FUNDAMENTAL AND APPLIED TOXICOLOGY 3 9 , 1 2 0 - 130 (1997)
ARTICLE NO FA972363
Physiologically Based Pharmacokinetics and the Dermal Absorption
of 2-Butoxyethanol Vapor by Humans1
R. A. Corley,* D. A. Markham.t C. Banks.t P. Delorme,$ A. Masterman,§ and J. M. Houle§
*Molecular Biosciences Department, Battelle Memorial Institute, Pacific Northwest Division, PO Box 999, MS P7-59, Richland, Washington 99352;
\The Toxicology Research Laboratory, Health and Environmental Research Laboratory, The Dow Chemical Company,
1803 Building, Midland, Michigan 48674; %Bw-Research Laboratories, Ltd., 87 Senneville Road, Senneville,
Quebec Canada H9X 3R3; and §LA.B. Pharmacological Research International, Inc.,
1000 St. Charles Avenue, Vaudreuil, Quebec Canada J7V 8P5
Received February 3, 1997; accepted August 12, 1997
gated (~67%) with glutamine, confirming recent reports. These
results, coupled with PBPK modeling of worst-case exposure scenarios (no clothing, 100% of the body was exposed), demonstrated
that no more than 15-27% (low-to-high relative temperatures and
humidities), not 75%, of the total uptake of BE could be attributed
It has generally been assumed that the skin contributes only to the skin of humans during simulated 8-hr exposures to the
minor amounts to the total uptake of solvent vapors, relative to ACGIH TLV concentration of 25 ppm. Even less of the total
the respiratory tract. Contrary to this assumption, the widely used uptake was attributed to the skin during simulations of exercise
glycol ether solvent, 2-butoxyethanol (BE), has been reported to with whole-body exposures (5-9%) or by more realistic exposures
be more effectively absorbed through the skin (75% of the total of only the arms and head (1-8%). As a result, humans are unuptake) than through the lungs of humans (Johanson and Boman, likely to reach hemolytic concentrations of the metabolite BAA in
1991, Br. J. Ind. Med. 48,788). The possibility that thefingerprick blood following vapor exposures to BE. o iw Sod«y of To
blood sampling technique used in the Johanson and Boman study
was confounded by locally high concentrations of BE at the site
of absorption was suggested using a previously developed PBPK
A common practice in physiologically based pharmacokimodel (Corley et al., 1994, Toxicol. Appl. Pharmacol 129,61). The netic (PBPK) modeling of vapor exposures is to assume
current study was conducted to verify the PBPK analysis and to that the skin does not contribute significantly to the total
determine whether or not the skin was the major site for absorption
uptake of solvents compared to the respiratory tract. This
of BE vapor by exposing one arm from each of six human volun13
practice
has been based largely on the relative differences
teers to 50 ppm C2-BE vapor for 2 hr. To evaluate the potential
in
surface
area, barrier thickness, and blood perfusion rates
consequences of blood sampling techniques, samples were taken
from both the unexposed arm (catheter; during and after exposure) which favor the absorption of solvent vapors via the lungs
and the exposed arm (finger prick; end of the exposure only) for over that of the skin. The limited number of available studies
analysis of both BE and its major metabolite, butoxyacetic acid with vapors have generally supported this assumption (Rii(BAA). Butoxyacetic acid is responsible for the hemolysis observed himaki and Pfaffli, 1978; McDougal et al., 1986, 1990;
in toxicity studies with laboratory animals. Humans, however, are Jacobs and Phanprasit, 1993). As a result, most PBPK modsignificantly less sensitive to this effect The concentration of BE els developed for volatile solvents have ignored this potenin the finger prick blood samples averaged 1500 times higher than
tial route of exposure.
the corresponding concentration in venous blood sampled from a
Recently, however, a human pharmacokinetic study was
catheter installed in the unexposed arm at the end of the exposure.
published
which contradicted this practice by concluding
Blood BAA levels were generally within a factor of 4 of each other
for the two techniques and, therefore, was considered a better that dermal absorption of vapor was far more significant
indicator of systemic absorption. Urine was collected for 24 hr and than respiratory uptake in humans exposed to 2-butoxyethaanalyzed for the following metabolites found in rat metabolism nol (BE), an ethylene glycol ether used in a variety of hard
studies: free and conjugated BE, BAA, ethylene glycol (EG), and surface cleaners (Johanson and Boman, 1991). In their study,
glycol ic acid (GA), with only BAA detected in the human urine. human volunteers were exposed mouth-only for 2 hr to 50
More importantly, urinary BAA was found to be extensively conju- ppm BE, followed by 1 hr of no exposure, followed by 2
hr of skin-only exposure to 50 ppm BE (i.e., volunteers were
exposed in a whole-body chamber while breathing fresh air
1
This work was presented in part at the 34th annual meeting of the
through a face-mask). Using the areas under the curve for
Society of Toxicology, March 5 - 9 , 1995, Baltimore, MD. This work was
the concentration of BE in finger prick blood samples for
supported by the Ethylene Glycol Ethers Panel of the Chemical Manufacturers Association, Washington, DC.
the two exposures and the known clearances of BE for each
Physiologically Based Pharmacokinetics and the Dermal Absorption of 2-Butoxyethanol Vapor by Humans. Corley, R. A.,
Markham, D. A., Banks, C , Delorme, P., Masterman, A., and
Houle, J. M. (1997). Fundam. Appl. Toxicol. 39, 120-130.
0272-0590/97 $25.00
Copyright C 1997 by the Society of Toxicology.
All rights of reproduction in any form reserved.
120
HUMAN DERMAL UPTAKE OF 2-BUTOXYETHANOL VAPOR
of the volunteers determined in a previous kinetic study
(Johanson et al., 1986), Johanson and Boman calculated that
approximately 75% of the total uptake of BE vapor was
attributable to the skin.
Corley et al. (1994) discussed the possibility that the finger prick blood sampling technique used by Johanson and
Boman was confounded by the local absorption of BE and
did not reflect the concentration of BE in "systemic blood,"
a necessary assumption for their classical, noncompartmental, pharmacokinetic analysis. By expanding their PBPK
model for BE to include the dermal uptake of vapor, the
data of Johanson and Boman were reanalyzed to explain
how such an unusual conclusion regarding dermal absorption
was reached. Simulations of their data were conducted assuming that the finger prick blood samples represented either
(a) systemic arterial blood or (b) venous blood draining the
skin compartment prior to systemic dilution, to estimate the
skin permeability coefficient for BE under each condition.
Additional simulations were conducted for another study
from the same laboratory on the kinetics of BE following
whole-body exposure to 20 ppm under exercise conditions
(Johanson et al., 1986). In each of these simulations, the
assumption that the finger prick blood samples represented
venous blood draining the skin compartment, and not systemic blood, provided the most consistent explanation of the
available data.
While the question of dermal vapor absorption was particularly important for 2-butoxyethanol, there existed the potentially larger issue of incorrectly assuming that dermal uptake
of certain solvent vapor is negligible compared to uptake in
the lungs. Thus, a human dermal vapor kinetic study was
conducted with BE to verify the previous simulations showing that the respiratory tract is far more important than the
skin for the uptake of BE. The PBPK model was used to
design the present study with particular attention to blood
sampling methods and time of collection. Blood and urine
samples were analyzed for the major metabolite, butoxyacetic acid (BAA), and its extent of conjugation by humans.
BAA is primarily responsible for the hemolysis of red blood
cells observed in laboratory animals (Ghanayem et al., 1987;
Bartnik et al., 1987; Ghanayem, 1989) although humans are
significantly less susceptible to this effect (Ghanayem, 1989;
Udden, 1993; Ghanayem and Sullivan, 1993). In addition,
analyses for metabolites identified in studies with laboratory
animals were included to improve the existing PBPK model.
METHODS
Experimental design. Six human volunteers were exposed, arm-only,
to 50 ppm I3C2-BE for 2 hr. Presumably due to the widespread use of
cleaners containing BE, variable background levels of BE metabolites (up to
20-50 ppb) were detected in control human urine samples during methods
development. As a result, the stable isotope, "Ci-BE, was used as the source
of the vapor exposure to unequivocally relate the concentrations of BE and
its metabolites to the controlled exposure and to achieve the very low
detection limits (low ppb) in blood and urine that were necessary to conduct
121
exposures of only one arm. Catheters were inserted into the antecubital
vein of the opposite, unexposed arm as the primary source of systemic
blood collection for the analysis of both BE and its major metabolite, BAA.
Finger prick blood samples were taken from the exposed arm only at the
end of the 2-hr exposure for analysis of both BE and BAA to compare with
die blood samples obtained from the catheter. Urine samples were collected
for metabolite analyses and to confirm a recent report that BAA is extensively conjugated with glutamine in humans (Rettenmeier et al., 1993).
Chemicals. The test atmospheres were generated using "C 2 -butoxyethanol, labeled in the ethylene glycol backbone (CDN Isotopes, Vaudreuil,
Quebec, Canada) with a purity of 99.54% as determined by capillary gas
chromatography with thermal conductivity detection (GC-TCD). The metabolites, l3C2-butoxyacetic acid with a purity of 94.2% (by GC-TCD) and
13
C2-ethylene glycol with a purity of 98.3% were synthesized by CDN
Isotopes and The Dow Chemical Co., respectively. The internal standards,
synthesized by The Dow Chemical Co., for quantitating BE, BAA and EG
in blood and urine samples consisted of the deuterated analogues d7-BE,
d7-BAA, and d4-EG, respectively. The identity of each of these compounds
was confirmed by capillary gas chromatography electron impact mass spectrometry (GC/EI/MS). Authentic N-butoxyacetylglutamine (NBGL) was obtained from Dr. Albert Rettenmeier (University of Tubingen, Germany).
Subjects. Six healthy male volunteers, 20 to 45 years of age, participated
in this study. Each subject received a complete physical examination and
laboratory tests of hematopoietic, hepatic, and renal function prior to participating in the study. Medical histories and demographic data including sex,
age, race, body weight, height, build, and smoking habits were recorded
(Table 1). In addition, subjects were asked whether they had followed a
normal diet in the previous 4 weeks, if they were taking medication regularly
for a chronic condition, and whether they had donated blood in the previous
6 months. Subjects were excluded from a larger pool of volunteers if they had
a history of significant cardiovascular, hepatic, CNS, renal, hematological, or
gastrointestinal disease; skin problems, particularly on the arm to be used
for exposure; alcoholism or drug abuse during the preceding year; abnormal
diet during the 4 weeks preceding the study; personality disorders that would
have precluded valid informed consent or compliance with protocol requirements; and if, through the completion of this study, they would have donated
in excess of 900 ml of blood over the previous 20 weeks. Subjects were
prohibited from receiving medication for 7 days preceding the study and
from using alcohol or xanthine-containing beverages or foods for 24 hr
prior to exposure and throughout the sample collection period Smoking was
prohibited from 1 hr before to 4 hr after initiation of the exposure.
Subjects were housed at the L.A.B. Bio-Research Clinical Research Center from die evening prior to exposure until after the final blood and urine
collection period (24 hr after initiation of the 2-hr exposure). The subjects
fasted under supervision for at least 10 hr prior to exposure until approximately 0.5 hr postexposure. Standard meals were provided at 0.5 and 6 hr
after exposure and at appropriate times thereafter. Water was withheld 2
hr prior to and during the exposure. Prior to exposure, each subject's arm
was examined by a physician to determine that skin was intact and a venous
catheter was installed in the unexposed arm.
This study was conducted in accordance with the clinical research guidelines established by the Medical Research Council of Canada, the Basic
Principles defined in United States (21 CFR Part 312.120), and the principles
enunciated in the Declaration of Helsinki (Hong Kong, 1989). The protocol
and informed consents were approved by the Bio-Research Institutional
Review Board which is constituted and operated in accordance with the
requirements for ethical considerations in research involving human subjects
(Medical Research Council of Canada, Report No. 6, 1978; U.S. Code of
Federal Regulations, 21 CFR Part 56).
Exposure conditions. Each subject exposed one arm to a target concentration of 50 ppm I3C2-BE for 2 hr using a 400-liter Rochester-type stainless
steel and glass inhalation chamber equipped with sealed entry ports for the
arms. A thin rubber strap was suspended from the top of die chamber to
support the subject's arm.
Test atmospheres were generated by pumping '3C2-BE into a glass co-
122
CORLEY ET AL.
TABLE 1
Demographic Data From Human Volunteers
Surface area (m 2 )'
Subject
no.'
1
2
3
4
5
6
Mean
SD
Age
(year)
Height
(cm)
Weight
(kg)
Body
Total arm
Arm exposed
23
23
36
20
36
45
178
163
170
171
168
168
83.8
60.0
62.5
74.7
66.6
82.3
:>.O2
.64
.72
.87
.75
.92
0.1818
0.1476
0.1548
0.1683
0.1575
0.1728
0.1430
0.0981
0.1338
0.1283
0.1289
0.1397
30.5
9.9
169.7
4.9
71.65
10.15
.82
0.14
0.1638
0.0127
0.1286
0.0161
° Subjects 1 - 5 were Caucasian; subject 6 was black.
' T h e total body surface area (SA) was estimated using the formula of DuBois and DuBois (1916):
Area (cm 2 ) = Weight (kg) 0 4 5 X Height (cm) 0 7 2 3 X 71.84
The SA for the whole arm was estimated to be 9% of the total body surface area (Boyarsky and Smith, 1957; Raisz and Scheer, 1959). The SA for
the exposed arm was calculated from total arm — unexposed part (circumference X length).
lumn filled with glass beads where heated air was passed to vaporize the
test material according to Miller et al. (1980). The concentration of I3 C 2 BE was monitored continuously using a Miran 1A infrared analyzer (Foxboro Co., Foxboro, MA). At 2-min intervals, the concentration was recorded
and time-weighted average concentrations were calculated for each subject
since each exposure session could accommodate three subjects at a time
with each subject inserting their arm at staggered intervals to allow for the
collection of blood samples.
Blood sampling and analysis.
Peripheral venous blood samples were
obtained via the antecubital vein catheter in the nonexposed arm before
exposure and at 10, 20, 30, 40, 60 min and 1.5, 2, 2.25, 2.5, 3, 3.5, 4, 8,
12, 16, and 24 hr after the initiation of the 2-hr exposure. The samples
were drawn into heparinized Vacutainers, divided into two polypropylene
tubes, and frozen at — 15°C until the last sample was collected. Immediately
after exposure, the hands used for finger prick sampling were immersed
twice in fresh, lukewarm water for approximately 30 sec to wash the surface
of the skin and produce capillary vasodilation according to the procedure
used by Johanson and Boman (1991). The finger prick blood samples were
then collected into two 700-^1 Microtainers containing heparin, transferred
to polypropylene tubes, and frozen at - 1 5 ° C until the last sample was
collected. The exposed arm was then rinsed under running water and
cleansed with a mild soap. All samples were then shipped on dry ice to
The Dow Chemical Co. where they were stored frozen (—80°C) until analyzed (within 2 weeks). Analyses of 13 C 2 -BE and I3 C 2 -BAA, as their respective pentafluorobenzoyl and pentafluorobenzyl derivatives, were conducted
by capillary gas chromatography with negative chemical ionization-mass
spectrometry according to the method of Bormett et al. (1995). Detection
limits for I 3 C 2 -BE and I 3 C 2 -BAA in blood were 1.5 ng/g for each material
or 0.013 and 0.011 fiM, respectively.
Urine collection and analysis.
A urine sample was obtained from each
subject before exposure and at 0- to 12- and 12- to 24-hr intervals following
initiation of exposure. The urine was refrigerated (~4°C) until the end of
each collection interval, the volume and pH of each interval were recorded,
and two aliquots ( ~ 1 0 ml each) were taken and frozen from each interval
for analysis of metabolites. Analyses of free I 3 C 2 -BE and 13 C 2 -EG in the
urine as their respective pentafluorobenzoyl derivatives and free I 3 C 2 -BAA
as a pentafluorobenzyl derivative were conducted by capillary gas chromatography with negative chemical ionization-mass spectrometry according
to the method of Markham et al. (1995). Total (free and conjugated) 13 C 2 -
BE, I3 C 2 -BAA, and I3 C 2 -EG were also determined after acid hydrolysis.
Detection limits for free and total I3 C 2 -BE, I3 C 2 -BAA, and I3 C 2 -EG in urine
were 35.8 ng/g (0.30 /IM), 84.0 ng/g (0.64 /XM), and 25.6 ng/g (0.41 UM),
respectively.
The analysis for glycolic acid (GA), a major metabolite of EG, in urine
was conducted after the original protocol was completed and was, therefore,
based on I2 C-GA external standard quantitation. Extraction and analyses
for the pentafluorobenzoyl and pentafluorobenzyl derivatives of alcohol and
acid functional groups of GA, respectively, were similar to the method used
for analysis of BE and BAA by capillary gas chromatography with negative
chemical ionization-mass spectrometry.
To confirm the identity of the NBGL conjugate, urine samples were
filtered (0.2 fim Anotop 10; Alltech Assoc., Inc., Deerfield, IL) and analyzed
by gradient reversed-phase HPLC with positive electrospray ionization/
mass spectrometry (+ESI/MS) according to Markham et al. (1995). The
detection limit for the NBGL conjugate was 200 ng/g (0.73 /JM).
Fortified control blood and urine shipment samples.
To ensure the
integrity of the blood and urine samples shipped from L.A.B. Bio-Research
to The Dow Chemical Co. for analysis of metabolites, replicate fortified
controls were prepared from each subject's blood and urine collected prior
to exposure. The blood samples were spiked with known amounts of BE
and BAA and the urine samples were spiked with BE, BAA, and EG. All
fortified samples were prepared, treated, and shipped along with the samples
obtained following exposure to I3 C 2 -BE.
Clinical pathology.
Blood samples for hematology were collected just
prior to exposure, at the end of exposure, and at approximately 24 hr after
initiation of exposure (see Table 2). Samples for serum chemistry were
collected approximately 24 hr after initiation of exposure (Table 2). Approximately 12 hr after initiation of exposure, a blood sample was taken for
creatinine clearance determination.
A sample of urine was taken approximately 24 hr after initiation of the
exposure for urinalysis (see Table 2). When the 24-hr collection interval
was completed and samples were taken for analysis of BE and metabolites,
all remaining urine was pooled for each subject and one aliquot was taken
for creatinine clearance determinations.
Calculations.
The concentrations of BE and BAA in blood samples
collected from the exposed arm by finger prick at the end of the exposure
were compared to samples collected from the unexposed arm via catheter
using a two-way analysis of variance (SAS for microcomputers. Version
123
HUMAN DERMAL UPTAKE OF 2-BUTOXYETHANOL VAPOR
TABLE 2
Hematological, Clinical Chemistry, and Urinalysis
Determinations
Hematology
Hemoblobin
Hematocrit
Total and differential leukocyte count
Red blood cell count
Platelet count
Erythrocyte count
Calculated RBC indices
Serum chemistry
BUN
Microscopic examination
Creatinine
Total bihrubin
Alkaline phosphatase
Alanine aminotransferase
Aspartate aminotransferase
LDH
Na
K
Ca
P
Glucose
Haptoglobin
Plasma hemoglobin
Urinalysis
PH
Specific gravity
Protein
Glucose
Ketones
Bilirubin
Blood
Nitrite
Urobilinogen
Urine Screen for Drugs of Abuse
Cocaine
THC
Other
HIV
6.04, SAS Institute, Cary, NC). The areas under the curve for both BE and
BAA were calculated by the trapezoidal rule from 0 to 24 hr (AUQ,.},)
and extrapolated to infinity (AUC) by addition of the quantity equal to the
last measurable concentration divided by the elimination rate constant (*"d).
The apparent first-order elimination rate constant (K^) was calculated from
a log-linear plot of the last three (or more) nonzero concentrations in the
blood concentration vs time curve. The terminal phase blood half-life was
obtained by dividing 0.693 by the elimination rate constant (Kel). The maximum measured blood concentration of BE and BAA ( C ^ and the time
of the maximum value ( i j were also determined. The permeability coefficient (ip) for the dermal uptake of BE was determined by fitting the model
to the concentration of BE and BAA in blood from this study and the
concentration of BE in finger prick blood from the Johanson and Boman
(1991) study using the PB-PK model of Corley et al. (1994).
RESULTS
Exposure conditions. The concentration of 13C2-BE vapor, determined by fixed-wavelength IR analysis, for each
of the subjects averaged 50.2 ppm with an overall nominal
concentration (total amount of material used divided by the
total volume of air) of 51.9 ppm (Table 3). The chamber
temperatures (23-25°C) were similar to the "normal" temperature used by Johanson and Boman (1991) (mean of
22.6°C) while the relative humidity (55-58%) was between
the "normal" (mean of 29%) and "raised" humidities
(mean of 71%).
Clinical pathology and health status. The total blood
volume collected in this study did not exceed 136 ml, except
for subject 3, whose prestudy hematological evaluation was
repeated, bringing his total blood volume collected to 141
ml. No significant deviations were observed in the postexposure hematological, serum chemistry, or urinalysis data even
with an emphasis on markers for hemolysis (data not shown).
Subjects were monitored throughout the 24 hr of confinement after the initiation of the exposure with no adverse
events reported.
BE toxicokinetics. BE was not detected in blood samples
from the unexposed arm until 30 min into the exposure
TABLE 3
Exposure Conditions and Toxicokinetics of BE and BAA
Chamber conditions
I3
C-BE in chamber air (ppm)
Analytical concentration
Nominal concentration
Temperature (°C)
Relative humidity (%)
BE Toxicokinetics
A U Q K Omol • hr/Ly
AUCV, (jimol • hr/I,)"
Half life in blood (hry
Cmax(/iM)
Tmax (hr)
Concentration in blood at 2 hr (/JM)
Finger prick, exposed arm
Catheter, unexposed arm
Ratio (finger prick/catheter)*
BAA toxicokinetics
AUQt, (/imol • hr/L)
AUC^. (/imol • hr/L)
Half life in blood (hr)
C m a x (JJLM)
Tmax (hr)
Concentration in blood at 2 hr (JIM)
Finger prick, exposed arm
Catheter, unexposed arm
Ratio (finger prick/catheter)''
Mean ± SD
Range
(min-max)
50.2 ± 0.6
51.9
24.1 ± 0.6
56.3 ± 0.9
49.5-50.9
—
23-25
55-58
0.091
0.158
0.66
0.07 ± 0.03
2.1 ± 0.3
0.039-0.134
0.152-0.163
0.6-0.72
0.03-0.11
1.5-2.25
35.5 ± 25.1
0.037 ±0.019
1488 ± 1908
19.5-85.9
0.016-0.067
374-5337
3.69
3.77
3.27
0.59
3.7
± 0.97
± 0.96
± 0.72
±0.18
± 0.4
2.54-5.39
2.65-5.45
2.35-4.38
0.44-0.92
3.0-4.0
0.97 ± 0.46
0.31 ± 0.15
3.8 ± 2.9
0.48-1.71
0.18-051
1.7-9.4
* No meaningful K^ was determined for subjects 1,4,5, and 6. Therefore,
AUC and apparent elimination phase half-lives could not be calculated for
BE in these individuals.
* Represents the mean ± SD of the individual subject ratios of finger
prick:catheter blood concentrations (rather than a ratio of the means provided in this table).
124
CORLEY ET AL.
0.12
0.1
0.08 •
-Subject 1
m
UJ
m
o
-Subject 2
-Subject 3
0.06 •
-Subject 4
-Subject 5
0.04 -
8
0.02 -
-Subject 6
itra
s
8
o1
2
Time (hr)
1.0 i
3
0.8 -
CD
c
0.6 -
-Subject 1
-Subject 2
-Subject 3
-Subject 4
0.4 -
-Subject 5
ion
"5
c
8
c
o
O
-Subject 6
0.2 -
0.0
12
16
20
24
Time (hr)
FIG. 1. Concentration of (a) 2-butoxyethanol (BE) and (b) butoxyacetic acid (BAA) in the blood of volunteers exposed (arm-only) to 50 ppm I3 CBE vapor for 2 hr. Blood samples were obtained from the antecubital vein of the unexposed arm.
(subjects 2 and 4, only) (Fig. la). By 1.5 hr, BE was detectable in all six subjects. An apparent steady state was reached
for the concentration of BE in systemic blood between 1.5
and 2 hr (Fig. la). Interestingly, the concentration of BE in
blood was nearly twofold higher in the first blood sample
taken 15 min after the exposure ceased than it was at the
end of the exposure in all but one volunteer (subject 2; Fig.
la). BE was rapidly cleared from the blood with an apparent
elimination half-life of 0.66 hr based upon two of the six
subjects data (Table 3). Due to the rapid loss of BE from
venous blood, no meaningful half-lives or AUC's could be
calculated for the other four subjects.
BAA toxicokinetics. As with BE, BAA was not detected
in blood samples from the unexposed arm until 30 min into
the exposure (subjects 2 and 4 - 6 ) with amounts detected in
all six subjects by 1 hr (Fig. lb). BAA concentrations ex-
125
HUMAN DERMAL UPTAKE OF 2-BUTOXYETHANOL VAPOR
CH
CH 2 -CH 2 -O-CH
2-Butoxyethanol
(BE)
Mixed Function
Oxidase
OH
\
Alcohol and Aldehyde
Dehydrogenase
HO - C H r C H r OH
Ethylene Glycol
(Rat Only?)
O
\
CH-,BE-Glucuronide
(Rat Only?)
- CH 2 - C H j O -
II
-C-OH
Butoxyacetic Acid (BAA)
(
BAA-Glu + BAA-Gly
(Human Only)
Expired Air
(10-20%)
Urine
(50 - 70%)
FIG. 2. Proposed metabolic scheme for 2-butoxyethanol in rats and humans. BAA-Glu and BAA-Gly are the glutamine and glycine conjugates of
BAA, respectively. According to Rettenmeier et al. (1993), the glutamine conjugate represents approximately 90% of the amino acid conjugates of BAA
in humans.
ceeded those of BE at all time points. Peak blood concentrations of 0.44-0.92 fiM were reached 3 - 4 hr after the start
of the 2-hr exposures (Table 3). BAA was less rapidly
cleared from the blood than BE with an apparent elimination
half-life of 3.27 ± 0.72 hr.
Comparison of blood sampling techniques. For the parent compound, BE, the ratio of the finger prick blood samples taken at the end of the 2-hr exposure to the corresponding blood sample taken from the catheter installed in the
unexposed arm, averaged nearly 1500:1 (Table 3), indicating a considerable influence by local absorption for the
finger prick blood samples. Moreover, the concentrations
of BE in the finger prick blood samples were, on average,
37 times higher than the concentrations of BAA. This is
the opposite of the catheter blood sample, in which the
concentrations of BE were approximately one-tenth the concentration of BAA at the end of the exposure. The concentrations of BAA in both blood samples were much more
consistent than that of BE and were, on average, within a
factor of four of each other.
Urinary metabolite profiles. No free BE was detected in
any urine sample (see Fig. 2). Following hydrolysis to cleave
acid-labile conjugates, BE was detectable in only two 0- to
12-hr urine samples at concentrations of 43 (subject 3) and
48 ng/g (subject 6). These concentrations were only slightly
higher than the detection limit of 35.8 ng/g BE in urine (data
not shown).
No free or acid-labile conjugates of ethylene glycol were
detected in any urine sample [25.6 ng/g (free) and 16.7
ng/g (acid-hydrolyzed) detection limits]. Glycolic acid, a
major metabolite of ethylene glycol, also was not detectable above a highly variable background concentration of
2.5 to 13 /xg/g.
Corresponding to the relatively short half-life in blood,
the major metabolite, BAA, was eliminated primarily in the
urine during the first 12-hr collection interval (Table 4).
Consistent with the observations of Rettenmeier et al.
(1993), approximately two-thirds of the total amount of BAA
excreted in the urine was in the form of an acid-labile conjugate, NBGL. The identity of the NBGL conjugate was confirmed by reversed-phased HPLC with positive electrospray
ionization/mass spectrometry (+ESI/MS).
Skin permeability coefficient. Using the data from this
study and the Johanson and Boman (1991) study, the best
estimates of kp, were 2 or 4 cm/hr for normal (23°C, 29%
RH) and elevated (33°C, 71% RH) temperatures and humidities and were unchanged from those determined previously
(Corley et al., 1994). Since there were no statistically identified differences between the normal and elevated temperature/humidity data of Johanson and Boman (1991), an average permeability coefficient (fcp = 3 cm/hr) was used for
general simulations in Figs. 3 and 4.
DISCUSSION
In a previous human vapor exposure study, a unique design was used to evaluate the relative contributions of the
126
CORLEY ET AL.
skin and the lungs to the total uptake of BE (Johanson and
Boman, 1991). This required a significant effort to control
for possible artifacts associated with the operation of respirators, chambers, and blood sampling methods, all while collecting physiological data on heart and ventilation rates. In
the current study, arm-only exposures were used to simplify
the design and minimize potential artifacts that could occur
with the use of respirators and whole-body exposures. 13CLabeled test material was necessary to achieve the very low
detection limits that were required to conduct an arm-only
vapor exposure study and to unequivocally identify BE and
its metabolites in blood and urine resulting from this controlled exposure.
The elimination kinetics of BE and BAA in this study
were consistent with previous studies and appeared to be
independent of the type of exposure. The half-lives for the
elimination of BE and BAA in this study average 0.66 hr
(range: 0.6-0.72 hr in the two subjects for which a meaningful half-life could be determined) and 3.27 hr (range:
2.35-4.38 hr), respectively. In the Johanson and Boman
study, the elimination half-life of BE averaged 0.53-0.6
hr following dermal absorption. Following whole-body exposures under exercise conditions, the elimination halflives for BE and BAA were 0.66 and 4 hr, respectively
(Johanson et al., 1986; Johanson and Johnsson, 1991). For
dermal exposures to neat liquids, the half-lives for the elimination of BE and BAA were 1.3 and 3.1 hr, respectively
(Johanson et al., 1988).
The kinetics of BE in the blood of one volunteer during
0.12
T3
O
m
0.08
LJJ
CD
0.04-
O
O
O
2
4
6
Time (hr)
1.0 -i
TABLE 4
Total Amounts of 13C-BAA Excreted in the Urine of Human
Volunteers That Exposed One Arm to 50 ppm "C-BE for 2 hr
Collection interval (hr)
Free BAA (mg)
(/ig/g Creatinine)
Conjugated BAA (mg)
(/ig/g Creatinine)
Total BAA (mg)
(fig/g Creatinine)
% Conjugated BAA
0-12
12-24
1.02 ± 0.35
(0.69-1.6)
0.11 ± 0.07
(0-0.17)
Total (0-24)
1.13 ± 0.40
(0.69-1.63)
68.8 ± 20.3
(46.3-94.3)
1.97 ± 0.21
0.23 ± 0.19
2.20 ± 0.31
(1.74-2.17)
(1.92-2.77)
(0.09-0.61)
136.0 ± 20.1
(112.6-169.9)
2.99 ± 0.38 0.34 ± 0.23
3.33 ± 0.47
(2.76-3.86)
(2.55-3.60)
(0.15-0.78)
204.8 ± 20.3
(173.6-233.2)
66.4 ± 8.1
65.8 ± 19.6
66.6 ± 8.6
(54.4-74.1)
(47.1-77.7)
(54.4-74.9)
Note. Urine samples were collected in two 12-hr intervals for each subject, sampled for analysis of free and total (acid hydrolyzable) BE, BAA,
ethylene glycol, and glycolic acid, then combined for analysis of creatinine.
Only BAA (free and conjugated) was detected in these samples. Ranges
for individual subject values are given in parentheses.
0.0
12
16
20
24
Time (hr)
FIG. 3. Experimental data (means ± SD) and simulations (lines) for
the concentrations of (a) BE and (b) BAA in the blood of humans exposed
arm-only to 50 ppm I3C-BE for 2 hr. Simulations were conducted assuming
that only one metabolic pathway (via the formation of BAA) operates in
humans In (a), the concentration of BE in blood was simulated using
"normal" physiological parameters throughout the simulation and by increasing the blood flow to the skin for 5 min postexposure to simulate
movement following 2 hr of restraint.
and immediately postexposure appeared different than the
other volunteers (Subject 2, Fig. 1). Peak blood concentrations of BE occured prior to the end of the exposure and no
postexposure increases in BE concentrations were observed.
The reasons for this difference are unknown and may reflect
either experimental variability or physiological differences.
However, the clearance of BE from this subject was similar
127
HUMAN DERMAL UPTAKE OF 2-BUTOXYETHANOL VAPOR
Vmax2C = 0
50 •
Vmax 2C = 5 mg/hr/kg
(Corteyetal., 1994)
"8 40 -
£
.E
5
30 20 -
10 -
0
2
4
6
8
Time (hr)
FIG. 4. Comparison between PB-PK simulations (lines) and the concentrations (symbols) of (a) BE and (b) BAA in the blood of humans exposed
to 20 ppm BE for 2 hr while exercising (50 W) on a bicycle ergonometer (Johanson et al, 1986; Johanson and Johnsson, 1991). Simulations were
conducted using the model structure of Corley et al. (1994) (dotted lines: V m i 2C = 5 mg/hr/kg) and by assuming that only one metabolic pathway (via
the formation of BAA) operates in humans (solid line: V mu 2C = 0).
to that of other subjects and had a half-life of BE in blood
similar to values reported in the literature.
Data from this study were used to validate the analysis
of Corley et al. (1994) indicating that the finger prick blood
sampling technique used by Johanson and Boman (1991)
resulted in an incorrect conclusion that dermal uptake of BE
is quantitatively more important than respiratory uptake in
humans exposed to BE vapor. Their conclusion was based
on the assumption that the finger prick blood samples represented systemic arterial blood concentrations of BE. In the
current study we verified that, when analyzing only for the
parent chemical, finger prick blood samples most likely represented both systemic concentrations and local concentrations of the compound in venous blood draining the skin. A
comparison of these two assumptions of the source of the
finger prick blood sample is depicted in Fig. 4 of Corley
(1996) where a significant overestimate of the uptake of BE
vapor through human skin occurred when finger prick blood
samples were assumed to represent systemic blood.
In addition, the concentrations of BE and BAA in the
finger prick blood samples taken from the exposed arms in
this study were not comparable to the concentrations seen
in blood samples taken from the catheter inserted in the
antecubital vein of the opposite, unexposed arm (Table 3).
In this case, the concentrations of BE averaged nearly 1500fold higher in the finger prick blood samples versus the
corresponding samples taken from the unexposed arm at the
end of the exposure. Differences in blood concentrations
between exposed and unexposed arms have also been documented for topical applications of liquid formulations. For
example, concentrations of MN-diethyl-m-toluamide
(DEET) in blood taken from the exposed arm were considered to be indicative of dermally absorbed DEET plus circulating DEET while the concentrations in blood taken from
the unexposed arm were considered to be indicative of only
systemically circulating DEET (Selim et al, 1995).
For the major metabolite, BAA, the concentrations were
within a factor of four of each other using the two blood
sampling techniques (Table 3). The reason for the higher
concentration of BAA in finger prick blood samples (statistically significant) than in blood obtained from the catheter is
unknown. This difference could reflect any combination of
local metabolism of BE to BAA in the skin, or a slight
preference of human skin versus blood to accumulate BAA
(due to differences in solubility or protein binding), or to
the fact that the finger prick blood samples were inevitably
taken a few minutes after the catheter samples while the
concentration of BAA in blood was rapidly increasing because of the time needed to wash and prepare the exposed
area for sampling. Regardless, this difference in BAA concentration was slight relative to the 1500-fold difference
between the two techniques for the parent compound.
128
CORLEY ET AL.
Lastly, the ratios of the concentration of BE to BAA in
finger prick blood samples taken at the end of the exposure
were the opposite (BE > BAA) of the ratios from the unexposed arm (BAA > BE), the latter of which are more consistent with the pharmacokinetics of BE and BAA (Table 3).
Had Johanson and Boman included BAA in their analyses,
it is likely that this discrepancy would have been uncovered.
Taken together, these data demonstrate that finger prick
blood sampling is an inappropriate technique for the analysis
of parent compounds following dermal exposures under resting conditions unless the source of the blood can be factored
into the analysis. This can be readily accomplished with a
PBPK model. Thus, the alternative assumption that finger
prick blood samples represented venous blood draining the
skin was used to calculate the permeability coefficient for
the uptake of BE vapor through the skin from the data of
Johanson and Boman. As described by Corley et al. (1994),
this assumption, as verified in this study, provided the most
consistent simulation of all of the available human data sets.
The confirmation of a glutamine conjugate of the major
metabolite, BAA, in human urine (Table 4) further demonstrates the difficulties in extrapolating toxicity data developed in rats to humans without a metabolic and kinetic
bridge between the species. Biochemical constants for the
formation of the NBGL conjugate from BAA and its elimination in urine would have to be added and verified in the
PBPK model if a complete description of the fate of the
hemolytic metabolite, BAA, in humans were desired. Preliminary analyses of the blood samples from this arm-only study
for the NBGL conjugate failed to find significant concentrations (compared with BAA concentrations; data not shown)
indicating that NBGL is rapidly cleared after its formation
(presumably in the liver) via glomerular filtration and not
reabsorbed intact. Incorporation of this pathway may account
from some of the previous overpredictions of the amounts
of BAA excreted in the urine (Corley et al., 1994). However,
current descriptions of total BAA (free plus protein-bound)
in the blood of rats and humans, which is relevant for hemolysis, are still adequate and can be used in cross-species,
dose-route, and high-to-low dose extrapolations.
Since no free or conjugated BE, ethylene glycol (free or
conjugated), or glycolic acid were found in the urine of
humans in the present study, these metabolic pathways,
which together account for approximately 25-30% of the
metabolites in rats (see Fig. 2 for the proposed metabolic
scheme) were eliminated from the human PBPK model developed previously (Corley et al., 1994). Simulations of the
data from the present arm-only study, along with the previous human, kinetic study of Johanson et al. (1986) and Johanson and Johnsson (1991), are shown in Figs. 3 and 4, respectively. Although the model predicted a higher area under the
curve and C ^ for BE in blood than was observed in this
arm-only study, the observed data and simulations were
within a factor of two of each other. More importantly for
health risk assessments based on hemolysis, the simulated
concentrations of BAA in blood were in very close agreement with the data. Adjustments to the permeability coefficient to improve the fit of the arm-only exposure data would
have resulted in a correspondingly decreased ability to fit
the BAA data from the arm-only study as well as the other
two human studies, which were well-described. Thus, the
choice of the skin permeability coefficient represented a balance between the different studies to achieve an overall description of the available data.
The increases observed in the concentration of BE in
blood immediately after the exposures ceased in five of the
six subjects were unexpected (Fig. la). However, when consideration was given to the fact that the subjects were partially immobilized during the exposure and were allowed to
move freely about only after the exposure ceased and finger
prick blood samples were obtained, it is reasonable to expect
a change in the local blood flow to the exposed arm. When
the PBPK model was adjusted by increasing the blood flow
to the skin for only 5 min postexposure (fourfold increase),
this change in the concentration of BE in blood was duplicated by the model (Fig. 3a). Alternatively, washing the arm
after the exposure ceased may have altered the permeability
of the skin and thus momentarily enhanced the absorption
of any residual BE on the surface of the skin into circulation.
However, in this vapor exposure, the amount of residual
BE on the surface should be minimal and the postexposure
increases can be explained by locally altered blood flow to
the skin.
Using the existing PBPK model for BE and BAA (Corley
et al., 1994) and deleting the secondary pathways for the
metabolism of BE via MFO (ethylene glycol pathway) and
conjugation (glucuronidation or sulfation), estimates of the
relative contribution of the skin to the total uptake of BE
by humans exposed to the ACG1H TLV concentration of 25
ppm for 8 hr were calculated for resting and exercise conditions (Table 5). Dermal exposures were simulated assuming
both worst-case conditions where no clothing is worn (100%
of the total body surface area was fully exposed) and more
realistic conditions where only 25% of the total body surface
area, such as when parts of both arms and the head, are
exposed. Under worst-case resting conditions, where respiratory rates are at their lowest and no clothing is worn, the skin
can account for no more than 15-27% (normal-elevated
temperatures and humidities as used by Johanson and Boman, 1991) of the total uptake of BE vapor, not 75% as
discussed previously. With more realistic exposures (25%
of the surface area), the skin accounted for 4.4-8.4% of
the total uptake using the lower respiratory rates of resting
conditions. Under simulations of exercise (50 W) with elevated respiratory rates and cardiac output, dermal uptake
accounted for 1.2-2.3% and 4.6-8.7% of the total uptake
assuming either 25 or 100% of the body surface area was
exposed, respectively. With the absorption of BE driven pri-
HUMAN DERMAL UPTAKE OF 2-BUTOXYETHANOL VAPOR
TABLE 5
Simulated Relative Contributions of the Skin
to the Total Uptake of BE
Permeability coefficient (cm/hr)
Surface area exposed (%): 25/100
3
25/100
4
25/100
6.4/21.4
215/256
49/58
41/35
8.4/26.7
220/275
50/62
40/32
1.8/6.7
779/820
183/192
11/10
2.3/8.7
784/839
184/195
11/10
Resting Physiology
Dermal uptake (% of total)
Total uptake (mg)
C m BAA (UM)
NOEL/C
4.4/15.4
210/238
48/54
42/37
129
sis that blood sampling techniques can have a significant
impact on pharmacokinetic analyses and interpretation of
controlled human exposure studies, especially when the dermal route of exposure is used and only the parent chemical
is analyzed. Consistent with other solvents, the respiratory
tract represents the most significant contributor to the total
uptake of BE vapor. As a result, the use of appropriate
respiratory protection will effectively reduce the overall internal dose of BE in situations where vapor exposures are
expected. Furthermore, the use of protective clothing will
further reduce the uptake of BE vapor by the skin and minimize exposure to liquids containing BE, the other potentially
significant route for human exposures.
50 W exercise
Dermal uptake (% of total)
Total uptake (mg)
C ^ BAA (JIM)
1.274.6
775/802
182/188
11/11
c. Simulations of a 70-kg human exposed whole-body to the ACGIH
TLV concentration of 25 ppm continuously for 8 hr under resting or exercise
(50 W) conditions with the PBPK model of Corley et al. (1994) using the
modifications described in the text. Dermal uptake was assumed to occur
under either worst-case conditions where no clothing was wom (100% of
the total body surface was fully exposed to BE vapor for the entire 8 hr)
or more realistic exposures where most of both arms and the head region
were exposed (25% of the total surface area). A series of skin permeability
coefficients were used to simulate the normal (>tp = 2 cm/hr; 23°C, 29%
RH) and elevated (/kp = 4 cm/hr, 33°C, 71% RH) relative temperatures and
humidities used by Johanson and Boman (1991). The average permability
coefficient (Jtp = 3 cm/hr) was also included, since it was used for general
simulations as described in the text. Peak blood concentrations ( C ^ ) of
BAA for each simulation were calculated for comparison to the in vitro
hemolytic no effect level of 2000 fiM for normal humans and humans with
potential susceptibility (Udden and Patton, 1993; Udden, 1993).
ACKNOWLEDGMENTS
The authors are grateful to Dr. Albert Rettenmeier (University of Tubingen, Germany) for supplying the glutamine conjugate and hydrolysis
conditions and Dr. Gunnar Johanson (National Institute of Occupational
Health, Sweden) for helpful discussions on possible interpretations of the
Johanson and Boman (1991) study. In additions the authors thank the members of The CMA Ethylene Glycol Ethers Panel and Toxicology Research
Task Group for funding this study and for their help in its design and
conduct.
REFERENCES
Bartnik, F. G., Reddy, A. K., Klecak, G., Zimmerman, V., Hostynek, J. J.,
and Kunstler, K. (1987). Percutaneous absorption, metabolism, and hemolytic activity of n-butoxyethanol. Fundam, Appl. Toxicol. 8, 59-70.
Bormett, G., Bartels, M. J., and Markham, D. A. (1995). Determination of
2-butoxyethanol and butoxyacetic acid in rat and human blood by gas
chromatography/mass spectrometry. J. Chromatogr. B. 665, 315-325.
Boyarsky, S., and Smith, H. W. (1957). Renal concentrating operation at
low urine flows. J. Urol. 78, 511 -524.
marily by the respiratory tract, the total uptake estimates and
peak blood concentrations for BAA for the different skin
permeabilities were not significantly different from each
other within each set of simulations for resting and for exercise conditions.
Peak blood concentrations of BAA were compared to the
in vitro no effect level of 2 ITIM in 4-hr incubations of BAA
with blood from normal humans and humans with potential
susceptibility to hemolysis (Udden and Patton, 1993; Udden,
1993). The in vitro NOEL was 32- to 42-fold higher than
peak blood concentrations predicted for resting physiology
and approximately 10- to 11-fold higher than peak blood
concentrations predicted under exercise conditions for an 8hr continuous exposure to 25 ppm BE (Table 5). More recent
in vitro studies with human blood suggest that prehemolytic
changes (increased osmotic fragility) are not reached until
concentrations are well above 5 mM in 4-hr incubations (Udden, personal communication), a concentration that is unlikely to be achieved except when large quantities of BE are
intentionally ingested.
In conclusion, data from this study confirmed our hypothe-
Corley, R. A. (1996). Assessing the risk of hemolysis in humans exposed
to 2-butoxyethanol using a physiologically-based pharmacokinetic model.
Occup. Hyg. 2, 45-55.
Corley, R. A., Bormett, G. A., and Ghanayem, B. I. (1994). Physiologicallybased pharmacokinetics of 2-butoxyethanol and its major metabolite, 2butoxyacetic acid, in rats and humans. Toxicol. Appl. Pharmacol. 129,
61-79.
DuBois, D., and DuBois, E. F. (1916). Clinical calorimetry. Tenth paper.
A formula to estimate the approximate surface area if height and weight
be known. Arch. Int. Med. 17, 863-871.
Ghanayem, B. I. (1989). Metabolic and cellular basis of 2-butoxyethanolmduced hemolytic anemia in rats and assessment of human risk in vitro.
Biochem. Pharmacol. 38, 1679-1684.
Ghanayem, B. I., Burka, L. T., and Matthews, H. B. (1987). Metabolic basis
of ethylene glycol monobutyl ether (2-butoxyethanol) toxicity: Role of
alcohol and aldehyde dehydrogenases. J. Pharmacol Exp. Ther. 242,
222-231.
Ghanayem, B. I., and Sullivan, C. A. (1993). Assessment of the haemolytic
activity of 2-butoxyethanol and its major metabolite, butoxyacetic acid,
in various mammals including humans. Human Exp. Toxicol. 12, 3 0 5 311.
Jacobs, R. R., and Phanprasit, W. (1993). An in vitro comparison of the
permeation of chemicals in vapor and liquid phase through pig skin.
Amer. Ind. Hyg. Assoc. J. 54(10), 560-575.
130
CORLEY ET AL.
proved methodology for generating controlled test atmospheres. Am. Ind
Johanson, G., and Boman, A. (1991). Percutaneous absorption of 2-butoxyethanol vapour in human subjects. Br. J. Indusi. Med. 48, 788-792.
Hyg. Assoc. J. 41, 844-846.
Johanson, G., Boman, A., and Dynesius, B. (1988). Percutaneous absorption
Raisz, L. G., and Scheer, R. L. (1959). Studies on the renal concentrating
of 2-butoxyethanol in man. Scand, J. Work Environ. Health 14, 101 —
mechanism. II. Effect of small acute changes in solute excretion. J. Clin.
109.
Invest. 38, 1-7.
Johanson, G., and Johnsson, S. (1991). Gas chromatographic determination
Rettenmeier, A. W., Hennigs, R., and Wodarz, R. (1993). Determination of
of butoxyacetic acid in human blood after exposure to 2-butoxyethanol.
butoxyacetic acid and N-butoxyacetylglutamine in urine of lacquerers
Arch. Toxicol. 65, 433-435.
exposed to 2-butoxyethanol. Int. Arch. Occup. Environ. Health 65, S 1 5 1 S153.
Johanson, G., Kronborg, H., Naslund, P. H., and Nordquist, M. B. (1986).
Toxicokinetics of inhaled 2-butoxyethanol (ethylene glycol monobutyl
Riihimaki, V., and Pfaffli, P. (1978). Percutaneous absorption of solvent
ether) in man. Scand. J. Work Environ. Health 12, 594-602.
vapors in man. Scand. J. Work Environ. Health 4, 73-85.
Markham, D. A., Bartels, M. J., Corley, R. A., and Rettenmeier, A. W.
Selim, S., Hartnagel, R. E., Osimitz, T. G., Gabriel, K. L., and Schoenig,
Analysis of 2-butoxyethanol and metabolites in human urine by GC &
G. P. (1995). Absorption, metabolism and excretion of N,N-diethyl-mESI/MS. 43rd ASMS Conference on Mass Spectrometry and Allied Toptoluamide following dermal application to human volunteers. Fundam.
ics, May 21-26, 1995. [Abstract WPG 170]
Appl. Toxicol. 25, 95-100.
McDougal, J. N., Jepson, G. W., Clewell, H. J., HI, Gargas, M. L., and
Udden, M. M., and Patton, C. S. (1993). Hemolysis and deformability of
Andersen, M. E. (1990). Dermal absorption of organic chemical vapors
erythrocytes exposed to butoxyacetic acid, a metabolite of 2-butoxyethain rats and humans. Fundam. Appl. Toxicol. 14, 299-308.
nol. I. Sensitivity in rats and resistance in normal humans. J. Appl. Toxicol. 14,91-96.
McDougal, J. N., Jepson, G. W., Clewell, H. J. HI, MacNaughton, M. G.,
and Andersen, M. E. (1986). A physiologically-based pharmacokinetic
Udden,M. M. (1993). Hemolysis and deformability of erythrocytes exposed
model for dermal absorption of vapors in the rat Toxicol. Appl. Pharmato butoxyacetic acid, a metabolite of 2-butoxyethanol. n. Resistance in
col. 85, 286-294.
red blood cells from humans with potential susceptibility. J. Appl. ToxMiller, R. R., Letts, R. L., Potts, W. J., and McKenna, M. J. (1980). Imicol. 14,97-102.