Clinical Chemistry 52:7
1346 –1355 (2006)
Liquid Chromatography–Tandem Mass
Spectrometry for Detection of Low
Concentrations of 21 Benzodiazepines,
Metabolites, and Analogs in Urine:
Method with Forensic Applications
Oscar Quintela,1,2 François-Ludovic Sauvage,1 Fabienne Charvier,1
Jean-Michel Gaulier,1 Gérard Lachâtre,1 and Pierre Marquet1*
Background: Commonly used methods for detecting
benzodiazepines (BZPs) and BZP-like substances, such
as zolpidem and zopiclone, may not detect low concen-
trations of these drugs. We developed a liquid chro-
matographic–tandem mass spectrometric method for
identifying these drugs and their relevant metabolites.
Methods: We extracted BZPs from urine by solid-phase
extraction with a mixed-mode phase (OASIS® HLB
cartridges). Chromatographic separation was performed
with a Waters XTerra MS C18 [150 ⴛ 2.1 mm (i.d.); bead
size, 5 ␮m] reversed-phase column with deuterated
analogs of the analytes as internal standards (IS). Detec-
tion was performed with a triple-quadruple mass spec-
trometer that monitored 2 specific transitions per com-
pound in the electrospray, positive-ion selected-reaction
monitoring mode. We tested this technique on urine
samples from 12 healthy volunteers and 1 forensic
sample obtained in a case of alleged drug-facilitated
sexual assault.
Results: Chromatographic separation was achieved
within 18 min. The linear dynamic ranges extended
from 0.02 or 0.1 ␮g/L (depending on the drug or metab-
olite) to 50 ␮g/L. Extraction recovery (range) was 77%–
110%. Limits of detection were <0.05 ␮g/L. No ion
1
Department of Pharmacology-Toxicology, Limoges University Hospital,
Limoges, France.
2
Forensic Toxicology Department, Institute of Legal Medicine, University
of Santiago de Compostela, C/San Francisco, Spain.
* Address correspondence to this author at: Department of Pharmacology-
Toxicology, Limoges University Hospital, 87042 Limoges, France. Fax 33-555-
05-61-62; e-mail pierre.marquet@unilim.fr.
Received December 21, 2005; accepted May 2, 2006.
Previously published online at DOI: 10.1373/clinchem.2005.065631
1346
Drug Monitoring and
Toxicology
suppression was seen except for alprazolam, for which
baseline decreased by almost 20%. In the forensic urine
sample, the method detected alprazolam (3.5 ␮g/L) and
its characteristic metabolite, ␣-hydroxyalprazolam
(0.17 ␮g/L).
Conclusion: This method measured low concentrations
of BZPs and BZP-like substances and might be useful
for analyses of urine in suspected drug-facilitated sex-
ual assault cases.
© 2006 American Association for Clinical Chemistry
The benzodiazepines (BZPs)3 and BZP-like substances
(e.g., zopiclone and zolpidem) are relatively safe drugs
that are frequently prescribed for the treatment of stress,
panic disorders, sleep disorders, muscle spasms, alcohol
withdrawal, and seizures (1 ). Because of the high con-
sumption of these substances world-wide, they are in-
volved in many forensic cases. The influence of BZPs on
driving and in jobs in which the presence of high serum
concentrations in employees could be dangerous has been
documented (2– 6 ). These drugs also play a role in drug-
facilitated sexual assault (DFSA) (7, 8 ).
Zopiclone and zolpidem belong to a new generation of
hypnotics and sedatives that are structurally quite differ-
ent from the BZPs (9 ). They have a rapid onset of action
and a short half-life and, unlike BZPs, have weak myore-
3
Nonstandard abbreviations: BZP, benzodiazepine; DFSA, drug-facili-
tated sexual assault; LC-MS, liquid chromatography–mass spectrometry; MS/
MS, tandem MS; ESI electrospray ionozation; IS, internal standard; SPE,
solid-phase extraction; SRM, single reaction monitoring; LLOQ, lower limit of
quantification; ULOQ, upper limit of quantification; MRE, mean relative error;
RT, retention time; and GC, gas chromatography.
 Clinical Chemistry 52, No. 7, 2006
laxant and anticonvulsant effects. They are prescribed as
hypnotics instead of BZPs (10 ).
Many widely used screening and confirmation meth-
ods do not detect lower concentrations of these drugs, and
in some cases, the laboratory may not have the capability
to detect a particular drug (11 ). New ionization tech-
niques have made liquid chromatography–mass spec-
trometry (LC-MS) and LC-tandem MS (LC-MS/MS) ex-
tremely effective for the analysis of BZPs and BZP-like
drugs. We developed an LC electrospray ionization (ESI)-
MS/MS method for screening and quantification of zopi-
clone, zolpidem, and several BZPs and their relevant
metabolites, with a focus on the analysis of urine samples
from DFSA cases.
Materials and Methods
chemicals, reagents, and solutions
The calibrators used were 1-hydroxymidazolam, 4-hy-
droxymidazolam, midazolam, 7-aminoclonazepam, des-
methylflunitrazepam, flunitrazepam, and 7-aminofluni-
trazepam (all from Lipomed); alprazolam, bromazepam,
clonazepam, triazolam, lorazepam, 3-hydroxyflunitraz-
epam, zolpidem, and zopiclone (all from Sigma); fluraz-
epam, desalkylflurazepam, ␣-hydroxyalprazolam, ␣-hy-
droxytriazolam, and the deuterated internal standards
(IS) 7-aminoflunitrazepam-d7, 7-aminoclonazepam-d4,
alprazolam-d5, triazolam-d4, ␣-hydroxyalprazolam-d5,
desmethylflunitrazepam-d4, ␣-hydroxytriazolam-d4, de-
salkylflurazepam-d4, and flunitrazepam-d7 (all from
Ceriliant); lormetazepam (kindly provided by Schering,
Levallois-Perret, France); and loprazolam (kindly pro-
vided by Sanofi-Aventis, Paris, France).
Propanol-2 (Normapur), dichloromethane (Norma-
pur), and ammonia (Normapur) were purchased from
Prolabo. HPLC-grade acetonitrile (Normapur) and meth-
anol (Normapur) were obtained from Carlo-Erba. Formic
acid and ammonium formate (⬎99% pure for both) were
purchased from Sigma. Deionized water was prepared
on a Direct-Q laboratory plant (Millipore). For sample
extraction, Waters Oasis® HLB extraction cartridges (3
mL/60 mg) were used.
We prepared a pH 7.4 Sorensen buffer (66.7 mmol/L)
by dissolving 9.07 g of potassium dihydrogen phosphate
into 1 L of deionized water and by dissolving 11.6 g of
disodium hydrogen phosphate anhydrous into 1 L of
deionized water. The disodium hydrogen phosphate so-
lution was then used to adjust the potassium dihydrogen
phosphate solution to pH 7.4. A solution of 0.5% ammonia
in methanol–water (40:60 by volume) was prepared by
adding 1.25 mL of concentrated ammonia to a mixture of
99.5 mL of methanol and 149.25 mL of water.
Stock solutions of desalkylflurazepam, ␣-hydroxytria-
zolam, and ␣-hydroxyalprazolam were prepared at 10
mg/L in methanol; all stock solutions of the deuterated
standards were prepared at 1 mg/L, and those of the
remaining drugs at 1 g/L in methanol. All solutions were
kept at ⫺20 °C in the dark. The drugs are stable for at least
1347
3 months under these conditions. Working solutions of
mixtures of the analytes at concentrations of 0.0004, 0.001,
0.002, 0.01, 0.04, 0.2, and 1 mg/L were freshly prepared
each day by appropriate dilution in methanol–water
(50:50 by volume). In the same way, a 0.1 mg/L working
solution was prepared from all the IS stock solutions.
sample preparation
We added 1 mL of Sorensen buffer, 50 ␮L of working IS
solution, and 100 ␮L of the appropriate working solutions
of the parent drugs and metabolites to 2 mL of urine to
obtain the following calibration concentrations: 0, 0.02,
0.05, 0.1, 0.5, 2, 10, and 50 ␮g/L. The tubes were then
vortex-mixed for 10 s. Solid-phase extraction (SPE) was
performed with Oasis HLB cartridges (3 mL/60 mg)
previously conditioned with 2 mL of methanol and equil-
ibrated with 2 mL of water. The mixtures of sample and
buffer were loaded on the cartridge and slowly passed
through the bed without vacuum. The SPE column was
rinsed sequentially with 3 mL of deionized water and
3 mL of 0.5% ammonia in methanol–water (40:60 by
volume). Reduced pressure was then applied to the
column (maximum of ⫺60 kPa) for 15 min to dry the
column. The retained drugs were eluted under gravity
with 3 mL of dichloromethane–propanol-2 (75:25 by vol-
ume) into 5-mL round-bottomed glass tubes. The eluate
was evaporated under nitrogen at 40 °C in a Zymark
TurboVap® LV Concentration Workstation. The dried
extracts were reconstituted in 50 ␮L of mobile phase, and
15 ␮L was injected into the chromatographic system.
hplc conditions
The HPLC system consisted of a Shimadzu LC 10 ADvp
pump with an SIL-Ht autoinjector and a CTO-10ASvp
column oven. The chromatographic separation was per-
formed on a Waters XTerra MS C18 [150 ⫻ 2.1 mm (i.d.);
bead size, 5 ␮m] reversed-phase column maintained at
30 °C. The flow rate was 0.2 mL/min. The mobile phase
was a gradient of a mixture of acetonitrile and 2 mmol/L
(pH 3) ammonium formate (90:10 by volume; solvent A)
and 2 mmol/L (pH 3) ammonium formate (solvent B),
programmed as follows: initial, 30% A, increased to 36%
in 4 min, maintained at 36% for 8.5 min, then increased to
90% A in 3.5 min, and finally, decreased to 30% A in the
last 1.5 min. All chromatographic solvents were degassed
with helium.
A divert valve was set to direct the LC flow initially to
the waste before 1.3 min and after 18 min to protect the
spectrometer from salts, polar compounds, and lipids.
ms detection
Detection was carried out with a Thermo-Electron TSQ
Quantum Discovery MS/MS system equipped with an
orthogonal ESI source and controlled by the Xcalibur
computer program. Ionization was achieved with electro-
spray in the positive ionization mode. The following
optimized conditions were used: spray voltage, 4 kV;
1348 Quintela et al.: LC-MS/MS Assay for BZPs
capillary temperature, 325 °C; sheath gas pressure (N2), 35
kPa; auxiliary gas pressure (N2), 25 kPa.
For selected reaction monitoring (SRM) of the individ-
ual compounds and their respective deuterated analogs,
the pseudomolecular ions [M⫹H]⫹ were selected in the
first quadrupole, and the collision energy was adjusted to
optimize the signal for the most abundant product ions. A
quantification and a confirmation transition were chosen
for each compound, except for the deuterated analogs, for
which only 1 transition was chosen (Table 1). The chro-
matographic run was divided into 3 segments to limit the
number of transitions monitored at a time in MS/MS
mode and to ensure sensitivity for quantification pur-
poses. In each segment, the compounds were monitored
according to their chromatographic separation (Table 1).
The total LC-ESI-MS/MS method was optimized to
detect 32 substances (including 10 IS).
validation procedures
We performed assay validation according to the guide-
lines of the US Food and Drug Administration (12 ). We
evaluated selectivity by analyzing urine samples from 10
healthy volunteers who did not take any of the targeted
compounds for several days before urine sampling and
checked for the absence of the compounds of interest by
analyzing the samples with the present technique before
using them as blanks. We investigated assay linearity by
constructing calibration curves (n ⫽ 5) using analysis data
from a blank urine sample and 8 urine samples enriched
with the analytes at concentrations of 0.02, 0.05, 0.1, 0.5, 2,
10, 25, and 50 ␮g/L. Stable isotope IS were used to correct
for variability in the extraction procedure and the effect of
the ion suppression produced by the matrix. When it was
not possible to obtain a commercially available deuterated
analog for any compound, the deuterated analog of
another close compound was selected. We determined the
best fit for accuracy for the calibration curves by applying
linear and quadratic regression and different types of
weighting (1/x, 1/x2).
Extraction recoveries were determined in triplicate at 3
concentrations for all compounds studied: the lower limit
of quantification (LLOQ) was 0.02 or 0.05 ␮g/L, depend-
ing on the molecule; the upper limit of quantification
(ULOQ) was 50 ␮g/L; and the intermediate limit was
2 ␮g/L. In the case of bromazepam and ␣-hydroxyalpra-
zolam, the extraction recoveries were studied only at 2
and 50 ␮g/L because the first concentration (0.02 or
0.05 ␮g/L) was outside their calibration ranges. For each
concentration, 3 urine samples were enriched with the
appropriate amounts of drugs and with the IS, and in 3
other samples, only the IS was added. After extraction
and evaporation, the first 3 samples were reconstituted
with 50 ␮L of mobile phase and the last 3 samples with
50 ␮L of mobile phase containing a nominal amount of the
compounds of interest. The relative extraction efficiency
was calculated as the quotient of the peak-area ratios of
the extracted samples with those of the unextracted
solutions, which represented 100% recovery.
We evaluated within-run precision and recovery by
analyzing 6 samples in the same batch, enriched with all
substances studied, for 3 or 4 concentrations of the cali-
bration set, depending on the linear dynamic range of the
compound. We evaluated between-run precision and re-
covery, as well as the linearity of the method, by prepar-
ing and analyzing 1 set of calibrators each day for 5 days.
Required specifications for within- and between-batch
experiments were a CV ⬍15% (precision) and a mean
relative error (MRE) ⬍15% from the nominal value at
every concentration studied, except for the LLOQ, where
20% was acceptable for both values (12–14 ). The limit of
detection was defined as the lowest concentration of the
drug giving a signal-to-noise ratio ⬎3:1 and was deter-
mined for all compounds by analysis of a blank urine
enriched with decreasing amounts of the analytes. The
LLOQ was the lowest concentration of the calibration
curve that could be measured with an imprecision and
deviation from target concentration ⬍20%, in terms of CV
and MRE, respectively (12–14 ).
We investigated ion suppression by injecting 5 extracts
of different blank human urine samples into the system
while a mixture of the BZPs in mobile phase was contin-
uously infused in parallel at 50 ␮L/min in the ionization
source through a PEEK tee-junction. We investigated the
stability of the extracts in the autosampler by extracting,
in quadruplicates, a urine sample enriched with the 21
compounds at 10 ␮g/L. The second extract was injected
6 h after the first, the third extract 12 h after, and the
fourth 15 h after.
applications
This technique was used in a clinical trial approved by
our local ethics committee. Twelve healthy volunteers
who gave informed consent were administered a single
dose of bromazepam, clonazepam, flunitrazepam, loraz-
epam, zolpidem, or zopiclone (2 patients per drug) at the
lowest dosage available. Urine samples were collected
every 12 h for 1 week after dosing. The technique was also
applied to a forensic sample obtained in a case of alleged
DFSA.
Results
Separation of the 21 compounds and their respective IS
was achieved in 18 min. A 6-min equilibration time was
necessary at the end of each run. The extracted ion
chromatograms of the 21 BZPs obtained for a urine
sample enriched at 2 ␮g/L are presented in Fig. 1. The
selectivity of the method was acceptable in terms of
absence of interference in the blank urine samples ana-
lyzed. On the total ion chromatogram, not all of the
chromatographic peaks resolved, but this was not a
problem because quantification and specific identification
were performed for each compound with the correspond-
ing transitions in the SRM mode. The combination of
 Clinical Chemistry 52, No. 7, 2006 1349

Table 1. MS and chromatographic characteristics for tested BZPs.

Segment of MS experiment Collision
Compound (I, II, or III) SRM transition,a m/z energy, V RT, min
Zopiclone I 389.13245.1 26 2.3
389.13217.2 42
7-Aminoclonazepam I 286.23222.1 30 3.5
286.23250.1 26
Zolpidem I 308.03263.1 34 2.8
308.03235.1 30
7-Aminoflunitrazepam I 284.03135.0 34 4.3
284.03226.0 38
4-Hydroxymidazolam I 342.13325.0 26 3.9
342.13297.0 30
1-Hydroxymidazolam I 342.13324.0 26 5.8
342.13203.0 34
Loprazolam I 465.43408.1 30 4.5
465.43252.2 46
Midazolam I 326.13291.2 32 4.8
326.13249.2 40
Bromazepam I 316.03181.1 30 6.1
316.03209.2 34
␣-Hydroxytriazolam/4-Hydroxytriazolam II 359.03331.0 32 8.4
359.03250.0 42
␣-Hydroxyalprazolam II 325.13297.1 32 8.5
325.13216.1 44
3-Hydroxyflunitrazepam II 330.03284.1 24 9.4
330.03238.1 36
Desmethylflunitrazepam II 300.03254.1 30 10.0
300.03225.1 40
Alprazolam II 309.13281.2 28 11.1
309.13205.2 46
Lorazepam II 321.03275.1 28 11.2
321.03303.1 20
Triazolam II 343.03308.0 32 12.1
343.03239.0 48
Clonazepam II 316.03270.0 30 12.2
316.03241.0 40
Desalkylflurazepam III 289.03140.1 36 14.3
289.03226.1 34
Flurazepam I 388.23315.1 26 4.8
388.23317.1 24
Flunitrazepam III 314.03268.0 32 14.9
314.03239.0 40
Lormetazepam III 335.03289.1 26 16.6
335.03177.1 46
7-Aminoclonazepam-d4 I 290.03226.2 30 3.5
7-Aminoflunitrazepam-d7 I 291.03138.1 34 4.2
␣-Hydroxytriazolam-d4 II 363.13176.1 32 8.2
␣-Hydroxyalprazolam-d5 II 330.03302.1 30 8.3
Desmethylflunitrazepam-d4 II 304.13258.1 28 9.8
Alprazolam-d5 II 314.03286.1 32 10.8
Triazolam-d4 II 346.93312.1 30 11.8
Desalkylflurazepam-d4 III 293.13230.0 32 14.1
Flunitrazepam-d7 III 321.13246.2 30 14.5
a
The selected quantification transitions are in bold.
1350 Quintela et al.: LC-MS/MS Assay for BZPs
Fig. 1. Analysis of blank urine samples enriched with 2 ␮g/L each of the studied compounds.
TIC, total ion chromatogram.
retention time (RT) and 2 transitions (and their relative
abundances) provided high specificity for all of the com-
pounds. However, with this analytical method, we could
not distinguish ␣-hydroxytriazolam from 4-hydroxytria-
zolam because we obtained exactly the same RT and SRM
transitions for both, with the same intensity ratio. We
therefore introduced ␣-hydroxytriazolam, but not 4-hy-
droxytriazolam, into the calibration and control samples.
In addition, 2 peaks showed up on the reconstituted
chromatogram of lorazepam: the first peak corresponded
to lorazepam and the second to the IS, flunitrazepam-d7
(transition m/z 321.13246.2). However, these 2 com-
pounds were perfectly separated, so that flunitraz-
epam-d7 could be kept as IS. The SRM transitions, opti-
mized collision energies, and RT are shown in Table 1.
The deuterated IS selected for each compound is
shown in Table 2. No cross-talk interference with the
deuterated IS was observed. The peak-area ratios of the
drugs to their respective deuterium-labeled IS were plot-
ted against the corresponding enriched concentrations.
The linearity of the method was verified over the concen-
tration range. After assaying different types of regression
and weighing factors, we constructed the calibration
curves using a 1/x weighted quadratic regression to
obtain the best fit across the calibration range, based on
the standard error of the fit and minimization of bias of
the low calibrators. The squared correlation coefficient
(r2), y-intercept, and slope of the regression line are
summarized in Table 2. The mean r2 (n ⫽ 5) was ⬎0.998
for all of the compounds except for 4-hydroxymidazolam
(r2 ⫽ 0.998).
Limits of detection between 0.01 and 0.02 ␮g/L were
observed for all of the compounds except for bromaz-
epam and ␣-hydroxyalprazolam (0.05 ␮g/L). The LLOQ
and ULOQ values for the compounds are shown as the
linear dynamic range in Table 2. The within-run impreci-
sion (CV) was ⬍17% in all cases for the LLOQ, and was
⬎15% for only 3 drugs: loprazolam, bromazepam, and
␣-hydroxytriazolam. For the other concentrations, only
one CV was ⬎15% (16.4% for loprazolam at the 2 ␮g/L
concentration). The within-run deviation from target was
satisfactory for all compounds under the noted criteria,
 Clinical Chemistry 52, No. 7, 2006 1351

Table 2. Calibration data and limits of detection for the examined compounds.
Linear dynamic range, Correlation
Compound IS ␮g/L LOD,a ␮g/L coefficient (r2)
Zopiclone 7-Aminoclonazepam-d4 0.05–50 0.01 0.999
7-Aminoclonazepam 7-Aminoclonazepam-d4 0.02–50 0.01 0.999
Zolpidem 7-Aminoclonazepam-d4 0.02–50 0.01 0.999
7-Aminoflunitrazepam 7-Aminoflunitrazepam-d7 0.02–50 0.01 0.999
4-Hydroxymidazolam 7-Aminoclonazepam-d4 0.05–50 0.02 0.998
1-Hydroxymidazolam 7-Aminoflunitrazepam-d7 0.05–50 0.02 0.999
Loprazolam 7-Aminoflunitrazepam-d7 0.05–50 0.02 0.999
Midazolam 7-Aminoflunitrazepam-d7 0.05–50 0.02 0.999
Bromazepam 7-Aminoflunitrazepam-d4 0.1–50 0.05 0.999
␣-Hydroxytriazolam/4-Hydroxytriazolam ␣-Hydroxytriazolam-d4 0.05–50 0.02 0.999
␣-Hydroxyalprazolam ␣-Hydroxyalprazolam-d5 0.1–50 0.05 0.999
3-Hydroxyflunitrazepam Desmethylflunitrazepam-d4 0.05–50 0.02 0.999
Desmethylflunitrazepam Desmethylflunitrazepam-d4 0.02–50 0.01 0.999
Alprazolam Alprazolam-d5 0.05–50 0.02 0.999
Lorazepam Triazolam-d4 0.02–50 0.01 0.999
Triazolam Triazolam-d4 0.02–50 0.01 0.999
Clonazepam Triazolam-d4 0.05–50 0.02 0.999
Desalkylflurazepam Desalkylflurazepam-d4 0.05–50 0.02 0.999
Flurazepam 7-Aminoclonazepam-d4 0.05–50 0.02 0.999
Flunitrazepam Desalkylflurazepam-d4 0.05–50 0.02 0.999
Lormetazepam Desalkylflurazepam-d4 0.02–50 0.01 0.999
a
LOD, limit of detection; defined as the lowest concentration of the drug giving a signal-to-noise ratio of 3:1.

except for loprazolam, for which the MRE value reached
24% at the LLOQ, and bromazepam, for which it reached
22.4% and 21.2% at the LLOQ and 2 ␮g/L, respectively.
The between-run CV was ⬍15% for all compounds at
all concentrations of the calibration range, including the
LLOQ, except for zolpidem, which had a slightly higher
result for the LLOQ (CV ⫽ 15%), although it was still
acceptable according to the Food and Drug Administra-
tion recommendations. For all substances, between-run
MRE values were ⱕ14% at all concentrations, including
the LLOQ.
The extraction recovery of the analytes at 3 concentra-
tions with the SPE procedure was 79.0%–99.3% at the
LLOQ, 77.2%–99.6% at the intermediate concentration,
and 88.5%–110.0% at the ULOQ (Table 3). No significant
ion suppression occurred during chromatographic runs,
except for alprazolam, for which the baseline decreased
almost 20% (see the Data Supplement that accompanies
the online version of this article at http:www.clinchem.
org/content/vol52/issue7). The stability of the extracts in
the autosampler, studied with an enriched urine sample
extracted in quadruplicate (Table 4), showed that 4-hy-
droxymidazolam may begin to deteriorate after 6 h and
midazolam and loprazolam after 12 h. This instability of
BZPs in solution, despite low temperatures (the auto-
sampler was cooled at 4 °C in this study) and protection
from light, is well known and probably unavoidable. As a
consequence, when large batches of samples were to be
analyzed, a portion of each extract was stored dry at 4 °C
before being reconstituted and placed in the autosampler
a few hours before analysis.
The clinical trial in healthy volunteers given a single
dose of bromazepam, clonazepam, flunitrazepam, loraz-
epam, zolpidem, or zopiclone (2 patients per drug) at the
lowest dosage available showed that lorazepam, broma-
zepam, and 7-aminoflunitrazepam could be detected in
urine up to 144 h (1 week) post dose, whereas flunitraz-
epam was not detected after 132 h, zopiclone after 84 h,
and zolpidem after 48 h in one volunteer and after 96 h in
the other. These results demonstrate the suitability of this
method, which we have used successfully in our labora-
tory for the investigation of actual DFSA cases.
The results obtained in a DFSA case are shown in Fig.
2. In this urine sample, the assay detected alprazolam at
3.5 ␮g/L and its characteristic metabolite, ␣-hydroxy-
alprazolam, at a low concentration of 0.17 ␮g/L.
Discussion
Because higher concentrations of drugs and their metab-
olites can usually be found in urine, it is the specimen of
choice for toxicologic testing in DFSA cases (15 ). One of
the most important issues in DFSA investigations is the
sensitivity of the screening and confirmatory techniques
because some of the compounds (e.g., BZPs) are typically
used in a single low dose (16, 17 ). Kintz et al. (18 )
demonstrated that the LC-MS and gas chromatography
(GC)-negative chemical ionization-MS/MS techniques
are the most sensitive for detecting low quantities of the
drug metabolite 7-aminoflunitrazepam several days after
administration of flunitrazepam. A highly sensitive GC-
negative chemical ionization-MS method for the simulta-
1352 Quintela et al.: LC-MS/MS Assay for BZPs

Table 3. Extraction recovery of the 21 BZPs and metabolites.

Relative recovery, % (n ⴝ 3)

Lowa Mediumb Highc

Compound concentration concentration concentration
Zopiclone 83.5 91.2 96.6
7-Aminoclonazepam 84.1 94.1 94.0
Zolpidem 79.0 96.3 92.0
7-Aminoflunitrazepam 81.4 95.7 98.5
4-Hydroxymidazolam 95.5 96.5 99.3
1-Hydroxymidazolam 97.3 93.4 95.6
Loprazolam 92.4 91.5 100.0
Midazolam 97.0 94.2 91.8
Bromazepam 77.2 88.5
␣-Hydroxytriazolam/4-Hydroxytriazolam 92.6 99.3 94.6
␣-Hydroxyalprazolam 99.6 93.3
3-Hydroxyflunitrazepam 94.2 95.4 96.0
Desmethylflunitrazepam 83.3 96.8 95.8
Alprazolam 95.6 96.6 96.0
Lorazepam 90.0 97.6 91.6
Triazolam 83.1 95.5 91.3
Clonazepam 89.7 96.4 96.3
Desalkylflurazepam 91.7 96.4 110
Flurazepam 99.3 85.1 88.8
Flunitrazepam 95.7 95.9 90.4
Lormetazepam 79.4 86.8 91.5
a
LLOQ (0.02 or 0.05, depending on compound).
b
2 ␮g/L for all compounds.
c
ULOQ (50 ␮g/L for all compounds).

neous quantification of clonazepam and its major metab- LC-MS/MS) assays have been reported for forensic appli-
olite in urine detected the metabolite 28 days after a single cations, including use in DFSA cases (19 –23 ).
dose (3 mg) of clonazepam (19 ). Other LC-MS (and We evaluated several previous procedures for extrac-

Table 4. Stability in the autosampler of extracts of the 21 BZPs and metabolites at 10 ␮g/L.
Recovery (%), with elapsed time before analysis of:

0 6h 12 h 15 h
Zopiclone 100 80 85 113
7-Aminoclonazepam 100 97 101 104
Zolpidem 100 81 84 90
7-Aminoflunitrazepam 100 85 78 96
4-Hydroxymidazolam 100 77 71 61
1-Hydroxymidazolam 100 99 93 83
Loprazolam 100 85 80 76
Midazolam 100 99 86 69
Bromazepam 100 96 88 90
␣-Hydroxytriazolam 100 105 103 106
␣-Hydroxyalprazolam 100 103 99 105
3-Hydroxyflunitrazepam 100 97 89 91
Desmethylflunitrazepam 100 103 106 97
Alprazolam 100 108 107 106
Lorazepam 100 99 101 97
Triazolam 100 100 101 100
Clonazepam 100 108 112 105
Desalkylflurazepam 100 105 105 106
Flurazepam 100 92 91 99
Flunitrazepam 100 114 111 100
Lormetazepam 100 104 111 89
 Clinical Chemistry 52, No. 7, 2006 1353

Fig. 2. Chromatograms of a forensic DFSA sample collected ⬍24 h after the assault.
The urine sample contained alprazolam and its metabolite ␣-hydroxyalprazolam at 3.5 and 0.17 ␮g/L, respectively.

tion of the BZPs and related compounds during develop- principal reason for our use of an SPE method. Variation
ment of the method we report here (24 ). The presence of in the pH of urine samples induced significant variations
compounds in a relatively wide polarity range was the in compound retention on the SPE column and in chro-
1354 Quintela et al.: LC-MS/MS Assay for BZPs
matogram background noise. The best results were ob-
tained by dilution of urine samples with Sorensen buffer
(pH 7.4). Because of the high sensitivity of our method,
which allowed detection of the parent drug and phase I
metabolite at very low concentrations, and the instability
of several analytes during incubation (mainly the amino
metabolites), we processed the urine samples without
enzymatic hydrolysis.
Our results demonstrate that this technique meets
method validation criteria for measuring analytes of in-
terest in urine. In the case of bromazepam and loprazo-
lam, certain values were not in the expected range, but
this should not present a problem for applying this
technique to urine samples in DFSA cases, where precise
quantification is not the most important goal.
The use of HPLC to separate the compounds before
detection by MS avoids the temperature-induced degra-
dation of certain BZPs that can occur when GC methods
are used (10 ). Another major advantage of HPLC for
analyzing BZPs is that no derivatization of the metabolites
is required (11 ).
Compared with other reports on analysis of several
BZPs (25–35 ), in the screening method described in this
study, the LLOQs for 19 BZPs or their major metabolites
and 2 related compounds (zopliclone and zolpidem) are
lower and seem to be appropriate for the detection and
quantification of these compounds in urine samples from
DFSA cases.
We thank Bea López and Anna Chasnoff for helpful
assistance in the preparation of this manuscript. This
work was supported by a grant from the Ministry of
Education and Science of Spain under the Formación de
Profesorado Universitario Program.
References
1. Stewart CH. Hypnotics and sedatives. In: Goodman Gilman A,
Goodman LS, Rall TW, Murad F, eds. Gilman’s The Pharmacolog-
ical Basis of Therapeutics, 7th ed. New York: Macmillan Publish-
ing Company, 1985:339 –71.
2. Currie D, Hashemi K, Fothergill J, Findlay A, Harris A, Hindmarch I.
The use of anti-depressants and benzodiazepines in the perpetra-
tors and victims of accidents. Occup Med 1995;45:323–5.
3. Ray WA, Fought RL, Decker MD. Psychoactive drugs and the risk of
injurious motor vehicle crashes in elderly drivers. Am J Epidemiol
1992;136:873– 83.
4. Neutel CI. Benzodiazepine-related traffic accidents in young and
elderly drivers. Hum Psychopharmacol 1998;13(Suppl 2):115–23.
5. Barbone F, McMahon AD, Davey PG, Morris AD, Reid IC, McDevitt
DG, et al. Association of road-traffic accidents with benzodiaz-
epine use. Lancet 1998;352:1331– 6.
6. Logan BK, Couper FJ. Zolpidem and driving impairment. J Forensic
Sci 2001;46:105–10.
7. Dowd SM, Strong MJ, Janicak PG, Negrusz A. The behavioral and
cognitive effects of two benzodiazepines associated with drug-
facilitated sexual assault. J Forensic Sci 2002;47:1101–7.
8. Villain M, Cheze M, Dumestre V, Ludes B, Kintz P. Hair to
document drug-facilitated crimes: four cases involving bromaz-
epam. J Anal Toxicol 2004;28:516 –9.
9. Drummer OH, Odell M, eds. The Forensic Pharmacology of Drugs
of Abuse. London: Arnold, 2001:103– 66.
10. Kraemer T, Maurer H. Therapeutic drugs: sedatives and hypnotics.
In: Bogusz MJ, ed. Handbook of Analytical Separations. Amster-
dam: Elsevier Science BV, 2000:197–223.
11. Salamone SJ, ed. Benzodiazepines and GHB: Detection and
Pharmacology. Totowa, NJ: Humana Press, 2001:160pp.
12. US Department of Health and Human Services, Food and Drug
Administration. Guidance for industry: bioanalytical method val-
idation. http: // www.fda.gov/ cder/ guidance/4252fnl.pdf (ac-
cessed March 2006).
13. Shah VP, Midha KK, Findlay JW, Hill HM, Hulse JD, McGilveray IJ,
et al. Bioanalytical method validation-a revisit with a decade of
progress. Pharm Res 2000;17:1551–7.
14. International Conference on Harmonization. Validation of analyti-
cal procedures: text and methodology. Q2(R1). http://www.
ich.org/ (accessed March 2006).
15. LeBeau M, Andollo W, Hearn WL, Baselt R, Cone E, Finkle B, et al.
Recommendations for toxicological investigations of drug-facili-
tated sexual assaults. J Forensic Sci 1999;44:227–30.
16. Negrusz A, Gaensslen RE. Analytical developments in toxicological
investigation of drug-facilitated sexual assault. Anal Bioanal Chem
2003;376:1192–7.
17. Couper FJ, Logan BK. Determination of ␥-hydroxybutyrate (GHB) in
biological specimens by gas chromatography-mass spectrometry.
J Anal Toxicol 2000;24:1–7.
18. Kintz P, Villain M, Cirimele V, Goulle JP, Ludes B. Crime under the
influence of psychoactive drugs: the problem of the duration of
detection. Acta Clin Belg 2001;1(Suppl):24 –30.
19. Negrusz A, Bowen AM, Moore CM, Dowd SM, Strong MJ, Janicak
PG. Elimination of 7-aminoclonazepam in urine after a single dose
of clonazepam. Anal Bioanal Chem 2003;376:1198 –204.
20. Marquet P, Lachâtre G. Liquid chromatography-mass spectrome-
try: potential in forensic and clinical toxicology. J Chromatogr B
Biomed Sci Appl 1999;733:93–118.
21. Kintz P, Villain M, Cirimele V, Pepin G, Ludes B. Windows of
detection of lorazepam in urine, oral fluid, and hair, with a special
focus on drug-facilitated crimes. Forensic Sci Int 2004;145:
131–5.
22. Kintz P, Villain M, Dumestre-Toulet V, Ludes B. Drug-facilitated
sexual assault and analytical toxicology: the role of LC-MS/MS: a
case involving zolpidem. J Clin Forensic Med 2005;12:36 – 41.
23. Quintela O, Cruz A, Castro A, Concheiro M, Lopez-Rivadulla M.
Liquid chromatography-electrospray ionization mass spectrometry
for the determination of nine selected benzodiazepines in human
plasma and oral fluid. J Chromatogr B Analyt Technol Biomed Life
Sci 2005;825:63–71.
24. Drummer OH. Methods for the measurement of benzodiazepines
in biological samples. J Chromatogr B Biomed Sci Appl 1998;713:
201–25.
25. Pirnay S, Ricordel I, Libong D, Bouchonnet S. Sensitive method for
the detection of 22 benzodiazepines by gas chromatography-ion
trap tandem mass spectrometry. J Chromatogr A 2002;954:235–
45.
26. Bishop SC, Lerch M, McCord BR. Micellar electrokinetic chromato-
graphic screening method for common sexual assault drugs
administered in beverages. Forensic Sci Int 2004;141:7–15.
27. Jinno K, Taniguchi M, Hayashida M. Solid phase micro extraction
coupled with semi-microcolumn high performance liquid chroma-
tography for the analysis of benzodiazepines in human urine.
J Pharm Biomed Anal 1998;17:1081–91.
28. McClean S, O’Kane E, Hillis J, Smyth WF. Determination of
1,4-benzodiazepines and their metabolites by capillary electro-
 Clinical Chemistry 52, No. 7, 2006
phoresis and high-performance liquid chromatography using ultra-
violet and electrospray ionization mass spectrometry. J Chro-
matogr A 1999;838:273–91.
29. Rittner M, Pragst F, Bork WR, Neumann J. Screening method for
seventy psychoactive drugs or drug metabolites in serum based
on high-performance liquid chromatography-electrospray ioniza-
tion mass spectrometry. J Anal Toxicol 2001;25:115–24.
30. Miki A, Tatsuno M, Katagi M, Nishikawa M, Tsuchihashi H.
Simultaneous determination of eleven benzodiazepine hypnotics
and eleven relevant metabolites in urine by column-switching
liquid chromatography-mass spectrometry. J Anal Toxicol 2002;
26:87–93.
31. Zweigenbaum J, Heinig K, Steinborner S, Wachs T, Henion J.
High-throughput bioanalytical LC/MS/MS determination of benzo-
diazepines in human urine: 1000 samples per 12 hours. Anal
Chem 1999;71:2294 –300.
32. Smink BE, Brandsma JE, Dijkhuizen A, Lusthof KJ, de Gier JJ,
Egberts AC, et al. Quantitative analysis of 33 benzodiazepines,
1355
metabolites, and benzodiazepine-like substances in whole blood
by liquid chromatography-(tandem) mass spectrometry. J Chro-
matogr B Analyt Technol Biomed Life Sci 2004;811:13–20.
33. Yuan H, Mester Z, Lord H, Pawliszyn J. Automated in-tube solid-
phase microextraction coupled with liquid chromatography-electro-
spray ionization mass spectrometry for the determination of
selected benzodiazepines. J Anal Toxicol 2000;24:718 –25.
34. Klette KL, Wiegand RF, Horn CK, Stout PR, Magluilo Jr J. Urine
benzodiazepine screening using Roche Online KIMS immuno-
assay with ␤-glucuronidase hydrolysis and confirmation by gas
chromatography-mass spectrometry. J Anal Toxicol 2005;29:
193–200.
35. Gunnar T, Ariniemi K, Lillsunde P. Determination of 14 benzodi-
azepines and hydroxy metabolites, zaleplon, and zolpidem as
tert-butyldimethylsilyl derivatives compared with other common
silylating reagents in whole blood by gas chromatography-mass
spectrometry. J Chromatogr B Analyt Technol Biomed Life Sci
2005;818:175– 89.