Rapid analysis of Lewisite metabolites in urine by high-performance liquid chromatography-inductively coupled plasma-mass spectrometry.

Abstract

A high-throughput method has been developed for determining Lewisite [dichloro(2-chlorovinyl)arsine] exposure by measuring the urine metabolite 2-chlorovinylarsonous acid (CVAA) and the oxidized metabolite 2-chlorovinylarsonic acid (CVAOA). The rapid sample preparation included a simple dilution of 400 microL of urine with 40 microL of water and 1 mL of buffer containing an internal standard and brief centrifugation prior to analysis by high-performance liquid chromatography-inductively coupled plasma-mass spectrometry (ICP-MS). CVAOA and CVAA were eluted isocratically with retention factors of approximately 3.0 and 4.2, respectively, from a reversed-phase polar embedded column with a cycle time of 5 min per sample. The dynamic reaction cell, typically used to remove polyatomic isobaric interferences, was not required for ICP-MS analysis because of the resolution of chloride from arsenical peaks of interest. This method was used to detect CVAA and CVAOA in the urine of a rat administered Lewisite up to 24 h after exposure. The method demonstrated linearity over at least three orders of magnitude and had a method detection limit of 1.3 microg/L as CVAA (1.4 microg/L CVAOA). The relative standard deviations for quality control samples ranged from 3 to 6%. The method was sensitive and selective with no false positives in 100 different urine samples collected from individuals with no known exposure to Lewisite. Ninety-six samples could be analyzed in an 8-h day.

Full text

Page 1

A high-throughput method has been developed for determining
Lewisite (dichloro(2-chlorovinyl)arsine) exposure by measuring the
urine metabolite 2-chlorovinylarsonous acid (CVAA) and the
oxidized metabolite 2-chlorovinylarsonic acid (CVAOA).The rapid
sample preparation included a simple dilution of 400 µL of urine
with 40 µL of water and 1 mL of buffer containing an internal
standard and brief centrifugation prior to analysis by high-
performance liquid chromatography–inductively coupled plasma-
mass spectrometry (ICP-MS). CVAOA and CVAA were eluted
isocratically with retention factors of approximately 3.0 and 4.2,
respectively, from a reversed-phase polar embedded column with a
cycle time of 5 min per sample. Use of the dynamic reaction cell,
typically used to remove polyatomic isobaric interferences, was
not required for ICP-MS analysis due to resolution of chloride from
arsenical peaks of interest.This method was used to detect CVAA
and CVAOA in the urine of a Lewisite-administered rat up to 24 h
after exposure.The method demonstrated linearity over at least
three orders of magnitude and had a method detection limit of 1.3
µg/L as CVAA (1.4 µg/L CVAOA).The relative standard deviations
for quality control samples ranged from 3 to 6%.The method was
sensitive and selective with no false-positives in 100 different urine
samples collected from individuals with no known exposure to
Lewisite. Ninety-six samples could be analyzed in an 8-h day.
Introduction
Dichloro(2-chlorovinyl)arsine was synthesized in 1903 by
Julius Arthur Nieuwland at Catholic University (Washington,
D.C.) by reacting acetylene and arsenic trichloride (1). Winford
Lee Lewis synthesized and purified dichloro(2-chlorovinyl)ar-
sine by adding hydrogen chloride before distillation and de-
tailed the toxicity of the compound (2). Dichloro(2-
chlorovinyl)arsine was subsequently named Lewisite.
Lewisiteisanoily,colorlessliquidinitspureform.Itcanap-
pearambertoblackinitsimpureformwithanodorresembling
geraniums (3,4). It is a highly toxic vesicant, causing irritation
of the skin and mucous membranes within minutes after con-
tact and can blister within hours (3–6). In addition to the blis-
teringeffects,dermalexposuretoLewisitecanresultinelevated
urinary arsenic levels and significant loss of blood pressure in
what is called “Lewisite Shock” (7). A low-dose exposure to
skin (14 µg) can cause vesication, and higher doses (30–50
mg/kg) can lead to death (4). Lewisite has been stockpiled as a
chemical warfare agent and has no medical or other use.
Lewisitewentintoproductionin1918,whichwastoolatefor
battlefielddeployment in World War I(WWI).More than 45,000
tons of Lewisite are estimated to have been produced since
1903 (2). Between WWI and World War II (WWII), at least 11
countries produced or possessed Lewisite (2). After WWII, most
of these countries disposed of their Lewisite stockpiles. There
are a few sites in the U.S. where Lewisite has been stored while
awaiting disposal or was discovered in former military facilities
(4,8,9). Several nations have allegedly used Lewisite- or
Lewisite-mustard-containing weapons in past conflicts (4,10).
Therefore, concerns remain that Lewisite might be used to
cause large numbers of injuries and deaths.
Urban and von Tersch used ab initio molecular orbital cal-
culations to predict that Lewisite would consist predominantly
of the trans-isomer, very little of the cis-isomer, and negli-
gible amounts of the geminal-isomer (11). Smith et al. (12) in-
terpreted1H and13C nuclear magnetic resonance spectroscopy
data and gas chromatography–mass spectrometry (GC–MS)
analysis of the 1,2-ethanedithiol derivative to imply that the
Lewisite samples they analyzed consisted of 95.2% trans-
isomer, 2.7% geminal-isomer, less than 1% cis-isomer, and a
small remaining unidentified fraction. In either case, the trans-
isomer is by far the predominant isomer in production mate-
rials.
The primary metabolite of Lewisite in animals is the hy-
drolysis product chlorovinylarsonous acid (CVAA) shown in
Figure 1 (9,13). Lewisite is rarely found in the environment be-
RapidAnalysis of Lewisite Metabolites in Urine by
High-Performance Liquid Chromatography–Inductively
Coupled Plasma-Mass Spectrometry*
Rayman D. Stanelle1,†, William J. McShane2, Elena N. Dodova3, R. Steven Pappas1, and Robert J. Kobelski1
1Centers for Disease Control and Prevention, 4770 Buford Hwy, MS F-44,Atlanta, Georgia 30341;2Battelle-Battelle Eastern
Science and Technology Center, 1204 Technology Dr.,Aberdeen, Maryland 21001; and3Battelle-Centers for Disease Control and
Prevention, 4770 Buford Hwy, MS F-50,Atlanta, Geogria 30341
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
1
Journal of AnalyticalToxicology,Vol. 34, April 2010
* Disclaimer: The use of trade names and commercial sources is for identification only and does not
imply endorsement by the Centers for Disease Control and Prevention or the U.S. Department of
Health and Human Services
†Author to whom correspondence should be addressed: Rayman D. Stanelle, Centers for
Disease Control and Prevention, 4770 Buford Hwy, MS F-44, Atlanta, GA 30341.
E-mail: fmk9@cdc.gov.
Abstract

Page 2

Journal of AnalyticalToxicology,Vol. 34, April 2010
2
cause of the rapid rate of hydrolysis to CVAA (14–17). In the en-
vironment, CVAA may be oxidized to 2-chlorovinylarsonic acid
(CVAOA) (14), which may also be found in urine (19). Both
CVAA and CVAOA still retain the toxic properties of Lewisite, al-
beit at lower potency (17–20). CVAA and CVAOA are not natural
environmental forms of arsenic and can be detected in urine
only after exposure to Lewisite or CVAA.
An acute exposure to Lewisite in rats can result in elevated
total arsenic in urine as shown in work presented here and
elsewhere (21). Total arsenic may be determined by induc-
tively coupled plasma (ICP)-MS (22,23). However, determina-
tion of total arsenic does not differentiate between CVAA, in-
organic species, and metabolically derived organoarsenic
species. In order to differentiate between the forms of arsenic
found in different compounds by ICP-MS, a separation tech-
nique must be employed (24–27).
Several analytical methods have measured CVAA to deter-
mine Lewisite exposure. Szostek and Aldstadt (28) used 1,3-
propanedithiol or 1,2-ethanedithiol derivatization with solid-
phase microextraction (SPME)-GC–MS to examine CVAA in
soil samples from a military installation. Wooten et al. (5) de-
termined CVAA in urine by using GC–MS for the 1,3-
propanedithiol derivative. Wada et al. (18) used reversed-phase
liquid chromatography (HPLC)–tandem MS to analyze the
degradation products of Lewisite in soil extracts. Although
GC–MS and SPME-GC–MS methods may have very good sen-
sitivity, ICP-MS is highly sensitive due to matrix destruction
and high ionization efficiency as a result of the plasma tem-
perature. GC–MS and SPME-GC–MS have advantages but may
also require derivatization and SPME steps that add to sample
preparation time.
Kinoshita et al. (19) used a C8reversed-phase column for
HPLC–ICP-MS analysis of CVAA and CVAOA in mouse urine
afteronlydilutionandpHadjustmentsamplepreparationsteps;
they used oxygen introduced as an ICP-MS reaction cell gas to
convert As+(mass 75) to AsO+(mass 91) to avoid interference
from ArCl+(mass 75). Addition of cell gases suppresses ICP-MS
polyatomic interference problems as well as the analyte signal.
We optimized instrument and chromatography conditions to
minimize formation of ArCl+and to resolve chloride and, thus,
any minimal remaining ArCl+from arsenicals of interest. We,
therefore, eliminated the need for a cell gas and the conse-
quential suppression of arsenic signal in the method presented
here.
Emergency response methods must be able to analyze hun-
dreds or thousands of samples in a single event. These methods
must be rugged and easily transferrable to other laboratories
and instruments. A rapid HPLC–ICP-MS method is presented
here for the quantitative determination of CVAA in urine with
concurrent measurement of CVAOA when necessary. The
method is capable of 12 samples/h throughput (96 samples in
an 8-h workday).
Experimental
Materials
The U.S. Army Medical Research Institute of Chemical
Defense (Aberdeen Proving Ground, MD) provided CVAA at a
concentration of 746 µg/mL. Lots of the L1 standard (one
chlorovinylgroupbondedtothearsenicatom)received95–98%
purity. Purity on the basis of arsenic content was verified upon
receipt using our established total arsenic method (22). Ultra-
pure >18 MΩ∙cm water and organic-free water came from an
Aqua Solutions water system (Jasper, GA). Tetrabutylammo-
nium hydroxide (TBAH, 40% aqueous solution), succinic acid,
2-propanol (IPA, electronics-grade), Fluka sodium cacodylate
(DMA), and arsenobetaine (AsB) were purchased from Sigma
Aldrich (St. Louis, MO). Sodium arsonoacetate (AA) was pur-
chased from TCI America (Portland, OR). Inorganic arsenic
[As(III) and As(V)] standards were purchased from SPEX Cer-
tiprep (Metuchen, NJ). Monosodium acid methane arsonate
was obtained from Chem Services (West Chester, PA). GFS
Chemicals double-distilled ammonium hydroxide and nitric
acidwereobtainedfromFisherScientific(Suwanee,GA).Tama-
pure 10 double-distilled 35% hydrogen peroxide was obtained
from Moses Lake Industries (Moses Lake, WA).
Preparation of standards, quality controls, and samples
Aqueous solutions of the CVAA intermediate standards were
prepared by dilution to concentrations of 100, 1608, 4026,
7962, and 31,847 µg/L. Quality control (QC) intermediate stan-
dards were prepared at 796, 12,170 and 20,018 µg/L. Short-
term storage of the CVAA spikingsolutions was at 2–4°C. Long-
term storage (>4 days) was at –20°C. 2-Chlorovinylarsonic acid
standards (CVAOA) were prepared from CVAA standards by hy-
drogen peroxide oxidation as described by Kinoshita et al. (19).
The instrument response of five individual base urine sam-
ples spiked at a concentration of 35 µg/L of CVAA were com-
pared to the response from five base urine samples spiked with
the same concentration of CVAA that had been oxidized to
CVAOA on two different days to confirm equivalence.
The intermediate internal standard solution was made by
adding 300 mg of sodium arsonoacetate to 1 L of water. The
diluent was prepared by diluting 4.0 mL of the intermediate in-
ternal standard solution with mobile phase to a final volume of
1.0 L to give a final concentration of 1.2 mg/L of sodium ar-
sonoacetate. The pH of the diluent was adjusted to 5.5 and
stored at 2–4°C.
Urine samples from anonymous donors were collected in
pre-screened polypropylene urine cups. Total arsenic concen-
tration for each sample was determined by inductively coupled
plasma-dynamic reaction cell-mass spectrometry (ICP-DRC-
Figure 1. Primary metabolite and hydrolysis product of Lewisite.

Page 3

Journal of AnalyticalToxicology,Vol. 34, April 2010
3
MS) (23). The urine samples with the lowest concentrations of
arsenic were pooled for use as base urine.
The matrix-matched CVAA standards and quality controls
were created by spiking 400 µL of base urine with 40-µL
aliquots of each of the intermediate CVAA calibration or QC so-
lutions. Then, 1000 µL of diluent with internal standard
sodium arsonoacetate was added to each standard or QC for a
total volume of 1440 µL. The final standard concentrations of
CVAA in urine were 9.09, 146, 366, 724, and 2895 µg/L with
quality control concentrations of 72.4, 1106, and 1820 µg/L.
The centrifuge tubes were capped, vortex mixed briefly, and
centrifuged for 4 min at 1.5 × 104× g. A 0.85-mL aliquot of the
supernatant was transferred to an autosampler vial and capped.
Urine samples were prepared for analysis in a similar manner,
except that the substitution of 40 µL ultrapure water came in
place of the CVAA standards.
Frozen rat urine collected over the final hour of the 4-h
time periods was obtained from a 2004 Lewisite exposure
study (21). Twelve rats were included in the study. Four rats
were exposed percutaneously to 0 (control group), 2.0, or 10
mg/kg of Lewisite according to an approved animal protocol
including anesthesia, buprenorphine administration to pre-
vent pain, and euthanasia. Urine was thawed prior to analysis.
Mobile phase and diluent preparation
The liquid chromatographic separations were accomplished
with an isocratic mobile phase consisting of 2% (v/v) IPA, 11.6
mM TBAH, and 5.0 mM succinic acid with the pH adjusted to
5.5 using nitric acid.
Instrumentation
Chromatography was performed using a Perkin Elmer
(Waltham, MA) 200 series HPLC (including pumps and de-
gasser) with autosampler at 4°C and column oven at 30°C.
Supelco Ascentis RP-Amide polar-embedded reversed-phase
HPLC columns (Sigma Aldrich) with a 4.6-mm internal diam-
eter and lengths of 50, 100, and 150 mm were used for chro-
matographic separations. The flow rate and injection volume
were 1 mL/min and 20 µL, respectively. Chromatographic data
acquisitionwascontrolledusingChromeraversion1.2software
(Perkin Elmer).
The eluent was introduced into a Perkin Elmer ELAN DRC II
ICP-MSthroughaGlassExpansionTwister50-mLcyclonicspray
chamber and SeaSpray nebulizer for HPLC (Pocasset, MA). The
instrument was operated in evacuated cell mode at 1,550 W
radio frequency (RF) power with nebulizer gas (typically 0.95
mL/min), and other parameters optimized. The ICP-MS was
controlledusingELANversion3.4software(Toulouse,France).
Integration of data
Peaks representing the arsonoacetate internal standard,
CVAA, and other metabolites were integrated by reprocessing
the data with Chromera 1.2 software to obtain peak-area/in-
ternal standard area ratios. A concentration regression model
was generated using the area ratios versus standard concen-
trations. The standard concentration regression model was
used to calibrate peak-area/internal standard area ratios for
samples with previously unknown concentrations.
Determination of method detection limits
The method detection limit (MDL) was determined based on
3 times S0, standard deviation at 0 µg/L CVAA concentration
(31). S0was approximated as total standard deviation [St=
(S(within run)2+ S(between run)2)½] because Chromera was not
capable of integrating a base urine CVAA baseline. Therefore,
S(within run)was calculated from the integrated results of 20 in-
dividual base urine samples spiked with the lowest concentra-
tion standard and run as samples versus the same blank and
calibration standards in the same run. S(between run)was calcu-
lated by extrapolation to 0 µg/L CVAA concentration from a re-
gression plot of between run standard deviations from twenty
runs on fourteen different days versus concentrations of the
lowest three standards (32). Limit of quantitation (LOQ) was
determined as 10 times S0(32).
Results and Discussion
In our initial studies, we examined the purity of the CVAA
standards used to quantify unknown samples. We determined
as much as 4% arsenious acid impurity relative to the mass of
CVAA in one standard preparation. It is probable that this small
fraction of inorganic As(III) resulted from hydrolysis of unre-
acted AsCl3remaining from the synthesis described briefly in
the introduction. On this basis, it is possible that AsCl3may
have been one component of the unidentified fraction of
Lewisite described by Smith et al. (12).
To ensure that analysis of CVAA would not be affected by ar-
senic compounds commonly found in urine, initial studies
were carried out using Supelco Ascentis RP-Amide columns
(150 × 4.6 mm, Figure 2A) with 3-µm particle sizes to resolve,
identify, and assign major peaks from spiked or pre-existing ar-
senic compounds in urine. Mass 35 (Cl+) and mass 75 (As+or
ArCl+) chromatograms of standards in urine with a 150-mm
column are shown in Figure 2A. When using the 150-mm
column, monomethylarsonate (MMAV), DMA, As(V), As(III),
AsB, and AA were resolved from CVAOA and CVAA. The main
objective of this method was to rapidly analyze samples for
CVAA without regard to isomers, although CVAA is predomi-
nantly the trans-isomer, while maintaining its elution separate
from AA and the compounds in urine that could otherwise in-
terfere with the accuracy of the results.
As can be clearly seen from Figures 2B–C, CVAOA, CVAA, and
AA were still resolved from other arsenic standards in the chro-
matograms obtained by using 100-mm or 50-mm 3-µm par-
ticle columns. To decrease the CVAOA and CVAA analysis time,
the 50 × 4.6-mm column with 3-µm particle size was em-
ployed. The analysis time decreased to 4 min as shown in
Figure 2C. The chloride peak was still adequately resolved
from CVAA, CVAOA, and AA peaks in chromatograms obtained
using all column lengths. Urine contains enough chloride to
potentially create a polyatomic isobaric interference by re-
acting with argon to produce ArCl+(m/z 75). Although chloride
clearly eluted prior to CVAA and CVAOA,ArCl+(m/z 75) was not
formed to a degree that was significant relative to the arsenic
concentrations in standards that had been spiked into human

Page 4

Journal of AnalyticalToxicology,Vol. 34, April 2010
4
urine nor relative to arsenicals present in pre-exposed rat urine
samples and to only a small degree in urine from rats exposed
to Lewisite. This was apparent by comparing the35Cl+traces to
the m/z 75 traces.
It is probable that the use of platinum tipped cones, high RF
power,optimizationofnebulizergasforlowoxides,and2%iso-
propanol in the mobile phase (33), all of which contribute to a
hotterplasmaatthecone,resultedinverylowArCl+ionforma-
tion. Ferguson and Houk describe a plasma condition in which
fewerpolyatomicionsoccurinthemassspectrumasaresultof
the kinetic gas temperatures (Tgas) value. These values corre-
spond to the measured dissociation constant (Kd) for a given
polyatomic ion that is higher than the expected 5500–6000 K
(34), such as may be the case when the plasma is optimized as
describedhere.Therefore,theneedforacellgasornitrogencar-
riergas(35)toeliminatetheArCl+ionwasobviated,andtheDy-
namicReactionCellTM(DRC),whichistypicallyusedtoremove
isobaric interferences, was operated in evacuated mode.
Twenty calibration curves, over the range of 9.09–2895 µg/L,
were run over a 6-week period. Correlation coefficients were at
a minimum > 0.99 and typically were > 0.999. Anticipating the
need for expanded analytical capacity in response to a large-
scale emergency, we must achieve acceptable reproducibility
and accuracy of results following transfer of the method be-
tween instruments and analysts. The method was transferred
to a different HPLC–ICP-MS for an inter-instrument and inter-
analyst comparison. A regression plot of the calibration stan-
dards and quality controls resulted in a slope of 1.03 and a cor-
relation coefficient of 0.9997, which indicates that comparable
results can be obtained with separate instruments and dif-
ferent analysts.
T-test of results from the comparison of instrument re-
sponse and calibration of five CVAA standard spikes in base
urine to five spikes of CVAOA on two different days demon-
strated no statistically significant difference in instrument re-
sponse on the basis of arsenic (p = 0.15, p = 0.68, respectively).
The MDL was determined to be 1.3 µg/L as CVAA (1.4 µg/L
CVAOA) on the basis of compound mass at the dilution de-
scribed. The LOQ was calculated as 4.2 µg/L as CVAA (4.6 µg/L
CVAOA) (32). The MDL and LOQ are of the order of concen-
trations of compounds reported on the basis of arsenic typically
determined in the U.S. population as a result of exposure from
food, water, and other environmental sources (36). Therefore,
the method is sufficiently sensitive for determination of CVAA
or CVAOA arsenic in urine at concentrations similar to those of
species that are found as a result of common environmental ar-
senic exposure.
As part of the validation process, quality control samples
were analyzed in twenty independent runs over a six-week
period. Quality control was evaluated according to National
Center for Environmental Health, Division of Laboratory Sci-
ences quality control policy, and was consistent with the West-
gard criteria(37,38).Qualitycontrolwasfoundto beacceptable
at all three levels as exemplified in the plot of data from the
lowest concentration QC in Figure 3. The average low,
medium, and high QC results (70.2, 1104, and 1841 µg/L, re-
spectively) were in good agreement with the calculated con-
centrations of 72.3, 1106, and 1820 µg/L. In addition, relative
standarddeviationforthe lowQC samplewas less than 6%,and
the medium and high QC samples were less than 3%.
Twenty urine samples from unexposed individuals were
Figure 2. Separation of standards spiked into human urine using a (A) 150
× 4.6 mm, 3 µm (B) 100 × 4.6 mm, 3 µm and (C) 50 × 4.6 mm, 3 µm RP-
Amide column.The 35Cl scan was inncludeed to demonstrate the elution
of chloride without formation of any interference resulting from the for-
mation of (40Ar35Cl)+. The identities of the peaks are AsB, As(III), DMA,
MMA, As(V), CVAOA, CVAA, and AA.
Figure 3. Plot of the low urine chlorovinylarsonous acid quality control
analyses: 70.2 ± 4.0 µg/L mean ± standard deviation vs. 72.3 µg/L ex-
pected. Twenty individual runs of standards and quality control samples
were obtained over a six-week period to demonstrate the long-term previ-
sion and accuracy of the method.

Page 5

Journal of AnalyticalToxicology,Vol. 34, April 2010
5
spiked with CVAA, which resulted in concentration levels
ranging from 10 to 900 µg/L. Each concentration level was
spiked in five individual urines, and the mean, accuracy, and
reproducibility were determined as shown in Table I. The
average recovery, a measure of the accuracy of the analysis,
rangedfrom96to101%.Therelativestan-
dard deviation values for the four concen-
trations were less than 6%. The accuracy
andreproducibilityofthespikedindividual
urine samples demonstrate the applica-
bilityofthismethodacrossadiverserange
of urine matrices and concentrations.
One hundred urine samples from indi-
viduals with no known exposure to
Lewisitewereanalyzedtoestablishaback-
ground reference range in the unexposed
population and to ascertain whether any
false-positives would result due to variation in the chloride
concentrations or environmental and dietary arsenic that
might be passed from one person to another. There were no de-
tectable levels of CVAA or CVAOA in urine from individuals
without any known exposure to Lewisite.
To determine method applicability to analysis of urine ex-
creted after exposure to Lewisite, urine samples collected from
ratsexposedina2004study(21)wereanalyzed.Chromatograms
resultingfromanalysisofurinefromaratimmediatelypriorto
exposureandat19–20hafterexposuretothelowLewisitedose
ona50×4.6-mminternaldiamtercolumnareshowninFigure
4A–4B.NoAspeakselutedinthevicinityofCVAAorCVAOAre-
tentiontimesduringtheanalysisofurineobtainedfromratsim-
mediately before Lewisite exposure (Figure 4A) or from control
rat urine collected at any time point. The As(III) peak is visible
inFigure4A,butthesmallAs(V)peakhadinsufficientintensity
forquantification,whichindicatesstabilityofAs(III)tooxidation
overthe1-hurinecollectiontime,whilethawedforanalysisand
frozenfor5years.TheDMAandtheAApeaksaretheonlyother
peaksvisibleabovebaselineinthechromatogramfromraturine
obtained immediately prior to exposure.
The chromatogram from analysis of a rat urine sample col-
lected 19–20 h after the low-dose Lewisite exposure (the time
at which concentration of total arsenic and CVAA excreted had
reached their zeniths) (21) in Figure 4B is markedly different
from the rat urine collected before exposure. A CVAA concen-
tration of 759 µg/L was determined from the analysis of the
urine collected during this time period, which indicates sig-
nificant transdermal and systemic Lewisite absorption. The
concentration of CVAA decreased to 314 µg/L in urine col-
lected between 23 and 24 h after exposure.
Most of the urine samples from exposed rats that were ana-
lyzed contained CVAOA with little or no CVAA remaining,
though inorganic As(III) remained unoxidized (Figure 4C).
With one exception, these samples also had less than 1-mL
total volume, which led to a high ratio of surface area exposed
to air for the volume. The analysis of urine from a rat that re-
ceived a high Lewisite dose on a 150 × 4.6-mm internal diam-
eter column is shown in Figure 4C. Kinoshita et al. (19) re-
ported that CVAOA but almost no CVAA remained in a 24-h
collection of urine from a mouse exposed to CVAA. The CVAA
and CVAOA metabolites observed in chromatograms from both
low and high dose rat urine samples in Figure 4B–4C were,
therefore,expectedandconfirmtheapplicabilityofthemethod
to measurement of these metabolites as a result of Lewisite ex-
posure. It was not clear whether the CVAOA occurred as a
Table I. Results of Spiked Urine Experiments
2-Chlorovinylarsonous
Acid Concentration
(µg/L)
Number of
Replicates
Mean Result
(µg/L)
Accuracy
(%)
Reproducibility
(%)
10
35
50
5
5
5
5
10.3
34.6
51.0
934.8
97.4
101.1
98.1
96.3
5.3
2.5
3.1
2.7900
Figure 4. Rat urine analysis at (A) time 0 and (B) 19–20 h after percutaneous
exposure to 2.0 mg/kg of Lewisite (low dose) performed on a 50 × 4.6 mm
i.d. column.The chromatogram from analysis of rat urine collected 19–20
h after percutaneous exposure to 10.0 mg/kg of Lewisite (high dose) (C) on
a 150 × 4.6 mm i.d. column shows that chlorovinylarsonous acid (CVAA)
has oxidized to CVAOA. The identities of other peaks are As(III). DMA,
MMA, As(V), ArCl+, and AA.

Page 6

Journal of AnalyticalToxicology,Vol. 34, April 2010
6
6
metabolite, as Kinoshita et al. (19) considered possible, or if it
was generated during the period of collection and analysis in
which the mouse or rat urine was thawed. However, these find-
ingsindicatethatthevinylarsenicinCVAAismuchmorelabile
to oxidation than inorganicAs(III). Therefore, the ability to de-
termine the oxidized metabolite is very important to Lewisite
exposure assessment.
CVAOA concentrations were 1634 µg/L CVAOA (equivalent to
1494 µg/L on the basis mass of CVAA on a molar basis) in rat
urine collected between 19 and 20 h after the high dose
Lewisite exposure. The urine concentration then decreased to
the equivalent of 1096 µg/L CVAA in urine (1199 µg/L CVAOA)
collected between 23 and 24 h after exposure. These data il-
lustrate the dose-dependent increase in combined CVAOA and
CVAA excretion after exposure to Lewisite. The data also
demonstrate the importance of analyzing both CVAA and
CVAOA as a measure of Lewisite exposure, especially consid-
ering the ambiguity in the literature on the extent of suc-
cessful derivatization of CVAOA with dithiols for gas chro-
matography (39).
It is beyond the scope of this study to fully explain whether
the urine chloride concentration increased sufficiently to cause
the appearance of the small ArCl+peak in Figure 4B–4C as a re-
sult of Lewisite and AsCl3hydrolysis alone. However, under the
instrumental conditions described, the ArCl+peak was small
and resolved from CVAOA and CVAA, so ArCl+was not a factor
that could compromise the CVAA or CVAOA results even when
chloride excretion was sufficiently elevated to cause detectable
ArCl+formation.
Kinoshita et al. (19) noted several major unknown and un-
resolved organoarsenic metabolites in mouse urine after expo-
suretoCVAA.Inthisstudy,weanalyzedurinesamplesfromrats
that had been exposed to Lewisite five years earlier, collected,
and stored frozen. We observed large increases in peak area for
the metabolites that eluted in less than 1.0 min (Figure 4B)
comparedtopre-exposureurinesamples.Determiningwhether
alloftherelativepeakareasofthearseniccontainingspeciesare
a function of active metabolism of Lewisite or due to other
causes is beyond the scope of this study. However, the peak
areasofAs(III)andDMAincreased867%and862%,respectively
(Figure 4B), in urine of the rat that received the low Lewisite
dose and 256% and 381%, respectively, in urine of the rat that
receivedthehighLewisitedose(Figure4C).Becausethesepar-
ticular metabolites have been described as stable while frozen
(29, 30) and no oxidation of As(III) to As(V) was observed, it is
highly unlikely that these metabolites were artifactual.
In both chromatograms, an additional major unidentified
metabolite was apparent. The unidentified metabolite that
eluted at 1.94 min from the 150 mm column (Figure 4C) had
peak intensity intermediate between As(III) and DMA in the rat
given a low Lewisite dose. Based on arsenic counts/s, in the
urine of the rat exposed to the high Lewisite dose (Figure 4C),
the unidentified peak had more than five times the intensity of
the DMA peak. Taken together, these could indicate possible
Lewisite dose-dependent metabolic or excretion differences
for some metabolites. Oxidation of our standards with hy-
drogen peroxide did not result in a peak that coeluted with the
major unidentified metabolite; nor did monitoring34S and
78Se along with the mass 75 chromatogram identify it as a
sulfur or selenium conjugate. Further investigation into the
identity of this metabolite is warranted, such as the possibility
of methylation of the chlorovinyl-arsenic atom because some
organoarsines have been found to be methylated microbially in
soil (40). The identification of all of the metabolites of Lewisite
is beyond the scope of this study, but those determined here in-
dicate that depending on dose, not all of the metabolic arsenic
from Lewisite exposure was reflected in the principal CVAA
and CVAOA metabolites. However, As(III) and DMA excretion
did not reflect the metabolic dose-dependence characteristic of
CVAA and CVAOA demonstrated in this study. Further, though
excretion of the unidentified metabolite increased withdose,its
relative intensity was not as high in urine from the rat ad-
ministered the low Lewisite dose as in the rat administered the
high Lewisite dose.
Without regard to the latter findings, the chromatographic
peaks for CVAA and CVAOA were well-resolved and, thus, still
quantifiableeveninthepresenceoftheseadditionalmetabolites
and with the ArCl+peak apparent only in urine from Lewisite-
exposed rats. Therefore, CVAA and CVAOA may be quantified
usingthismethodevenwhenincreasedconcentrationsofthese
additionalcharacterizedanduncharacterizedarsenicalmetabo-
lites are present. Though further study into dose-dependent
metabolism of organoarsenic compounds is needed to clarify
theseissues,forthepurposeofrapidlydeterminingLewisiteex-
posure,CVAAandCVAOAremainthespeciesofchoicefortheir
utilization as major biomarkers of Lewisite exposure.
Conclusions
A high through-put LC–-ICP-MS method for analysis of
CVAA, a major urine metabolite of Lewisite, and its oxidation
product, CVAOA, has been developed and is capable of 12 sam-
ples per h and 96 samples per 8-h workday through-put. The
data indicates that this method provides a rapid, robust, and
sufficiently sensitive means of to assess Lewisite exposure via
the CVAA and CVAOA metabolites in urine over a wide range of
concentrations. The method is simple, accurate, and repro-
ducible. Small volumes of urine (400 µL), simple sample prepa-
ration, and interinstrument reproducibility establish capability
for transferring the method to additional public health labo-
ratories for expanded capacity in the event of small- or large-
scale Lewisite exposures.
References
1. J.A. Nieuwland. Some Reactions of Acetylene, Doctoral
Thesis. University of Notre Dame, Notre Dame, IN, 1904.
2. J.A. Vilensky. Dew of Death: The Story of Lewisite, America’s
World War I Weapon of Mass Destruction, Indiana University
Press, Bloomington, IN, 2005, pp 23,108–118.
3. M. Goldman and J.C. Dacre. Lewisite: its chemistry, toxicology,
and biological effects. In Reviews of Environmental Contamination
and Toxicology,Volume 110, Springer-Verlag Berlin and Heidel-
berg GmbH & Co. K, Berlin, Germany, 1989, pp 75–115.
4. AmericanToxic Substances and Diseases Registry, Medical Man-
agement Guidelines for Blister Agents: Lewisite (L) (C2H2AsCl3)

Page 7

Journal of AnalyticalToxicology,Vol. 34, April 2010
77
Mustard-Lewisite Mixture (HL). http://www.atsdr.cdc.gov/MHMI/
mmg163.pdf, pp 1–20.
5. J.V. Wooten, D.L. Ashley, and A.M. Calafat. Quantitation of
2-chlorovinylarsonous acid in human urine by automated solid-
phase microextraction-gas chromatography–mass spectrometry.
J. Chromatogr. B 772: 147–153 (2002).
6. J.A. Romano, B.J. Lukey, and Harry Salem. Chemical Warfare
Agents: Chemistry, Pharmacology, Toxicology, and Therapeutics,
CRC Press, Boca Raton, FL, 2008, p 306.
7. F.R. Sidell, J.S. Urbanetti,W.J. Smith, and C.G. Hurst, Textbook of
Military Medicine, Office of the Surgeon General atTMM Publi-
cations, Borden Institute, Washington D.C., 1997, p 219.
8. N. Gussman. Lewisite: never-used chemical weapon becomes
hazardous waste. Chem. Eng. Prog. 101: 64 (2005).
9. A. Fidder, D. Noort, A.G. Hulst, L.P.A. de Jong, and H.P. Ben-
schop. Biomonitoring of exposure to lewisite based on adducts to
haemoglobin. Arch. Toxicol. 74: 207–214 (2000).
10. J. Perera. Lewisite—new gas weapon in Gulf War.New Sci. 105:
8–8 (1985).
11. J.J. Urban and R.L. vonTersch. Conformational analysis of the iso-
mers of Lewisite. J. Phys. Org. Chem. 12: 95–102 (1999).
12. J.R. Smith,T.P. Logan, and L.L. Szafraniec. Spectroscopic charac-
terization of the geminal isomer of Lewisite. Anal.Lett. 28: 1541–
1554 (1995).
13. T.P. Logan, J.R. Smith, E.M. Jakubowski, and R.E. Nielson.Verifi-
cation of lewisite exposure by the analysis of 2-chlorovinyl ar-
sonous acid in urine. Toxicol. Meth. 9: 275–284 (1999).
14. S.L. Bartelt-Hunt, D.R.U. Knappe, and M.A. Barlaz. A review of
chemical warfare agent simulants for the study of environmental
behavior. Criti. Rev. Env. Sci. Tech. 38: 112–136 (2008).
15. B.A.Tomkins, G.A. Sega, and C.-H. Ho. Determination of lewisite
oxide in soil using solid-phase microextraction followed by gas
chromatography with flame photometric or mass spectrometric
detection. J. Chromatogr.A 909: 13-28 (2001).
16. N.B. Munro, S.S.Talmage, G.D. Griffin, L.C.Waters,A.P.Watson,
J.F. King, and V. Hauschild. The sources, fate, and toxicity of
chemical warfare agent degradation products. Environ. Health.
Persp. 107: 933–974 (1999).
17. W.K. Fowler, D.C. Stewart, and D.S. Weinberg. Gas chromato-
graphic determination of the lewisite hydrolysate, 2-chloroviny-
larsonous acid, after derivatization with 1,2-ethanedithiol. J.Chro-
matogr. A 558: 235–246 (1991).
18. T. Wada, E. Nagasawa, and S. Hanaoka. Simultaneous determi-
nation of degradation products related to chemical warfare agents
by high-performance liquid chromatography/mass spectrometry.
Appl. Organomet. Chem. 20: 573–579 (2006).
19. K. Kinoshita, O. Shikino,Y. Seto, and T. Kaise. Determination of
degradation compounds derived from Lewisite by high perfor-
mance liquid chromatography/inductively coupled plasma-mass
spectrometry. Appl. Organomet. Chem. 20: 591–596 (2006).
20. G.R. Cameron, H.M. Carleton, R.H.D. Short. Pathological changes
induced by lewisite and allied compounds. J.Pathol.Bateriol. 58:
411–422 (1948).
21. J.V. Wooten, S.B. Stanfill, P.J. Sabourn, D.L. Ashley, and J.R. Barr.
Quantification of 2-chlorovinylarsonous acid (CVAA) in urine
from male Wistar rats percutaneously exposed to Lewisite.
Manuscript in progress.
22. W.J. McShane, R.S. Pappas, and D. Paschal. Analysis of total ar-
senic, total selenium and total chromium in urine by inductively
coupled plasma-dynamic reaction cell-mass spectrometry. J.Anal.
Atom. Spectrom. 22: 630–635 (2007).
23. J.M. Jarrett, R.L. Jones, K.L. Caldwell, and C.P.Verdon.Total urine
arsenic measurements using inductively coupled plasma mass
spectrometry with a dynamic reaction cell. Atom.Spectrosc. 28:
113–122 (2007).
24. J.R. Garbarino, A.J. Bednar, and M.R. Burkhardt. Methods of
Analysis by the U.S. Geological Survey, National Water Quality
Laboratory—Arsenic Speciation in Natural-Water Samples Using
Laboratory and Field Methods, Water-Resources Investigations
Report 02-4144, U.S. Geological Survey, Denver, CO, 2002, pp
1–47.
25. L.S. Milstein, A. Essader, E.D. Pellizzari, R.A. Fernando, J.H.
Raymer, K.E. Levine, and O. Akinbo. Development and applica-
tion of a robust speciation method for determination of six arsenic
compounds present in human urine. Environ.Health Persp. 111:
293–296 (2003).
26. T.I. Todorov, J.W. Ejnik, F.G. Mullick, and J.A. Centeno. Arsenic
speciation in urine and blood reference materials. Michrochim.
Acta 151: 263–268 (2005).
27. C.P. Verdon, K.L. Caldwell, M.R. Fresquez, and R.L. Jones. De-
termination of seven arsenic compounds in urine by HPLC-ICP-
DRC-MS: a CDC population biomonitoring method. Anal.
Bioanal. Chem. 393: 939–947 (2009).
28. B. Szostek and J.H. Aldstadt. Determination of organoarsenicals
in the environment by solid-phase microextraction gas chro-
matography mass spectrometry. J. Chromatogr. A 807: 253–263
(1998).
29. Y.C. Chen, C.J. Amarasiriwardena,Y.M. Hsueh, and D.C. Chris-
tiani. Stability of arsenic species and insoluble arsenic in human
urine. Canc. Epi. Bio. Prev. 11: 1427–1433 (2002).
30. J. Feldmann,V.W.M. Lai,W.R. Cullen, M.S. Ma, X.F. Lu, and X.C.
Le. Sample preparation and storage can change arsenic speciation
in human urine. Clin. Chem. 45: 1988–1997 (1999).
31. L.H. Keith, W. Crummett, J. Deegan, J.K. Taylor, R.A. Libby, and
G.Wentler. Principles of environmental-analysis. Anal.Chem.55:
2210–2218 (1983).
32. J.K.Taylor, Quality Assurance of Chemical Measurements, Lewis,
Boca Raton, LA, NewYork, NY,Washington, D.C., USA; London,
England, 1987, pp. 79–81.
33. E.H. Larsen and S. Stürup. Carbon-enhanced inductively cou-
pled plasma mass spectrometric detection of arsenic and selenium
and its application to arsenic speciation. J. Anal. Atom. Spec-
trom. 9: 1099–1105 (1994).
34. J.W. Ferguson and R.S. Houk. High resolution studies of the ori-
gins of polyatomic ions in inductively coupled plasma-mass spec-
trometry, Part I. Identification methods and effects of neutral gas
density assumptions, extraction voltage, and cone material.
Spectro.Acta B 61: 905–915 (2006).
35. S. Branch, L. Ebdon, M. Ford, M. Foulkes, and P. O’neill. Deter-
mination of arsenic in samples with high chloride content by in-
ductively coupled plasma mass spectrometry. J.Anal.Atom.Spec-
trom. 6: 151–154 (1991).
36. K.L. Caldwell, R.L. Jones, C.P.Verdon, J.M. Jarrett, S.P. Caudill, and
J.D. Osterloh. Levels of urinary total and speciated arsenic in the
US population: National Health and Nutrition Examination Survey
2003–2004. J. Expo. Sci. Env. Epid. 19: 59–68 (2009).
37. S.P. Caudill, R.L. Schleicher, and J.L. Pirkle. Multi-rule quality con-
trol for the age-related eye disease study. Stat. Med. 27: 4094–
4106 (2008).
38. J.O. Westgard, P.L. Barry, M.R. Hunt, and T. Groth. A multi-rule
Shewhart chart for quality control in clinical chemistry. Clin.
Chem. 27: 493–501 (1981).
39. M. Harper. The use of thermal desorption in monitoring for the
chemical weapons demilitarization program. J.Environ.Monitor.
4: 688–694 (2002).
40. F.-A. Pitten, G. Müller, P. König, D. Schmidt, K. Thurow, and
A. Kramer. Risk assessment of a former military base contaminated
with organoarsenic-based warfare agents: uptake of arsenic by
terrestrial plants. Sci. Tot. Environ. 226: 237–245 (1999).
Manuscript received September 1, 2009;
revision received October 22, 2009.