Titier.
qxd:JATLynneTemplate 6/13/11 10:56 AM Page 1
Journal of Analytical Toxicology, Vol. 35, July/August 2011
Suicide with Cisatracurium and Thiopental:
Forensic and Analytical Aspects
Nadège Castaing1,2, Larbi Benali2,3, Dominique Ducint1,2, Mathieu Molimard1,2, Sophie Gromb2,3,
and Karine Titier1,2,*
1Laboratoire de Pharmacologie et Toxicologie, Hopital Pellegrin, Bordeaux, France; 2Université de Bordeaux, Bordeaux, France;
and 3Service de Médecine Légale, Hopital Pellegrin, Bordeaux, France
Abstract
The suicide of a 43-year-old male by intravenous injection of
cisatracurium, a non-depolarizing neuromuscular blocking agent,
and thiopental, an ultra-short-acting barbiturate, is presented.
Systematic toxicological screening by gas chromatography–mass
spectrometry (GC–MS), liquid chromatography (LC)–diode-array
detection, and LC–MS–MS confirmed the presence of thiopental.
A large peak in the GC–MS chromatogram was matched by the
Pfleger-Maurer library as corlumine, but neither atracurium
neither its metabolite, laudanosine, were detected. To confirm the
absence or the presence of laudanosine in the blood sample, an
ultra-performance liquid chromatography–MS–MS method for
cisatracurium and laudanosine quantification was developed.
The calibration range was 2.5–500 ng/mL for laudanosine and 10–
500 ng/mL for cisatracurium. The biases were lower than 12.3%.
Intraday and interday precisions, expressed as coefficient of
variation, were lower than 13.3%. This method allowed to confirm
the presence of laudanosine and measurement of laudanosine
in all samples. The femoral blood concentration was therapeutic
(0.46 μg/mL). This case report documents a possible analytical
pitfall and describes a simple and fast method for cisatracurium
determination. Moreover, the purpose of this case report was to
document the postmortem redistribution of cisatracurium and
laudanosine, which could help make it possible to interpret tissue
or cardiac blood concentrations in forensic cases where femoral
blood is not available.
Introduction
Cisatracurium is 1 of 10 stereoisomers of atracurium, and it
has about 3 times the potency of the racemic mixture. It is a
synthetic quaterny ammonium derivative employed as a non-
* Author to whom correspondence should be addressed. Laboratoire de Pharmacologie et
Toxicologie, Hôpital Pellegrin, Place Amélie Raba-Léon, 33076 Bordeaux cedex, France.
Email: karine.titier@chu-bordeaux.fr.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Case Report
depolarizing neuromuscular blocking agent to facilitate endo-
tracheal intubation and to provide skeletal muscle relaxation
during surgery or mechanical ventilation. It is supplied as a be-
sylate salt in 2–20 mg/L solutions for intravenous injection.
Adult doses are usually 0.15–0.20 mg/kg followed by 0.03
mg/kg maintenance dose if needed during longer procedures.
Under physiological conditions, atracurium and
cisatracurium are very quickly hydrolyzed by Hoffmann’s re-
action, a non-enzymatic breakdown process occurring at phys-
iological pH and temperature to give laudanosine and an acry-
late moiety. None of the metabolites has significant
pharmacological activity. Cisatracurium is excreted in urine
mostly as metabolites, and the elimination half-life is about 30
min. Adverse effects of cisatracurium include itching,
wheezing, hives, bronchospasm, laryngospasm, prolonged neu-
romuscular block, hypotension, respiratory failure, and death.
Few fatal intoxications following non-depolarizing neuromus-
cular blocking agent administration have been reported in the
literature (1–4). Cisatracurium was involved in only one case
(3), but the forensic data were poor.
A recent case involving the suicide of an anesthesiologist is
presented here. In this case, corlumine and not cisatracurium
or laudanosine was detected by systematic toxicological anal-
ysis, which could have falsely indicated the absence of
cisatracurium administration. The aims of this case report
were to document a possible analytical pitfall, to describe a fast
and simple ultra-performance liquid chromatography–tandem
mass spectrometry (UPLC–MS–MS) method for cisatracurium
and laudasonine quantification, and to document the post-
mortem redistribution of these compounds.
Case History
A 43-year-old man was found dead in a hotel room by a
cleaner. The corpse was fully dressed, in a bath with a pillow
under his head. Next to the body in the bathroom were found
375
Titier.qxd:JATLynneTemplate 6/13/11 10:56 AM Page 2
used intravenous equipment (gauzes, intravenous cannulas,
rubbing alcohol, tapes, syringes, giving sets, intravenous in-
fusion lines), 3 broken vials of 2 mg/mL Nimbex ®
(cisatracurium), 2 empty vials of 1 mg thiopental, and Emla®
(lidocaine) patches usually used as a local skin anesthetic to
minimize pain prior to intravenous access. The deceased was
found to be an anesthesiologist.
The autopsy was carried out at the Medico-Legal Institute of
the Bordeaux Teaching Hospitals. Autopsy findings were prick
on the back of the hand, bruised on the edges, diffuse conges-
tion of the internal organs, and intense pulmonary edema.
Blood (cardiac and peripheral), urine, vitreous humor, gas-
tric contents, bile, and tissues (heart, liver, kidney, brain, lung)
were collected at autopsy and stored at –20°C. A systematic tox-
icological screening was performed. Immunoassay screening
for methadone, cocaine, opiates, amphetamines, cannabinoids,
benzodiazepines, tricyclic antidepressants, and barbiturates
was performed on blood and urine samples using CEDIA
reagents (Microgenics). Buprenorphine determination was
performed on the urine sample using ELISA (Microgenics).
Toxicological screening by gas chromatography (GC)–MS,
liquid chromatography–diode-array detection (LC–DAD), and
liquid chromatography (LC)–tandem MS was carried out on
blood, urine, and gastric content samples, after acidic and
neutral liquid–liquid extraction. Ethanol and others volatiles
(methanol, acetone, isopropanol) were searched in peripheral
blood and vitreous humor by headspace GC–flame-ionization
detection (FID). Specific determinations of psychoactive drugs
(benzodiazepine, antidepressants, antipsychotics, barbiturates)
and narcotics (cannabinoids, cocaine, opiates, amphetamines)
by LC–MS–MS and GC–MS were carried out on peripheral
blood and urine. Carboxyhemoglobin determination was per-
formed using spectrophotometry on blood.
Experimental
Chemicals
Cisatracurium was obtained from Sanofi Synthelabo (Paris,
France). Laudanosine was kindly donated by Dr. Kintz (Labo-
ratoire ChemTox, Illkirch, France), and methylrisperidone,
used as the internal standard (I.S.), was kindly donated by
Table I. MRM Transitions for the Detection of Laudanosine, Cisatracurium, and Internal Standard by UPLC–MS–MS
Retention Time Parent Ion
Compound (min) (m/z)
Laudanosine 1.71 358.2
Cisatracurium 2.65 464.5
Methylrisperidone (I.S.) 2.35 422.0
* The transition ion for quantification is underlined.
376
Journal of Analytical Toxicology, Vol. 35, July/August 2011
Lundbeck (Copenhagen, Denmark). HPLC-grade solvents were
purchased from Prolabo (Paris, France). Drug-free blood from
healthy volunteers was provided by the local branch of the na-
tional blood service (Etablissement Français du Sang, Bor-
deaux, France).
Sample preparation
Stock solutions were prepared in methanol for laudanosine
(1 mg/mL) and methylrisperidone (1 mg/mL) and in acetoni-
trile for cisatracurium (1 mg/mL). These stock solutions were
stored at –20°C in the dark, conditions under which they were
stable for at least three months. Working solutions of the mix-
ture of both analytes at 10 µg/mL were freshly prepared by di-
lution with 0.05 N HCl. A working solution of the internal
standard at 10 µg/mL was also freshly prepared by dilution
with methanol. Routine daily calibration curves and controls
were prepared for each analytical batch in drug-free whole
blood. Calibration standards were made to yield concentra-
tions of 2.5, 5, 10, 30, 50, 100, 250, and 500 ng/mL. For the
preparation of quality controls (QC) used for the validation of
the assay, an independent stock solution was prepared and
further diluted to achieve concentrations of 2.5, 80, and 400
ng/mL for laudanosine and 10, 80, and 400 ng/mL for
atracurium. Tissue samples were homogenized in isotonic
buffer (1:2, w/v) using a Precellys 24 homogenizer (Bertin
Technologies, France). Final quantification of each sample was
based on an appropriate dilution in deionized water if neces-
sary.
QC, calibration curve, blank blood samples, and unknown
samples were extracted using a liquid–liquid extraction. Whole
blood (250 μL) was fortified with 20 µL of internal standard (10
µg/mL) and 20 µL of HClO4 (0.5 M). Five milliliters of ethy-
lacetate was added to the mixture, which was then shaken for
10 min prior to being centrifuged at 5000 rpm for 10 min. The
organic layer was decanted into another tube, where it was
evaporated to complete dryness under a nitrogen stream at
room temperature. Samples were reconstituted with 100 µL of
0.05 N HCl/ACN (50:50, v/v), then vortex mixed for 10 s, and 1
µL was injected into UPLC–MS–MS system.
Instrumentation
The UPLC unit consisted of an Acquity UPLC® separation
module (Waters, Milford, MA) controlled by Masslynx® soft-
Daughter Ions* Relative Cone Collision Energy Dwell Time
(m/z) Intensity (V) (eV) (ms)
206.0 2.1 25 15 50
151.0 30 50
189.0 35 20 50
307.2 18 50
201.2 35 35 50
Titier.qxd:JATLynneTemplate 6/13/11 10:56 AM Page 3

Journal of Analytical Toxicology, Vol. 35, July/August 2011

ware. Separations were performed at 50°C on an Acquity UPLC mean of each set of the concentrations, the standard deviation
BHE C18 column (50 × 2.1 mm, 1.7 µm, Waters). The mobile (SD), and coefficient of variation (CV) were determined. The
phase consisted of a gradient of acetonitrile/ammonium for- precision expressed as CV should be less than 10% for each QC
mate buffer (4 mmol/L, pH 3.2). Acetonitrile was set at 15% for sample except for the lowest (less than 20%). Bias was mea-
0.5 min. Then acetonitrile was increased linearly for 15% to sured as the percentage of difference from theoretical concen-
45% over 2.5 min and then set at 45% for 0.5 min. The system tration. The bias should be less than 10% for all QC except for
was then set to the initial conditions that were achieved in 10 the lowest (less than 20%). The limit of quantification (LOQ)
seconds and was re-equilibrated during 0.9 min before the was determined as the lowest concentration of a given drug
next injection. The total run time of the analysis was 4.5 min giving a response that could be quantified with both interassay
at a flow rate of 0.37 mL/min.
An Acquity TQD® detector (Waters) with electrospray ion-
ization (ESI) in positive ion mode was used for detection. The
MS was operated in the multiple-reaction monitoring (MRM)
mode. The ESI source was operated at a temperature of 120°C.
The desolvation temperature was set at 350°C. The cone gas
flow was set at 50 L/h and the desolvation gas flow at 650 L/h.
The capillary voltage was 3.0 kV. The cone voltages were opti-
mized for each compound. The MS collision gas was argon at
3.2 × 10–3 mbar, and the collision energies were optimized for
each compound. For the quantitative determination, two mass-
to-charge ratios corresponding to the quantification and con-
firmation ions were monitored for each compound (Table I).

Validation of the method

The recoveries were evaluated at two levels by comparing the
peak areas of the extracted samples (n = 5) with those of un-
extracted reference standards prepared at the same concen-
tration. For the linearity study, four calibration curves were an-
alyzed. A weighted least-squares linear regression model was
used to calculate the equation relating the peak-area ratio
(drug vs. internal standard) and the concentration of each
compound in the calibrators. The inverse of the concentration
(1/x) was used as weighting factor. To determine the adequate
fit linear model, the difference between the observed values and
the fitted y values expressed as coefficient of variation (CV) was
examined for each standard concentration. The CV should be
less than 10% for each plot of all calibration curves. The pre-
cision of the developed method was determined by analysis
three concentrations of quality control over four days. Intraday
variation of the assay was assessed by injecting five quality Figure 1. Representative chromatogram of blood sample spiked with 50
controls at each concentration on the same day. Interday vari- ng/mL of laudanosine and cisatracurium (TIC: total ion current, I.S.: internal
ation was assessed by injecting a further five samples of each standard).
concentration on three subsequent days. Subsequently, the

Table II. Intraday and Interday Precision and Accuracy Data

Intraday Interday
Theoretical concentration Observed concentration CV Observed concentration CV Bias
Compound (ng/mL) (mean ± SD, n = 5) (%) (mean ± SD, n = 20) (%) (%)

Laudanosine 2.5 2.6 ± 0.2 8.6 2.8 ± 0.03 10.2 11.2

80 74.0 ± 5.3 7.1 77.1 ± 5.9 7.7 –3.6
400 416.0 ± 29.9 7.2 406.0 ± 24.7 6.1 1.5
Cisatracurium 10 8.8 ± 1.1 13.0 8.8 ± 1.2 13.3 –12.3
80 76.9 ± 4.5 5.9 75.7 ± 6.5 8.5 –5.4
400 390.0 ± 8.5 2.2 390.8 ± 13.9 3.6 –2.3

377
Titier.qxd:JATLynneTemplate 6/13/11 10:56 AM Page 4

Journal of Analytical Toxicology, Vol. 35, July/August 2011

Table III. Extraction Recovery and Matrix Effect

Extraction Recovery Matrix Effect: Relative Recovery

Nominal Concentration Mean ± SD CV Mean ± SD CV
Compound (ng/mL) (n = 5) (%) (n = 6) (%)

Laudanosine 80 73.4 ± 1.9 2.5 116.3 ± 10.4 8.9

400 71.8 ± 1.8 2.5 108.1 ± 4.1 3.8
Methylrisperidone (I.S.) 52.4 ± 2.4 4.6 99.5 ± 2.4 2.4
Cisatracurium 80 41.3 ± 1.6 3.9 97.8 ± 4.4 4.6
400 44.8 ± 1.9 4.3 100.2 ± 2.8 2.8

variation and bias lower than 20%. The

lack of interferences from blood endoge-
nous compounds was examined by ana-
lyzing different blank human blood ex-
tracts without internal standard. Six
different blank human blood samples
from hospitalized patients were tested to
check for the lack of interferences with
each drug and the internal standard. The
lack of interference from each drug and
internal standard with the MRM channels
of the other drugs was checked by in-
jecting pure solutions of each drug sepa-
rately and searching for the total absence
of signal in the MRM channels of the
other drugs. Ion suppression was investi-
gated through the addition of the ana-
lytes to blank matrix extracts. If the re-
sponse of the analytes is compared to an
unextracted standard solution at the same
nominal concentration, any difference
from 100% recovery could be attributed
to matrix effect. Two concentrations were
analyzed in replicates; six different human
blood samples have been tested and ana-
lyzed in different batches. To be accept-
able, relative recoveries should be be-
tween 80 and 120%.

Results and Discussion

Validation
The retention times of laudanosine,
I.S., and cisatracurium were 1.73, 2.35,
and 2.65 min, respectively (Figure 1).
Under the chromatographic conditions
used, no interference was seen from en-
dogenous substances in drug-free human
Figure 2. GC–MS spectra of laudanosine and corlumine with their chemical structures: GC–MS spectrum
blood at the times of interest. In addition,
from the case report (A), GC–MS spectrum of corlumine from Wiley library (B), and GC–MS spectrum of there was no interference from the in-
laudanosine from laboratory library (C). ternal standard with the MRM channels of
the drugs. Atracurium is not differenti-

378
Titier.qxd:JATLynneTemplate 6/13/11 10:56 AM Page 5

Journal of Analytical Toxicology, Vol. 35, July/August 2011

ated from cisatracurium. Calibration curves were linear in the 73% for laudanosine (Table III). The ion suppression assay did
range of 2.5–500 ng/mL for laudanosine and 10–500 ng/mL for not show any significant reduction in the response of each
cisatracurium (r > 0.993) with a slope variation coefficient (n drug (Table III).
= 4) lower than 10%. The results obtained for precision and ac-
curacy are presented in Table II. The method showed an in- Case report
traday and interday precision (CV) of less than 13.3%. A bias of Toxicological screening by LC–DAD in blood, urine, and
less than 12.3% was achieved. The LOQ was 2.5 ng/mL for gastric content samples revealed the presence of thiopental
laudanosine and 10 ng/mL for cisatracurium. The extraction and its metabolite pentobarbital. Toxicological screening by
recoveries were around 41% for cisatracurium and around GC–MS using automatic comparison to the Wiley, Pfleger-
Maurer, and laboratory libraries resulted in the presence of
corlumine, pentobarbital and thiopental in blood, urine, and
gastric content samples. Corlumine is an alkaloid of the plant
Corydalis species (papaveracaea) that grow in the Himalayan
region. Alkaloid of this plant are reported to have a spasmolytic
activity. There are no data published on the toxicology of this
compound. As shown in Figure 2, corlumine and laudanosine
have a similar chemical structure and the same electron impact
mass spectra. A manually process was used to identify the large
peak in the GC–MS chromatogram previously identified as
corlumine. With this manual process, laudanosine was
matched. Automatic comparison procedures could have falsely
indicated the absence of cisatracurium injection. Knowledge of
this case may help to avoid such an analytical pitfall.
The method described allowed us to confirm the presence of
laudanosine and measurement of laudanosine in all samples.
Because of their different molecular weight, corlumine, lau-
danosine and cisatracurium have different MRM transition.
Consequently, there is no interference between these com-
pounds using LC–MS–MS. Figure 3 shows the chromatogram
obtained from urine sample. The concentrations found in fluid
and tissues are presented in Table IV. The laudanosine femoral
blood concentration was therapeutic (0.46 μg/mL). The ther-
apeutic range of laudanosine after an intubating dose of
cisatracurium is around 0.2–0.3 μg/mL (5). High levels of lau-
danosine were measured in the kidney and in the liver. In the
vitreous humor, the concentration of laudanosine was low,
Figure 3. Urine chromatogram of the case report (TIC: total ion current, I.S.: which is in accordance with the fatal cases reported by Kintz et
internal standard). The urine sample contained 3.9 μg/mL of laudanosine al. (2) and Martinez et al. (3). Therefore, vitreous humor could
and 0.4 μg/mL of cisatracurium.
not be used to estimate the blood concentration. The small
amount of laudanosine recovered from
gastric contents may be due to post-
Table IV. Toxicological Findings: Fluid (µg/mL) and Tissue (µg/g) Concentrations mortem redistribution. Quantification of
thiopental and pentobarbital, was carried
Sample Laudanosine Cisatracurium Pentobarbital Thiopental out using GC–MS. The concentrations
found in fluid and tissues are presented in
Peripheral blood 0.46 ND* 0.36 10.62 Table IV. Thiopental peripheral blood con-
Cardiac blood 0.84 ND 0.34 15.10 centration is therapeutic (10.62 μg/mL).
Urine 3.90 0.36 0.92 6.90 The postmortem tissue concentrations
Vitreous humor 0.09 ND 1.00 3.30 reported in the literature were 11–26
Heart 1.00 ND 7.53 11.51 μg/mL in blood, 13 μg/mL in urine, 8–70
Brain 0.49 ND 0.74 20.70 μg/mL in brain, 32–79 μg/mL in liver, and
Kidney 3.60 0.03 0.56 23.40 16–41 μg/mL kidney (6). The postmortem
Lung 1.20 0.05 0.85 16.00
concentrations found here were consis-
Liver 2.10 0.03 6.71 78.00
Bile 1.20 0.05 0.26 5.70
tent with these results.
Gastric content 0.08 ND – – In this case, thiopental and cistar-
curium concentrations measured in the
* ND, not detectable. peripheral blood were in the therapeutic
ranges. However, in the absence of me-

379
Titier.qxd:JATLynneTemplate 6/13/11 10:56 AM Page 6
chanically controlled ventilation, such concentrations are fatal.
Finally, the cause of death was intense pulmonary edema fol-
lowing thiopental and cisatracurium administration. The
manner of death was suicide.
Conclusions
The simple and rapid method presented here is useful for de-
termination of laudanosine and cisatracurium concentrations
in biological materials. This case confirms the high tropism of
laudanosine for the liver and the kidney.
References
1. V. Cirimele, M. Villain, G. Pepin, B. Ludes, and P. Kintz. Screening
procedure for eight quaternary nitrogen muscle relaxants in blood
380
Journal of Analytical Toxicology, Vol. 35, July/August 2011
by high-performance liquid chromatography–electrospray ion-
ization mass spectrometry. J. Chromatogr. B Analyt. Technol.
Biomed. Life Sci. 789(1): 107–113 (2003).
2. P. Kintz, A. Tracqui, and B. Ludes. The distribution of laudanosine
in tissues after death from atracurium injection. Int. J. Legal Med.
114(1-2): 93–95 (2000).
3. M.A. Martinez, S. Ballesteros, and E. Almarza. Anesthesiologist
suicide with atracurium. J. Anal. Toxicol. 30(2): 120–124 (2006).
4. H. Sayer, O. Quintela, P. Marquet, J.L. Dupuy, J.M. Gaulier, and
G. Lachatre. Identification and quantitation of six non-depolar-
izing neuromuscular blocking agents by LC–MS in biological
fluids. J. Anal. Toxicol. 28(2): 105–110 (2004).
5. V. Nigrovic and J.L. Fox. Atracurium decay and the formation of
laudanosine in humans. Anesthesiology 74(3): 446–454 (1991).
6. R.C. Baselt. Disposition of Toxic Drugs and Chemicals in Man,
8th ed. Biomedical Publications, Foster City, CA, 2008.
Manuscript received September 20, 2010;
revision received November 29, 2010.