
FINAL
9-28-2009
Model Descriptions for likelihoods calculated
Model A1:	Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2 :	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi = Mu + e(i)
Var{e(i)} = Sigma^2
Likelihoods of Interest
Model	Log(likelihood)	# Param's	AIC
A1	29.048402	5	-48.096804
A2	30.246635	8	-44.493271
A3	29.048402	5	-48.096804
fitted	28.933200	4	-49.866400
R	16.206159	2	-28.412317
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (A1 vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Tests of Interest
Test
-2*log(Likelihood Ratio) Test df
p-value
Test
Test
Test
Test
28.081
2.39647
2.39647
0.230404
<.0001
0.4943
0.4943
0.6312
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .1. A homogeneous variance
model appears to be appropriate here
The p-value for Test 3 is greater than .1. The modeled variance appears
to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems
to adeguately describe the data
28

-------
FINAL
9-28-2009
Benchmark Dose Computation
Specified effect =	1
Risk Type	=	Estimated standard deviations from the control mean
Confidence level =	0.95
BMD =	222.31
BMDL =	81.4084
29

-------
FINAL
9-28-2009
Model Fitting Results for Decreased RBCs in Female Rats (Johnson et al., 2005)
Following the above procedure, continuous-variable models in the EPA BMDS
(version 2.1) were fit to the data shown in Table 2 (main text) for decreased RBCs in female rats.
In the absence of a biologically relevant response level, the BMR was chosen to be 1 SD from
the control mean, as recommended by U.S. EPA (2000). As shown in Table A-2, the linear and
power models (linear model shown in Figure A-2) provide identical fit to the data and yielded
AIC values that were significantly lower than the other models fitted. Thus, the estimated
BMDL for this endpoint is 280 mg/kg-day.
Table A-2. Model Predictions for Decreased RBCs in Female Rats"
Data Set/Model
Variance
7?-Valueb
Means
p-\alueb
AIC
bmd1sd
(mg/kg-day)
BMDL1sd
(mg/kg-day)
Linear (constant variance)0
0.3052
0.2239
-69.286
364.222
279.607
2-Degree Polynomial (constant variance)0
0.3052
0.0862
-67.3357
408.969
280.868
3-Degree Polynomial (constant variance)0
0.3052
0.08802
-67.3694
429.854
281.283
Power (constant variance)d
0.3052
0.2239
-69.286
364.876
280.266
Hill (constant variance)d
0.3052
0.08352
-67.2847
364.222
279.604
aJohnson et al. (2005)
bValues <0.10 fail to meet conventional goodness-of-fit criteria
Coefficients restricted to be negative
dPower restricted to >1
30

-------
FINAL
9-28-2009
Linear Model with 0.95 Confidence Level
Linear
8.4
8.2

c
o
CL

d)
q:
7.8
7.6
7.4
BMDL
BMD
0
200
400
600
800
1000
Dose
16:32 07/10 2008
Figure A-2. Fit of Linear Model to Data on Decreased RBCs in Female Rats
(Johnson et al., 2005)
BMDs and BMDLs indicated are associated with a change of 1 SD from the control and are in units mg/kg-day
Hill Model. (Version: 2.14; Date: 06/26/2008)
Input Data File: C:\USEPA\BMDS2l\Data\hilRBCmaleHill.(d)
Gnuplot Plotting File: C:\USEPA\BMDS21\Data\hilRBCmaleHill.plt
Fri Aug 28 16:57:54 2009
BMDS Model Run
The form of the response function is:
Y[dose] = intercept + v*dose^n/(k^n + dose^n)
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
Power parameter restricted to be greater than 1
A constant variance model is fit
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
31

-------
FINAL
9-28-2009
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha =	0.05775
rho =
intercept =
v =
n =
k =
0 Specified
8.26
-0.72
0.970124
490.566
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho -n
have been estimated at a boundary point, or have been specified by
the user,
alpha
intercept
and do not appear in the correlation matrix )
alpha
1
0.00028
0.0044
-0.0044
intercept
0.00028
1
0. 061
-0.065
v
0.0044
0.061
1
-1
-0.0044
-0.065
-1
1
Interval
Variable
Limit
alpha
0.0805658
intercept
8 .28808
v
11523.5
n
k
1.	85051e+007
Estimate
0.056016
8.19343
-240.216
1
369305
Parameter Estimates
Std. Err.
0.0125257
0.048292
6001.98
NA
9.25312e+006
NA - Indicates that this parameter has hit a bound
implied by some inequality constraint and thus
has no standard error.
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
0.0314661
8.09878
-12003.9
-1.77 665e+007
Table of Data and Estimated Values of Interest
Dose
Obs Mean
Est Mean Obs Std Dev Est Std Dev Scaled Res.
0
50
250
1000
10
10
10
10
8.26
8.06
8.07
7.54
8.19
8.16
8.03
7.54
0.22
0.31
0.24
0.17
0.237
0.237
0.237
0.237
0.889
-1.35
0.522
-0.0632
32

-------
FINAL
9-28-2009
Model Descriptions for likelihoods calculated
Model A1:	Yij = Mu(i) + e(ij)
Var{e(ij)} = SigmaA2
Model A2 :	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi = Mu + e(i)
Var{e(i)} = Sigma^2
Likelihoods of Interest
Model	Log(likelihood)	# Param's	AIC
A1	39.139849	5	-68.279698
A2	40.951049	8	-65.902099
A3	39.139849	5	-68.279698
fitted	37.642370	4	-67.284740
R	21.815106	2	-39.630212
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (A1 vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Tests of Interest
Test
-2*log(Likelihood Ratio) Test df
p-value
Test
Test
Test
Test
38 .2719
3.6224
3.6224
2.99496
<.0001
0.3052
0.3052
0.08352
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .1. A homogeneous variance
model appears to be appropriate here
The p-value for Test 3 is greater than .1. The modeled variance appears
to be appropriate here
The p-value for Test 4 is less than .1. You may want to try a different
model
Benchmark Dose Computation
33

-------
Specified effect
Risk Type
Confidence level
BMD
BMDL
FINAL
9-28-2009
1
Estimated standard deviations from the control mean
0. 95
364.222
279.607
34

-------
FINAL
9-28-2009
Model Fitting Results for Decreased Hemoglobin Concentration in Male Rats
(Johnson et al., 2005)
Following the above procedure, continuous-variable models in the EPA BMDS
(version 1.4.1c) were fit to the data shown in Table 3 (main text) for decreased hemoglobin
concentration in male rats. In the absence of a biologically relevant response level, the BMR
was chosen to be 1 SD from the control mean, as recommended by U.S. EPA (2000). As shown
in Table A-3, the linear, polynomial, and power models provide identical linear fit to the data.
Fit to the Hill Model cannot be tested due to insufficient degrees of freedom. Fit to the linear
model is illustrated in Figure A-3. For this endpoint, the estimated BMDL is 328 mg/kg-day.
Table A-3. Model Predictions for Decreased Hemoglobin in Male Ratsa
Data Set/Model
Variance
p-\alueb
Means
p-\alueb
AIC
bmd1sd
(mg/kg-day)
BMDLisd
(mg/kg-day)
Linear (constant variance)0
0.8494
0.2992
-23.84147
439.943
328.015
2-Degree Polynomial (constant variance)0
0.8494
0.2992
-23.84147
439.943
328.015
3-Degree Polynomial (constant variance)0
0.8494
0.2992
-23.84147
439.943
328.015
Power (constant variance)d
0.8494
0.2992
-23.84147
439.943
328.015
Hill (constant variance)d
Fit cannot be tested for this model due to inadequate degrees of
freedom
aJohnson et al. (2005)
bValues <0.10 fail to meet conventional goodness-of-fit criteria
Coefficients restricted to be negative
dPower restricted to >1
35

-------
FINAL
9-28-2009
Linear Model with 0.95 Confidence Level
16.2
16
15.8
S 15.6
C
o
§" 15.4
a.
S 15.2
2
15
14.8
14.6
14.4
0	200	400	600	800	1000
Dose
16:41 07/10 2008
Figure A-3. Fit of Linear Model to Data on Decreased
Hemoglobin Concentration in Male Rats (Johnson et al., 2005)
BMDs and BMDLs indicated are associated with a change of 1 SD from the control and are in units mg/kg-day
Polynomial Model. (Version: 2.13; Date: 04/08/2008)
Input Data File: C:\USEPA\BMDS2l\Data\linRBCmaleLinear.(d)
Gnuplot Plotting File: C:\USEPA\BMDS21\Data\linRBCmaleLinear.plt
Fri Aug 28 17:11:28 2009
BMDS Model Run
The form of the response function is:
Y[dose] = beta_0 + beta_l*dose + beta_2*dose/s2 + ...
Dependent variable = Mean
Independent variable = Dose
rho is set to 0
The polynomial coefficients are restricted to be negative
A constant variance model is fit
Linear
BMDL
BMD
36

-------
FINAL
9-28-2009
Total number of dose groups = 4
Total number of records with missing values = 0
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
alpha =	0.1825
rho =	0 Specified
beta_0 =	15.7086
beta 1 = -0.000949416
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -rho
have been estimated at a boundary point, or have been specified by
the user,
alpha
beta_0
beta 1
and do not appear in the correlation matrix )
alpha
1
-1.3e-009
1.7e-009
beta_0
-1.3e-009
1
-0.63
beta_l
1.7e-009
-0. 63
1
Interval
Variable
Limit
alpha
0.250925
beta_0
15.8752
beta_l
0.000626445
Estimate
0.174464
15 .7086
-0.000949416
Parameter Estimates
Std. Err.
0.0390113
0.0850278
0.000164784
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
0.0980032
15.5419
-0.00127239
Table of Data and Estimated Values of Interest
Dose
Obs Mean
Est Mean Obs Std Dev Est Std Dev Scaled Res.
0
50
250
1000
10
10
10
10
15.8
15.7
15.3
14.8
15 .7
15 .7
15 .5
14.8
0.5
0.4
0.4
0.4
0.418
0.418
0.418
0.418
0. 692
0.295
-1.3
0.309
Model Descriptions for likelihoods calculated
Model A1:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
37

-------
FINAL
9-28-2009
Model A2 :	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma(i)^2
Model A3:	Yij = Mu(i) + e(ij)
Var{e(ij)} = Sigma^2
Model A3 uses any fixed variance parameters that
were specified by the user
Model R:	Yi = Mu + e(i)
Var{e(i)} = Sigma^2
Likelihoods of Interest
Model	Log(likelihood)	# Param's	AIC
A1	16.127312	5	-22.254625
A2	16.527404	8	-17.054808
A3	16.127312	5	-22.254625
fitted	14.920736	3	-23.841473
R	2.835616	2	-1.671231
Explanation of Tests
Test 1: Do responses and/or variances differ among Dose levels?
(A2 vs. R)
Test 2: Are Variances Homogeneous? (A1 vs A2)
Test 3: Are variances adeguately modeled? (A2 vs. A3)
Test 4: Does the Model for the Mean Fit? (A3 vs. fitted)
(Note: When rho=0 the results of Test 3 and Test 2 will be the same.)
Tests of Interest
Test
-2*log(Likelihood Ratio) Test df
p-value
Test
Test
Test
Test
27.3836
0.800183
0.800183
2.41315
0.0001227
0.8494
0.8494
0.2992
The p-value for Test 1 is less than .05. There appears to be a
difference between response and/or variances among the dose levels
It seems appropriate to model the data
The p-value for Test 2 is greater than .1. A homogeneous variance
model appears to be appropriate here
The p-value for Test 3 is greater than .1. The modeled variance appears
to be appropriate here
The p-value for Test 4 is greater than .1. The model chosen seems
to adeguately describe the data
Benchmark Dose Computation
Specified effect =	1
Risk Type	=	Estimated standard deviations from the control mean
Confidence level =	0.95
38

-------
BMD =	439.943
BMDL =	328.015
FINAL
9-28-2009
39

-------
FINAL
9-28-2009
Model Fitting Results for Decreased Hemoglobin Concentration in Female Rats
(Johnson et al., 2005)
Following the above procedure, continuous-variable models in the EPA BMDS
(version 1.4.1c) were fit to the data shown in Table 3 (main text) for decreased hemoglobin
concentration in female rats. In the absence of a biologically relevant response level, the BMR
was chosen to be 1 SD from the control mean, as recommended by U.S. EPA (2000). As shown
in Table A-4, none of the models provides an adequate fit to the data, even with the highest dose
dropped. Dropping the highest dose reduces the data set to controls and two dose groups;
therefore, it is not possible to drop further dose groups to attempt to identify a fit. Therefore, this
data set is not amenable to BMD modeling.
Table A-4. Model Predictions for Decreased Hemoglobin in Female Ratsa
Data Set/Model
Variance
/7-Valueb
Means
p-\alueb
AIC
bmd1sd
(mg/kg-day)
BMDL1sd
(mg/kg-day)
All Doses
Linear (constant variance)0
0.652
0.05345
-35.534
579.683
408.287
2-Degree Polynomial (constant variance)0
0.652
0.05345
-35.534
579.683
408.287
3-Degree Polynomial (constant variance)0
0.652
0.05345
-35.534
579.683
408.287
Power (constant variance)d
0.652
0.05345
-35.534
579.683
408.287
Hill (constant variance)d
0.652
0.03974
-35.163
127.36
11.921
Highest Dose Dropped
Linear (constant variance)0
0.5991
0.03223
-22.281
331.251
170.263
2-Degree Polynomial (constant variance)0
0.5991
0.03223
-22.281
331.251
170.263
Power (constant variance)d
0.5991
0.03223
-22.281
331.251
170.263
Hill (constant variance)d
# observations < # parameters: cannot run this model
aJohnson et al. (2005)
bValues <0.10 fail to meet conventional goodness-of-fit criteria
Coefficients restricted to be negative
dPower restricted to >1
40

-------
FINAL
9-28-2009
Modeling Procedure for Hepatocyte Vacuolization in Females (Gushow et al., 1984)
The BMD modeling for dichotomous data was conducted with the EPA's BMD software
(BMDS version 2.1). For all the dichotomous data, the incidence data on hepatocyte
vacuolization with duration-adjusted concentrations (see Table 4 of main text) were modeled
with eight dichotomous models (i.e., Gamma, Multistage, Logistic, Log-logistic, Probit,
Log-probit, Weibull, and Quantal-linear models) with a BMR of 10% extra risk. An adequate
model fit was judged based on the goodness-of-fit p-walue (p> 0.1), scaled residual at the range
of BMR, and visual inspection of the model fit. As shown in Table A-5 and Figures A-4 and
A-5, all the models provided adequate goodness-of-fit ^-values (>0.1), and estimated BMCLs
from these models are sufficiently close (ranging from 0.32 to 0.60 mg/m3). Therefore, the
BMCL of 0.32 mg/m3 estimated from the Multistage model (Figure A-5), with the lowest AIC
from these five models, was considered to be an appropriate estimate for this endpoint.
Table A-5. Model Predictions for Increased Hepatocyte Vacuolization in Female Ratsa
Data Set/Model
Goodness-of-Fit />-Valueb
AIC
Scaled
Residual
BMC10
(mg/m3)
BMCL10
(mg/m3)
Gamma0
0.9917
38.18
0.001
2.0
0.32
Logistic
0.8180
36.58
0.367
0.9
0.52
Log-logisticd
0.8523
38.24
0.053
2.1
0.50
Multistage6
0.9889
36.20
-0.119
1.7
0.32
Probit
0.8330
36.54
0.338
0.8
0.55
Log-probitd
0.9398
38.19
0.017
2.1
0.60
Weibullc
0.9999
38.18
0.000
1.9
0.32
Quantal-linear
0.5132
37.60
0.360
0.48
0.27
aGushow et al. (1984)
bValues <0.10 fail to meet conventional goodness-of-fit criteria
°Power restricted to >1
dSlope restricted to >1
"Betas restricted to >0
41

-------
FINAL
9-28-2009
Quantal Linear Model with 0.95 Confidence Level
Quantal Linear
0.8
0.6
0.4
0.2
EMDL
BMD
0
5
10
15
20
Dose
13:56 07/28 2009
Figure A-4. Fit of Quantal-Linear Model to Data
on Increased Hepatocyte Vacuolization in Female Rats (Gushow et al., 1984)
BMCs and BMCLs indicated are associated with a change of 10% extra risk from the control, and are in units mg/m3
42

-------
FINAL
9-28-2009
Multistage Model with 0.95 Confidence Level
0.8
T3
aj
|	0.6
<
C
o
'¦4—'
o
03
0.4
0.2
14:53 07/28
Figure A-5. Fit of Multistage Model to Data
on Increased Hepatocyte Vacuolization in Female Rats (Gushow et al., 1984)
BMCs and BMCLs indicated are associated with a change of 10% extra risk from the control, and are in units mg/m3
Multistage Model. (Version: 3.0; Date: 05/16/2008)
Input Data File: C:\USEPA\BMDS2l\Data\mstHepaticmulti.(d)
Gnuplot Plotting File: C:\USEPA\BMDS21\Data\mstHepaticmulti.plt
Fri Aug 28 17:38:06 2009
BMDS Model Run
The form of the probability function is:
P[response] = background + (1-background)*[1-EXP(
-betal*dose/sl-beta2*dose/s2) ]
The parameter betas are restricted to be positive
Dependent variable = Response
Independent variable = Dose
Multistage
2009
0	5	10	15	20
Dose
43

-------
FINAL
9-28-2009
Total number of observations = 4
Total number of records with missing values = 0
Total number of parameters in model = 3
Total number of specified parameters = 0
Degree of polynomial = 2
Maximum number of iterations = 250
Relative Function Convergence has been set to: le-008
Parameter Convergence has been set to: le-008
Default Initial Parameter Values
Background =	0
Beta(l) =	0
Beta(2) = 2.27362e+017
Asymptotic Correlation Matrix of Parameter Estimates
( *** The model parameter(s) -Beta(l)
have been estimated at a boundary point, or have been specified by
the user,
and do not appear in the correlation matrix )
Background	Beta (2)
Background	1	-0.41
Beta (2)	-0.41	1
Parameter Estimates
Interval
Variable
Limit
Background
Beta(1)
Beta(2)
Estimate
0.288586
0
0.0381152
Std. Err.
95.0% Wald Confidence
Lower Conf. Limit Upper Conf.
Indicates that this value is not calculated.
Model
Full model
Fitted model
Reduced model
Analysis of Deviance Table
Log(likelihood)
-16.0896
-16.1007
-25.8979
# Param's
4
2
1
Deviance Test d.f.
0.0222818
19.6166
P-value
0.9889
0.0002038
AIC:
36.2015
Dose
Est. Prob.
Goodness of Fit
Expected Observed	Size
Scaled
Residual
0.0000
2.3000
0.2886
0.4185
2.886
4.185
3.000
4.000
10
10
0. 080
-0.119
44

-------
FINAL
9-28-2009
7.1000	0.8958	8.958	9.000	10	0.043
21.3000	1.0000	10.000 10.000	10	0.000
Chi^2 = 0.02	d.f. = 2	P-value = 0.9889
Benchmark Dose Computation
Specified effect =	0.1
Risk Type	=	Extra risk
Confidence level =	0.95
BMD =	1.662 61
BMDL =	0.318188
45

-------