Journal of Analytical Toxicology, Vol. 34, July/August 2010
A Validated Method for Simultaneous Screening and Quantification of Twenty-Three Benzodiazepines and Metabolites Plus Zopiclone and Zaleplone in Whole Blood by Liquid–Liquid Extraction and Ultra-Performance Liquid Chromatography– Tandem Mass Spectrometry Kirsten Wiese Simonsen1,*, Sigurd Hermansson2, Anni Steentoft1, and Kristian Linnet1 1Section
of Forensic Chemistry, Department of Forensic Medicine, Faculty of Health Sciences, University of Copenhagen, Frederik V’s vej 11, 3. DK-2100, Denmark and 2Waters Sweden AB, Djupdalsvagen 12-14, SE-191-24 Sollentuna, Sweden
Introduction
Abstract An ultra-performance liquid chromatography–tandem mass spectrometry (UPLC–MS–MS) method for detection of 23 benzodiazepines and related compounds in whole blood was developed and validated. The method is used for screening and quantitation of benzodiazepines in whole blood received from autopsy cases and living persons. The detected compounds were alprazolam, bromazepam, brotizolam, chlordiazepoxide, demoxepam, clobazam, clonazepam, 7-aminoclonazepam, diazepam, nordiazepam, estazolam, flunitrazepam, 7-aminoflunitrazepam, lorazepam, lormetazepam, midazolam, nitrazepam, 7-aminonitrazepam, oxazepam, temazepam, triazolam, zaleplon, and zopiclone. Whole blood from drug-free volunteers was used for all experiments. Blood samples (0.200 g) were extracted with ethyl acetate at pH 9. Target drugs were quantified using a Waters ACQUITY UPLC system coupled to a Waters Quattro Premier XE triple quadrupole in positive electrospray ionization, multiple reaction monitoring mode. The use of deuterated internal standards for most compounds verified that the accuracy of the method was not influenced by matrix effects. Extraction recoveries were 73–108% for all analytes. Lower limits of quantification ranged from 0.002 to 0.005 mg/kg. Long-term imprecision (CV%) ranged from 6.0 to 18.7%. We present a fully validated UPLC–MS–MS method for 23 benzodiazepines in whole blood with a run-time of only 5 min and using only 0.200 g of whole blood.
* Author to whom correspondence should be addressed: Section of Forensic Chemistry, Department of Forensic Medicine, Faculty of Health Sciences, University of Copenhagen, Frederik V’s vej 11, 3. DK-2100, Denmark. E-mail:
[email protected].
332
Benzodiazepines belong to the group of tranquilizers that are used to treat conditions with anxiety and unrest. Although benzodiazepines may be used as hypnotics, the so-called benzodiazepine agonists (zolpidem, zopiclone, and zaleplone) are primarily used for this purpose (z-hypnotics). Most benzodiazepines and z-hypnotics are safe when taken in larger doses, but some (e.g., clonazepam, triazolam, flunitrazepam, and zopiclone) can cause toxicity or even death (1–3). Because of their lethargic effects, benzodiazepines can be a problem when used by drivers. Other CNS depressants, such as alcohol or morphine (heroin), can potentiate or add to the effect of benzodiazepines, causing considerable impairment of driving (4), and benzodiazepines can contribute to a fatal outcome taken in combination with CNS depressants. Recently, fixed limits for blood concentrations of drugs of abuse for drivers have been introduced into the Danish legislation. Seventeen different benzodiazepines as well as zopiclone and zolpidem were considered. Thus to aid enforcement of these regulations, a fast and sensitive analytical method covering the most frequently used benzodiazepines is required. In this context, we considered development of an ultra-performance liquid chromatography–tandem mass spectrometry (UPLC–MS–MS) method for this class of compounds. We are only familiar with a single earlier paper describing separation and detection of benzodiazepines with the UPLC–MS–MS technique (5). In this method, solid-phase extraction (SPE) of 43 benzodiazepines was performed from 1 mL plasma. A gradient system with formic acid in water and acetonitrile, respectively,
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Journal of Analytical Toxicology, Vol. 34, July/August 2010
was used for separation of the benzodiazepines with a run-time of 17 min. The lower limit of quantification (LLOQ) ranged from 0.5 to 10 ng/mL. Herein, we describe our UPLC–MS–MS method for measurement in blood of 23 benzodiazepines (including zopiclone and zaleplone) and their metabolites based on liquid–liquid extraction with ethyl acetate and a short chromatographic gradient finished within 5 min.
Materials and Methods Chemicals and reagents
The following compounds were purchased from Lipomed (Bad Säckingen, Germany): bromazepam, flunitrazepam, lorazepam, nordiazepam, 7-aminonitrazepam, 7-aminoclonazepam, 7-aminoflunitrazepam, zopiclone, 7-aminoflunitrazepam-d3, and flunitrazepam-d3. Alprazolam and triazolam were received from Pfizer (Ballerup, Denmark). Temazepam and clonazepam were obtained from Roche (Hvidovre, Denmark). Oxazepam and diazepam came from Durascan Medical Products (Odense, Denmark). Midazolam and demoxepam were obtained from Alpharma (Copenhagen, Denmark). Nitrazepam and chlordiazepoxide were purchased from Nycomed Danmark (Roskilde, Denmark). The following benzodiazepines were received from different pharmaceutical companies: brotizolam (Boehringer Ingelheim, Copenhagen, Denmark), clobazam (Hoechst, Hørsholm, Denmark), estazolam (Lundbeck, Valby, Denmark), lormetazepam (Schering, Albertslund, Denmark), and zaleplone (Wyeth, Glostrup, Denmark). From Cerilliant (Round Rock, TX), we obtained the following substances: diazepam-d5, demethyldiazepam-d5, nitrazepam-d5, oxazepam-d5, alprazolam-d5, clonazepam-d4, 7-aminoclonazepam-d4, estazolam- d 5, and triazolam-d 4. 7-Aminonitrazepam-d5 and zopiclone-d8 were both obtained from Toronto Research Chemicals (Toronto, ON, Canada). All the reference standards were of ≥ 98% purity. Ethyl acetate and LC–MS-grade methanol and acetonitrile were obtained from Fisher Scientific (Leicestershire, U.K.). Boric acid was purchased from VWR (Albertslund, Denmark). Aqueous ammonia (25%) and sodium carbonate used for adjustment of the borate buffer solution were obtained from Merck (Darmstadt, Germany). Purified water was obtained with a Milli-Q system (Millipore, Copenhagen, Denmark). The mobile phase used for the LC system was prepared weekly. We performed the analyses on whole blood stabilized with sodium fluoride and potassium oxalate or CDP adenine. Most of the experiments were carried out using pooled human blood obtained from blood donors. Investigations of matrix effects and extraction efficiency were based on authentic samples negative for all kinds of licit and illicit drugs received by the department, either from autopsy cases or from living persons. The whole blood was stored at –20°C until use. Preparation of standard solutions
All standard compounds and deuterated analogs were dissolved in methanol or acetonitrile as recommended by the manufacturer to concentrations of either 100 or 1000 mg/L
and stored in ampoules at −20°C before mixing. A single working solution in purified water containing all of the compounds at concentrations of 10 mg/L (except for zopiclone, which was stored in acetonitrile and diluted separately just before extraction) were prepared every three weeks and stored at 4°C. On the day of analysis, diluted standard solutions containing all 23 compounds for spiking of calibrators and quality controls (QCs) were prepared by further dilution of the 10 mg/L stock solution in purified water. Calibrators were made by spiking 0.200 g of whole blood with 25 µL of standard solutions, yielding a final calibration range of 0.0025, 0.025, 0.25, and 0.5 mg/kg. QCs containing all compounds were prepared in pooled whole blood and stored at −20°C. An internal standard (IS) solution including all of the deuterated standards was prepared in purified water monthly and stored at 4°C. The IS solution was adjusted to a concentration of 0.125 mg/L and 30 µL of IS was added to 0.200 g of whole blood. The IS concentration was chosen based on the second calibrator concentration. The IS solution was used for all validation experiments, calibrators, QCs, and samples. LC chromatographic conditions
The chromatography was performed using an ACQUITY UPLC system (Waters, Milford, MA). The column used was an Acquity UPLC BEH C 18 (100 mm × 2.1 mm, 1.7 µm), which was maintained at a column temperature of 65°C and a constant flow rate of 0.4 mL/min. The mobile phase was composed of solvents A [0.1% aqueous ammonia (25%) in purified water] and B [0.1% aqueous ammonia (25%) in methanol]. The gradient program is shown in Table I. The injection volume was 10 µL. MS
MS was performed using a Quattro Premier XE triple quadrupole (Waters). Positive electrospray ionization mode (ESI+) was used for all MS analyses. The ionization parameters were a capillary voltage of 3.6 kV and source and desolvation temperatures of 120 and 450°C, respectively. Cone and desolvation gas (N2) flows were set at 50 and 1100 L /h, respectively. Argon was used as the collision gas at a pressure of 9.25 × 10−3 mBar, corresponding to a flow of 0.20 mL/min. Determination of the most suitable multiple reaction monitoring (MRM) transitions, cone voltages, and collision energies for all analytes and deuterated analogues were obtained by tuning on the analytes in standard solutions in the concentration of 1 mg /L dissolved in Milli-Q water and methanol (60:40, v/v). The comTable I. UPLC Gradient Program* Time (min)
%A
%B
0.0 0.20 4.00 4.50 5.00
60 60 25 5 60
40 40 75 95 40
* Total run-time 5.5 min.
333
Journal of Analytical Toxicology, Vol. 34, July/August 2010
pounds were injected into the MS using the syringe pump coupled to the UPLC system with a Tee fitting. The UPLC system delivered a constant flow of 0.4 mL /min. MassLynx 4.1 (Waters) software with automated data processing (QuanLynx) was used running in the MRM mode. The analytes were identified by two characteristic MRM transitions: their ion ratio and retention time. Tolerance was set to ± 20% for the ion ratio and ± 1% for the retention time. Quantification was performed by integration of the area under the curve from the specific MRM chromatograms of the analytes and their IS. The response (the ratio of the integrated area of the analyte and the corresponding IS) was compared to the calibration curve. The IS chosen for each analyte, retention times, and MRM transitions are shown in Table II. Sample preparation
Whole blood samples (0.200 g) were spiked with 30 µL of IS solution, and then 125 µL 0.63 M borate buffer (pH 9) was added before extraction with 1500 µL ethyl acetate. After centrifugation, 500 µL organic phase was taken, evaporated at room temperature under a stream of nitrogen, and redissolved in 250 µL 40% methanol in Milli-Q water. Method validation
Calibration curve. To determine the linearity for each compound in whole blood, we prepared the calibration curves with seven concentration points including a blank. The concentration points ranged from 0.0015 to 1.0 mg/kg. The samples were prepared in 0.200 g of whole blood spiked with 25 µL of stock solution diluted in Milli-Q water to appropriate concentrations, and subsequently, 30 µL of IS solution was added. Precision and accuracy. To evaluate precision and accuracy, we analyzed four replicates at four concentration levels on two different days. The four concentration levels analyzed were 0.002, 0.005, 0.05, and 0.5 mg/kg. A calibrator series was freshly prepared for every run based on 0.200 g of whole blood spiked with all analytes, yielding the concentration points 0, 0.0025, 0.025, 0.25, and 0.50 mg/kg. Prior to analysis, four different stock samples (5 g each) representing the four concentration levels were prepared by spiking pooled whole blank blood with all of the analytes. On day one of analysis, four samples (0.200 g of blood) were taken from each of the four stock samples. All 16 samples (four replicates for each concentration level)
334
Table II. Retention Time, MRM Transitions, and Operating Parameters for the Analyzed Drugs* Retention Time (min)
MRM Transitions (m/z)
Function 1 7-Aminoclonazepam
1.11
7-Aminoflunitrazepam
1.31
7-Aminonitrazepam
1.10
286.1 > 121.0 286.1 > 222.2 284.1 > 135.1 284.1 > 227.3 252.15 > 121.0 252.15 > 94.0
Function 2 Demoxepam
2.20
Clonazepam
2.27
Zaleplone
1.99
Function 3 Bromazepam
2.30
Nitrazepam
2.30
Zopiclone
2.37
Function 4 Flunitrazepam
Compound
Cone Collision Voltage Energy (V) (eV) 40 40 40
287.0 > 180.0 287.0 > 104.9 316.1 > 270.1 316.1 > 214.1 306.1 > 236.1 306.1 > 264.1
35
318.0 > 182.1 318.0 > 209.2 282.1 > 236.2 282.1 > 180.1 389.0 > 245.0 389.0 > 217.1
45
2.57
314.1 > 268.1 314.1 > 239.2
Function 5 Clobazam
2.85
Estazolam
2.80
Oxazepam
2.89
Function 6 Chlordiazepoxide
3.36
Nordiazepam
3.44
Lorazepam
2.89
Function 7 Alprazolam
3.05
Triazolam
3.01
Function 8 Brotizolam
3.19
Lormetazepam
3.26
Temazepam
3.16
Function 9 Diazepam
3.70
Midazolam
3.70
30 25 25 25 25 40
Internal Standard 7-Aminoclonazepam-d4 7-Aminoflunitrazepam-d3 7-Aminonitrazepam-d5
20 20 25 40 25 20
Diazepam-d5
30 25 25 35 20 35
Diazepam-d5
45
25 35
Flunitrazepam-d3
301.0 > 224.1 301.0 > 104.9 295.1 > 267.1 295.1 > 205.1 287.0 > 241.2 287.0 > 104.0
35
30 36 25 40 20 35
Diazepam-d5
300.1 > 283.1 300.1 > 227.1 271.1 > 139.9 271.1 > 164.9 321.0 > 275.1 321.0 > 229.1
25
15 25 25 28 20 33
Diazepam-d5
309.05 > 281.1 309.05 > 205.1 343.0 > 308.1 343.0 > 315.0
45
25 40 25 25
Alprazolam-d5
45 40
40 20
40 30
60 30
50
394.9 > 279.1 394.9 > 314.0 335.0 > 289.0 335.0 > 177.1 301.0 > 255.1 301.0 > 177.1
40
285.1 > 154.0 285.1 > 193.1 326.1 > 291.1 326.1 > 249.2
45
30 30
45
Clonazepam-d4 Diazepam-d5
Nitrazepam-d5 Zopiclone-d8
Estazolam-d5 Oxazepam-d5
Nordiazepam-d5 Diazepam-d5
Diazepam-d5
30 25 20 40 20 40
Diazepam-d5
25 30 25 35
Diazepam-d5
Diazepam-d5 Diazepam-d5
Diazepam-d5
* MRM transitions are listed for each analyte with quantifier transition at top and qualifier transitions below.
Journal of Analytical Toxicology, Vol. 34, July/August 2010
Table III. Validation Parameters for Benzodiazepines in Whole Blood
Analyte Alprazolam
Fixed Concentration Calibration Limit Range (mg/kg) (mg/kg)
Correlation Coefficient Polynomal (n = 7) Regression
LOD (mg/kg)
Theoretical LLOQ Concentration (mg/kg) (mg/kg)
Measured Concentration (n = 8) (mg/kg)
Bias (%)
Precision (%CV)
5 2 1 –2.7
Extraction Recovery (%)
ME (%)
13.5 10 5.6 6
85
–1.5
0.005
0–1.0
0.9976
0.375
0.0002
0.002
0.002 0.005 0.050 0.50
0.0021 0.0051 0.051 0.49
Bromazepam
0.050
0–1.0
0.9942
0.638
0.0004
0.002
0.002 0.005 0.050 0.50
0.0022 0.0044 0.046 0.45
9 –11 –9 –10
11 16.6 6.3 2.8
84
–15.2
Brotizolam
0.002
0–1.0
0.9974
0.173
0.0005
0.002
0.002 0.005 0.050 0.50
0.0016 0.0051 0.052 0.49
–19 2 3 –2.4
20.8 14.2 7 8
81
3.6
Chlordiazepoxide
0.20
0–1.0
0.9949
0.590
0.0011
0.002
0.002 0.005 0.050 0.5
0.0025 0.0056 0.058 0.54
24 11.5 15 8
19 12 7 6
79
–0.2
0–1.0
0.9924
0.827
0.0012
0.002
0.002 0.005 0.050 0.50
0.0016 0.0047 0.048 0.46
–18 –7 –5 –8
11.8 9.3 11.3 5.0
93
–3.1
`
Demoxepam
Clobazam
0.10
0–1.0
0.9968
0.058
0.0008
0.005
0.005 0.050 0.50
0.0055 0.054 0.51
10 7 3
15 4 4
89
–0.6
Clonazepam
0.005
0–1.0
0.9949
0.020
0.0007
0.002
0.002 0.005 0.050 0.50
0.0019 0.0049 0.055 0.45
–6 –2 9.7 –9.7
23 8 7 7
85
–0.9
0–1.0
0.9959
0.577
0.0003
0.002
0.002 0.005 0.050 0.50
0.0016 0.0048 0.046 0.50
–21 –5 –8.2 0
13 4 4 4.1
105
15.4
9.6 6.5 3.7 2.2
7-Aminoclonazepam
Diazepam
0.10
0–1.0
0.9962
0.610
0.0003
0.002
0.002 0.005 0.050 0.50
0.0020 0.0050 0.052 0.48
0 0 3 –4
Estazolam
0.050
0–1.0
0.9963
0.802
0.0007
0.002
0.002 0.005 0.050 0.50
0.0021 0.0054 0.053 0.47
5.6 7.2 5.5 –5.2
19 6 6.7 11
87
3.3
Flunitrazepam
0.005
0–1.0
0.9976
0.078
0.0007
0.002
0.002 0.005 0.050 0.50
0.002 0.0053 0.055 0.49
1 6 9 –2
15 15 13 8
73
–6.1
0–1.0
0.9960
0.537
0.0004
0.002
0.002 0.005 0.050 0.50
0.0019 0.0046 0.043 0.48
–5.6 –7.2 –15 –4
108
15.8
85
–3.4
7-Aminoflunitrazepam
11 2 4.1 4.2 20 9 11
82.5
–0.2
Lorazepam
0.020
0–1.0
0.9833
0.027
0.0015
0.005
0.005 0.050 0.50
0.0050 0.049 0.49
0 –2 –2.4
Lormetazepam
0.005
0–1.0
0.9951
0.112
0.0008
0.002
0.002 0.005 0.050 0.50
0.0024 0.0052 0.049 0.46
20 3 –1.2 –7
8 7.6 5 4
83
–1.3
Midazolam
0.050
0–1.0
0.9951
0.069
0.0004
0.002
0.002 0.005 0.050 0.50
0.002 0.0053 0.057 0.46
0 6.2 14.2 –8.5
4.6 4 4.2 2
82
–1.0
Table III continued on next page
335
Journal of Analytical Toxicology, Vol. 34, July/August 2010
and the calibrators were spiked with 30 µL of internal standard, as described earlier, and subjected to liquid–liquid extraction (LLE). The procedure was repeated on day two of analysis. Another spiked sample at a concentration level of 0.001 mg/kg was prepared and used for determination of LLOD. Four replicates were analyzed. This was repeated the next day, and LOD was calculated from the eight results as 3 × SD. High and low control samples were included in each run. The long-term precision was calculated from these measurements. Uncertainty (coefficient variation, CV) was calculated for all drugs as a combination of uncertainty related to drug purity, preparation of calibrator, and long-term precision: CVresult = [CVpurity2 + CVcal. preparation2 + CVlong-term precision2]½
Eq.1
Matrix effects, extraction recoveries, and ion suppression. We evaluated the matrix effects (ME) of whole blood on the peak-area responses. The experiments were performed as described by Matuszewski et al. (6,7). Two sets of six whole blood samples, obtained from six different authentic samples (autopsy cases and living cases), received by our department, and screened negative for a broad variety of drugs including ben-
zodiazepines, were extracted according to the LLE procedure. Set 1 was spiked with all analytes after extraction (B), and set 2 was spiked before extraction (C) to a corresponding concentration of 0.03 mg/kg in whole blood. All blood samples had a final concentration of 0.008 mg/L after extraction and re-solution in mobile phase. Three replicates of 0.008 mg/L reference solutions in mobile phase (A) were analyzed directly with the UPLC–MS–MS system. We calculated the ME for each analyte by comparison of the absolute peak areas. The ME results obtained in this study were calculated as follows: ME = (1− (B/A)) × 100%
Eq. 2
where A equals the peak area of standards in mobile phase and B is the peak area obtained for blank whole blood samples spiked with analytes after extraction. An ME value > 0 indicates ionization suppression and a value < 0 indicates ionization enhancement. Extraction recoveries (RE%) were calculated as the mean absolute peak areas of all six samples spiked before LLE (C) and compared with absolute peak areas from samples spiked after LLE (B):
Table III. Validation Parameters for Benzodiazepines in Whole Blood (Continued)
Analyte Nitrazepam
Fixed Concentration Calibration Limit Range (mg/kg) (mg/kg) 0.020
7-Aminonitrazepam
Correlation Coefficient Polynomal (n = 7) Regression
LOD (mg/kg)
Theoretical LLOQ Concentration (mg/kg) (mg/kg)
Measured Concentration (n = 8) (mg/kg)
Bias (%)
Precision (%CV)
Extraction Recovery (%)
ME (%)
85
0.7
0–1.0
0.9953
0.780
0.0002
0.002
0.002 0.005 0.050 0.50
0.0024 0.0056 0.0536 0.49
20 11.2 7.2 –2.4
8 9 12 10
0–1.0
0.9914
0.168
0.0004
0.002
0.002 0.005 0.050 0.50
0.0020 0.0051 0.048 0.49
0.6 2.3 –3.3 –2.7
5 9.2 9.5 10
106
11.3
Nordiazepam
0.10
0–1.0
0.9910
0.231
0.0007
0.002
0.002 0.005 0.050 0.50
0.0021 0.0050 0.051 0.47
6 0 1 –6
14.3 8.3 6.8 4.1
88
6.8
Oxazepam
0.10
0–1.0
0.9964
0.107
0.001
0.002
0.002 0.005 0.050 0.50
0.0016 0.0048 0.052 0.52
–18 –4.2 4.5 4.1
19.6 15.5 8.0 10.2
82
4.5
Temazepam
0.020
0–1.0
0.9920
0.601
0.0003
0.002
0.002 0.005 0.050 0.50
0.0020 0.0047 0.044 0.44
–1.9 –5.7 –12.6 –12.8
7.7 4 5.3 4.9
94
–1.4
Triazolam
0.002
0–1.0
0.9956
0.129
0.0006
0.002
0.002 0.005 0.050 0.50
0.0020 0.0050 0.053 0.50
0 –1 7 –0.5
19.7 5.3 2.0 5.3
85
1.5
0–1.0
0.9953
0.037
0.0005
0.002
0.002 0.005 0.050 0.50
0.0025 0.0048 0.043 0.50
23.8 –3.7 –14.6 0.3
7.3 8 13.3 15
89
–0.9
0–1.0
0.9982
0.845
0.0005
0.002
0.002 0.005 0.050 0.50
0.0023 0.0048 0.048 0.48
14 –5 –5 –5
19.8 7.6 11.9 5.7
91
Zaleplone
Zopiclone
336
0.010
–20
Journal of Analytical Toxicology, Vol. 34, July/August 2010
RE (%) = (C/B) × 100%
Eq. 3
We tested the impact of ion suppression and enhancement from ionization of components for all analytes and IS (7). The analytes and IS were injected continuously into the MS in mixtures of a maximum of six compounds, which were selected so that all of the compounds in the mix had similar responses, and
so that none of the compounds had a ∆[M+H]+ less than m/z 3. Furthermore, all of the compounds had baseline resolution chromatography. To produce a constant elevated response in both MRM channels for each analyte, the compounds were injected post-column (0.1 mg/L at a constant flow rate of 2 µL/min) using the syringe pump and “Tee-fitting” connected to the UPLC system (delivering a constant flow of 0.4 mL/min). The slightly elevated baseline responses were monitored following injection (10 µL) from the autosampler with extracted blank blood samples from the pooled drug-free volunteers. Furthermore, all analytes in mobile phase (0.13 mg/L) were injected individually from the autosampler simultaneously with the flow of analytes from the syringe pump. We performed individual injection of the analytes to evaluate possible enhancement or depression from co-eluting analytes. We then compared the acquired post-injection baseline responses to the baseline response after injection of a blank mobile phase. We also looked for interference in all traces of the analytes after the individual injection of the analytes.
Results Calibration curve
Figure 1. Quantitative transitions for a spiked whole blood sample at 0.0025 mg/kg.
We investigated the analyte/IS peakarea response ratio in whole blood (Table III). The calibration curve was fitted to a linear regression curve (1/x). The linearity was tested using a polynomial regression approach (8). Non-linearity was indicated by a coefficient for the quadratic term that deviated significantly from zero (p < 0.01). The calibration range obtained for all analytes in blood was 0.0015–1.0 mg/kg except for lorazepam and zaleplone, which had a range of 0.0025–1.0 mg/kg (Table III). Even though the calibration curves were tested linear up to 1.0 mg/kg, the correlation coefficient for all compounds was below 0.999 (Table III). In the interval 0.0025– 0.50 mg/kg, the calibration curves have correlation coefficients > 0.999, so we decided to evaluate this as the measuring interval. All of the samples with concentrations higher than the upper limit of quantification (ULOQ) were diluted with purified water (1+9). Figure 1 shows a chromatogram of a whole blood sample spiked with 0.0025 mg/kg of the compounds. Two cases with detection of sev-
337
Journal of Analytical Toxicology, Vol. 34, July/August 2010
eral benzodiazepines are presented in Figure 2 and 3. Figure 2 was an autopsy case where the cause of death was a fall from a window. The following benzodiazepines were detected: alprazolam (0.054 mg/kg), bromazepam (2.1 mg/kg), clonazepam (0.49 mg/kg), 7-aminoclonazepam (0.22 mg/kg), diazepam (0.10 mg/kg), nordiazepam (0.22 mg/kg), oxazepam (0.27 mg/kg), temazepam (0.010 mg/kg), nitrazepam (0.038 mg/kg), and 7aminonitrazepam (0.10 mg/kg). Besides the benzodiazepines, methadone, ketamine, mirtazapine, and phenobarbital were detected. Nordiazepam, oxazepam, and temazepam are presumably metabolites originating from diazepam as nordiazepam, and temazepam are not marketed as medical drugs in Denmark. Figure 3 is an example of a fixed concentration limit case. The benzodiazepines bromazepam (0.58 mg/kg), clonazepam (0.068 mg/kg), 7-aminoclonazepam (0.094 mg/kg), flunitrazepam (0.002 mg/kg), and 7-aminoflunitrazepam (0.007 mg/kg), were detected. Bromazepam and clonazepam exceeded
the fixed concentration limit. The corresponding deuterated internal standards are also shown in Figure 3. Limits of quantification, precision, and trueness
The LLOQ was determined as the lowest concentration yielding precision (CV) of ≤ 20% and bias of ±20% with fulfillment of retention time and ion ratio tolerances (9,10). LLOQ was determined to be 0.002 mg/kg for all analytes, except clobazam and lorazepam (0.005 mg/kg) (Table III). The LLOQ was low enough to be in compliance with the traffic limits and to meet the lower therapeutic limit of the compounds. The CV and accuracy were determined at four concentration levels, including the LLOQ and ULOQ and two intermediate concentrations. The CV and bias were generally accepted at a maximum of 15% (LLOQ 20%) (9). All analytes fulfilled the precision criteria at all concentration levels, except lorazepam. Lorazepam had a CV estimate exceeding the limit (35%) at the 0.002 mg/kg level (Table III). Clonazepam had a CV slightly above the criteria at LLOQ (23%). The accuracy values were satisfactory and not significantly different from the limits for all tested concentrations, except clobazam, which had a bias of 30% at the LLOQ, and chlordiazepoxide and zaleplone, which had a bias slightly above the criteria at LLOQ (24%). The LLOQ of lorazepam and clobazam was then set to 0.005 mg/kg. ME, ion suppression, extraction recoveries, and carryover
Figure 2. A case with alprazolam (0.054 mg/kg), bromazepam (2.1 mg/kg), clonazepam (0.49 mg/kg), 7aminoclonazepam (0.22 mg/kg), diazepam (0.10 mg/kg), nordiazepam (0.22 mg/kg), oxazepam (0.27 mg/kg), temazepam (0.010 mg/kg), nitrazepam (0.038 mg/kg), and 7-aminonitrazepam (0.10 mg/kg).
338
The ME provided as percentages for all analytes are listed in Table III. All analytes had ME within ±20%, and so we concluded that ME was of minor significance. Minor ME will be compensated by the deuterated IS because ME for the deuterated IS are of the same levels as corresponding analytes. The extraction recoveries (RE) were determined in six different sources of whole blood. Extraction recoveries were all above 70% (Table III). The recoveries for the IS were of the same order of magnitude as the corresponding compounds. Ion suppression was also tested by infusion experiments for all analytes using pooled whole blood from blood donors. The experiments showed that there was no major ion suppression or enhancement in whole blood. No interference was observed for co-eluting compounds. Triazolam gave rise to a signal in its corresponding deuterated IS (triazolam-d4) trace. We also found a depression performed by triazolam-d4 in the MRM trace of triazolam. These interferences between triazolam and its deuterated IS gave rise
Journal of Analytical Toxicology, Vol. 34, July/August 2010
to problems in the precision and trueness study, which could not fulfill the demands. Triazolam-d4 was therefore excluded as an IS, and diazepam-d5 was used instead. QC samples and uncertainty
The system ran very stable as indicated by the long-term impression of the controls (Table IV). At the low control level, the CVs ranged from 6.3 to 18.7%. At the high levels, the CVs ranged from 4.2 to 14.5% (N = 30–45). Uncertainty calculated as an average of the two QCs varied from 5.7 to 15.5% (Table IV).
Discussion
an LC equipped with a quatro Premier tandem MS system (Waters). The LLOQ ranged from 1 to 2 ng/mL in blood. Smink et al. (19) used SPE for extraction of 33 benzodiazepines, metabolites, zolpidem, and zopiclone in whole blood. An ion trap LC–MS technique, which was completed in 45 min, was used for identification and quantification. LLOQ ranged from 0.4 to 41.9 ng/mL with recoveries of 40–114%. A recent paper described an SPE extraction method for 22 benzodiazepines from 100 µL serum, which were then detected by an LC–MS– MS method with a run-time of 25 min (20). Thus, the LC–MS– MS technique has shown its applicability for benzodiazepine analysis in blood. Except for one method with a short run-time of 10 min (17), all these methods had rather long analysis times (25–45 min) (18–20). Low LLOQs were reported, for example, 1–2 ng/mL (18), 0.5–200 ng/mL (17), and 0.4–41.9 ng/mL (19). The next generation of LC–MS, UPLC–MS, came on the market a few years ago. The small particle size columns used give narrow peaks, making the system useful for sensitive detection and separation of many compounds in a short time. Rapid detection methods are important in the detection of drugs in blood specimens collected from potentially impaired individuals in DUI cases, which requires a quick response for
Several methods have been described for extraction and quantification or screening of one or more benzodiazepines in urine, plasma, or whole blood. The techniques used have included gas chromatography (GC) with dual column (11), GC– MS (12–15), high-performance liquid chromatography (HPLC) (16), LC–MS–MS (17–20), and UPLC–MS–MS (5). Few papers have covered a broad variety of benzodiazepines and benzodiazepine-like hypnotics (z-compounds) in whole blood. A multimethod analyzing 22 benzodiazepines in blood and urine was described by Pernay and colleagues (14). The importance of using derivatization was shown, and the use of trimethylsilylate lowered the detection threshold considerably. Recently, a GC–MS method covering the determination of 14 benzodiazepines, including zaleplone and zolpidem, in whole blood was published; after extraction of the benzodiazepines with butyl acetate, different derivatization reagents were tested (15). El Mahjoub and Staub (16) developed a column-switching HPLC–DAD method and analyzed five different benzodiazepines in serum/plasma by direct injection. The introduction of LC–MS–MS to forensic toxicology has resulted in easier sample preparation and better sensitivity. A screening and quantification method of 23 benzodiazepines, flumazenil, and z-hypnotics in plasma by use of LC–MS and APCI has been described (17). An aliquot (0.5 mL) was used for LLE with diethyl ether/ethyl acetate (1:1). The LLOQ ranged from 0.5 to 200 ng/mL, and the run-time was 10 min. Laloup et al. (18) also used LLE for extraction of 26 benzodiazepines and metabolites in blood, urine, and hair. Blood (250 µL) was extracted with 1-chlorobutane. By Figure 3. A fixed concentration limit case. Bromazepam (0.58 mg/kg), clonazepam (0.068 mg/kg), employing a gradient with a total run7-aminoclonazepam (0.094 mg/kg), flunitrazepam (0.002 mg/kg), and 7-aminoflunitrazepam (0.007 time of 35 min, these investigators were mg/kg). Bromazepam and clonazepam exceeded the fixed concentration limit. able to detect the benzodiazepines using
339
Journal of Analytical Toxicology, Vol. 34, July/August 2010
Table IV. Long-Term Precision (Quality Controls) and Uncertainty in the Measuring Interval Determined in Blank Whole Blood Compound
Tested Levels (mg/kg)
Long-Term Precision
Uncertainty (%)
Alprazolam Bromazepam Brotizolam Chloridazepoxi Demoxepam Clobazam Clonazepam 7-AMC Diazepam Estazolam Flunitrazepam 7-AMF Lorazepam Lormetazepam Midazolam Nitrazepam 7-AMN Nordiazepam Oxazepam Temazepam Triazolam Zaleplon Zopiclon
0.005, 0.050 0.05, 0.50 0.005, 0.050 0.05, 0.50 0.05, 0.50 0.05, 0.50 0.005, 0.050 0.005, 0.050 0.05, 0.50 0.05, 0.50 0.005, 0.050 0.005, 0.050 0.005, 0.050 0.005, 0.050 0.05, 0.50 0.05, 0.50 0.005, 0.050 0.05, 0.50 0.05, 0.50 0.05, 0.50 0.005, 0.050 0.005, 0.050 0.005, 0.050
10.2, 14.5 13.9, 10.8 15.7, 7.9 15.2, 10.7 15.3, 13.6 10.8, 6.0 11.4, 6.8 9.0, 8.8 6.3, 4.2 12.5, 9.9 12.2, 10.5 10.9, 9.4 18.7, 12.3 9.5, 8.7 7.0, 4.3 7.6, 5.2 11.2, 9.5 10.2, 10.59 11.8, 9.9 11.5, 8.3 15.9, 10.7 8.7, 9.0 16.1, 8.8
12.32 12.32 11.82 12.92 15.37 8.79 9.17 9.14 5.81 11.23 11.42 10.40 15.52 9.14 5.69 6.42 11.31 10.44 11.05 10.13 13.35 8.87 12.62
the investigating police, thus creating a demand for multimethods with a short run-time. As mentioned earlier, one UPLC–MS–MS method for benzodiazepines has been published (5). The method described in this paper meets the requirements for the fixed concentration limits according to the Danish law (Table III), and the LLOQs were low enough to meet the lower therapeutic limit of the compounds. Our LLE is very simple and yielded extraction recoveries higher than 70%. This feature combined with low LLOQs and a short run-time of 5 min makes our method suitable for routine traffic analysis. LLOQ was 0.002 mg/kg for all compounds except lorazepam and clobazam, which had an LLOQ of 0.005 mg/kg. We chose to validate the most common benzodiazepines and metabolites seen in our laboratory. The method may also detect more seldomly used benzodiazepines like flurazepam, halazepam, medazepam, phenazepam, prazepam, quazepam, tetrazepam, and zolazepam. Additionally, flumazenil and zolpidem are also detected by the method. Prescription and use of zolpidem is quite frequent in Denmark, but as zolpidem is analyzed in another accreditated method in our laboratory, we only screen for zolpidem with the present method.
References 1. A. Steentoft and K. Linnet. Blood concentrations of clonazepam and 7-aminoclonazepam in forensic cases in Denmark for the
340
period 2002–2007. Forensic Sci. Int. 184: 74–79 (2009). 2. A. Steentoft and K. Worm. Cases of fatal triazolam poisoning. J. Forensic. Sci. Soc. 33: 45–48 (1993). 3. H. Druid and P. Holmgren. A compilation of fatal and control concentrations of drugs in post-mortem femoral blood. J. Forensic Sci. 42: 79–87 (1997). 4. M.W. van Laar and E.R. Volkerts. Driving and benzodiazepine use. Evidence that they do not mix. CNS Drugs 10: 383–396 (1998). 5. T. Ishida, K. Kudo, M. Hayshida, and N. Ikeda. Rapid and quantitative screening method for 43 benzodiazepines and their metabolites, zolpidem and zopiclone in human plasma by liquid chromatography/mass spectrometry with a small particle column. J. Chromatogr. B 877: 2652–2657 (2009). 6. B.K. Matuszewski, M.L. Constanzer, and C.M. Chavez-Eng. Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC–MS–MS. Anal. Chem. 75: 3019– 3030 (2003). 7. T.M. Annesley. Ion suppression in mass spectrometry. Clin. Chem. 49: 1041–1044 (2003). 8. Clinical and Laboratory Standards Institute/NCCLS. Evaluation of the Linearity of Quantitative Measurement Procedures: A Statistical Approach. Approved Guideline. CLSI/NCCLS Document EP6-A. Clinical and Laboratory Standards Institute, Wayne, PA, 2003. 9. F.T. Peters, O.H. Drummer, and F. Musshoff. Validation of new methods. Forensic Sci. Int. 165: 216–224 (2007). 10. V.P. Shah, K.K. Midha, J.W.A. Findlay, H.L Hill, J.D. Hulse, I.J. McGilveray, G McKay, K.J. Miller, R.N. Patnaik, M.L. Powell, A. Tonelli, C.T. Viswanathan, and A. Yacobi. Bioanalytical method validation—a revisit with a decade of progress. Pharm. Res. 17: 1551–1557 (2000). 11. I. Rasanen, M. Neuvonen, I. Ojanperä, and E.Vuori. Benzodiazepine findings in blood and urine by gas chromatography and immunoassay. Forensic Sci. Int. 112: 191–200 (2000). 12. D. Borrey, E. Meyer, W. Lambert, C. Van Peteghem, and A.P. De Leenheer. Simultaneous determination of fifteen low-dosed benzodiazepines in human urine by solid-phase extraction and gas chromatography–mass spectrometry. J. Chromatogr. B 765: 187–197 (2001). 13. U. Staerk and W.R. Külpmann. High-temperature solid-phase microextraction procedure for the detection of drugs by gaschromatography–mass spectrometry. J. Chromatogr. B 745: 399–411 (2000). 14. S. Pernay, I. Ricordel, D. Libong, and S. Bouchonnet. Sensitive method for the detection of 22 benzodiazepines by gas chromatography–ion trap tandem mass spectrometry. J. Chromatogr. A 954: 235–245 (2002). 15. T. Gunnar, K. Ariniemi, and P. Lillsunde. Determination of 14 benzodiazepines 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 818: 175–189 (2005). 16. A. El Mahjoub and C. Staub. High-performance liquid chromatographic method for the determination of benzodiazepines in plasma or serum using the column-switching technique. J. Chromatogr. B 742: 381–390 (2000). 17. C. Kratzsch, O. Tenberken, F.T. Peters, A.A. Weber, T. Kraemer, and H. Maurer. Screening, library-assisted identification and validated quantification of 23 benzodiazepines, flumazenil, zaleplone, zolpidem, and zopiclone in plasma by liquid chromatography–mass spectrometry with atmospheric pressure chemical ionization. J. Mass Spectrom. 39: 856–872 (2004). 18. M. Laloup, M. del Mar Ramirez Fernandez, G. De Boeck, M. Wood, V. Maes, and N. Samyn. Validation of a liquid chromatography–tandem mass spectrometry method for the simultaneous determination of 26 benzodiazepines and metabolites, zolpidem and zopiclone in blood, urine and hair. J. Anal. Toxicol. 29: 616–626 (2005).
Journal of Analytical Toxicology, Vol. 34, July/August 2010
19. B.E. Smink, J.E. Brandsma, A. Dijkhuizen, K.J. Lusthof, J.J. de Gier, A.C.G. Egberts, and D.R.A. Uges. Quantitative analysis of 33 benzodiazepines, metabolites and benzodiazepine-like substances in whole blood by liquid chromatography–tandem mass spectrometry. J. Chromatogr. B 811: 13–20 (2004). 20. M. Nakamura, T. Ohmori, Y. Itoh, M. Terashita, and K. Hirano. Simultaneous determination of benzodiazepines and their metabolites in human serum by liquid chromatography–tandem
mass spectrometry using a high-resolution octadecyl silica column compatible with aqueous compounds. Biomed. Chromatogr. 23: 357–364 (2009).
Manuscript received January 26, 2010; revision received March 2, 2010.
341
