Forensic Science International113 (2000) 367–373 www.elsevier.com/ locate / forsciint

Screening and determination of benzodiazepines in whole bloodusing solid-phase extraction and gas chromatography/mass

spectrometrya , a a b a*Hiroyuki Inoue , Yoshitaka Maeno , Mineo Iwasa , Ryoji Matoba , Masataka Nagao

aDepartment of Legal Medicine, Nagoya City University Medical School, Kawasumi, Mizuho-ku, Nagoya 467-8601, JapanbDepartment of Legal Medicine, Osaka University Medical School, Yamada-oka, Suita City, Osaka 565-0871, Japan

Abstract

Benzodiazepines are one of the most widely prescribed drugs for the treatment of a wide spectrum of clinical disorders.They are used as anticonvulsants, anxiolytics, hypnotics or muscle relaxants with different duration of action. In this paper, asimple and sensitive method for the determination of benzodiazepines in whole blood using solid-phase extraction and gaschromatography/mass spectrometry (GC/MS) is described. The drugs spiked in whole blood were extracted with an OasisEHLB solid-phase extraction cartridge (Waters), which contains a copolymer designed to have a hydrophilic–lipophilicbalance. GC/MS analysis was performed using a Shimadzu QP-5000 equipped with a BPX5 capillary column (15 m30.32mm I.D., film thickness 0.25 mm, SGE). Nineteen benzodiazepines and two thienodiazepines were well separated from eachother on their SIM chromatograms and also on the TIC with the exception of oxazolam to cloxazolam separation. The blankextract from whole blood gave no peaks that interfered with all benzodiazepines and thienodiazepines on the chromatogram.The calibration curves for selected benzodiazepines with fludiazepam as an internal standard showed excellent linearity overthe concentration range 5–500 ng/ml blood with a correlation coefficients of .0.995. The detection limits ranged from 0.2to 20 ng/ml blood. The method is simple and sensitive for the determination of benzodiazepines in whole blood and seemsto be useful in the practice of forensic science. 2000 Elsevier Science Ireland Ltd. All rights reserved.

Keywords: Benzodiazepines; Solid-phase extraction; Gas chromatography–mass spectrometry

1. Introduction [1–3]. They are used as anticonvulsants, anxiolytics,hypnotics or muscle relaxants with different duration

Since the discovery of diazepam and chlor- of action. More recently, a new generation ofdiazepoxide more than three decades ago, several benzodiazepines belonging to the class of triazolo-new benzodiazepines have been introduced for the benzodiazepines, such as triazolam and estazolam,treatment of a wide spectrum of clinical disorders have gained widespread use due to their potent

action at low doses. They are also abused and oftenused for suicidal and criminal purposes [4,5].

*Corresponding author. First Chemistry Section, National Re- Current methods for the determination of the drugssearch Institute of Police Science, 6-3-1 Kashiwanoha, Kashiwa,

available in clinical and forensic toxicology includeChiba 277-0882, Japan. Tel.: 181-471-35-8001; fax: 181-471-gas chromatography (GC) [6,7], gas chromatog-33-9159.

E-mail address: inoue@nrips.go.jp (H. Inoue). raphy/mass spectrometry (GC/MS) [8,9], high-per-

0379-0738/00/$ – see front matter 2000 Elsevier Science Ireland Ltd. All rights reserved.PI I : S0379-0738( 00 )00226-7

368 H. Inoue et al. / Forensic Science International 113 (2000) 367 –373

formance liquid chromatography (HPLC) [10] and 2.2. Preparation of standard solutionsHPLC/MS [11]. Prior to application of these in-strumental analyses, drugs are normally extracted All drug stock solutions were prepared by dissolv-with an appropriate procedure to remove possible ing 25 mg of the individual compound in 25 ml ofinterfering materials. In recent years, the solid-phase methanol (1 mg/ml). Three different working stan-extraction approach, in which solid-phase adsorbents dards in methanol were used in this study. Workingare used as the media for retaining the compounds of standard solution A contained medazepam,interest, followed by selective elution, has been flutazolam, oxazolam, midazolam, bromazepam,intensely developed and used due to its several haloxazolam, flurazepam, clonazepam, alprazolamadvantages such as less organic solvent usage, no and triazolam. Solution B contained diazepam, chlor-foaming problems, shorter sample preparation time, diazepoxide, cloxazolam, flunitrazepam, nimetaze-etc. [12]. There are several reports on the solid-phase pam, mexazolam, nitrazepam, estazolam, etizolamextraction of benzodiazepines in urine, serum and and brotizolam. Solution C contained medazepam,plasma [8,9,11]. However, these would not be ap- diazepam, chlordiazepoxide, midazolam, flunitraze-plied directly to whole blood encountered commonly pam, nimetazepam, mexazolam, estazolam, al-in forensic laboratories. prazolam and triazolam, and was used for prepara-

OasisE HLB is a new polymeric reversed-phase tion of calibration curves.sorbent for solid-phase extraction designed to have ahydrophilic–lipophilic balance. We describe here a 2.3. Extraction proceduremethod for the screening and determination ofbenzodiazepines in whole blood using an Oasis HLB Samples of outdated whole blood were obtainedcartridge and GC/MS. from Nagoya City University Hospital. One milliliter

of blood containing a known amount of benzo-diazepines (added as a methanol solution) was mixed

2. Materials and methods with 4 ml of distilled water. The mixture was thencentrifuged at 10,000 g for 20 min and the superna-

2.1. Chemicals tant was mixed with 3 ml of 0.05 M phosphatebuffer, pH 7.0.

Diazepam, fludiazepam, medazepam, clonazepam Oasis HLB cartridges were pretreated by passing 5and nimetazepam were obtained from Sumitomo ml of CH Cl , 5 ml of methanol and 5 ml of distilled2 2

Chemical (Osaka, Japan); flutazolam from Mitsui water. The sample solution was poured into thePharmaceuticals (Tokyo, Japan); chlordiazepoxide pretreated cartridge at a flow-rate of less than 1and estazolam from Takeda Chemical (Osaka, ml /min. The cartridge was then washed with 3 ml ofJapan); oxazolam, cloxazolam, haloxazolam and distilled water. Finally, benzodiazepines were elutedmexazolam from Sankyo (Tokyo, Japan); with 5 ml of CH Cl . Eluates were centrifuged at2 2

midazolam, flunitrazepam, bromazepam and 1600 g for 5 min at 48C. The remaining aqueousflurazepam from Nippon Roche K.K. (Tokyo, layer was discarded with a pipette and the organicJapan); nitrazepam, alprazolam and triazolam from layer was evaporated to dryness at 458C under aUpjohn (Kalamazoo, MI, USA); etizolam from gentle stream of nitrogen. The residue was dissolvedYoshitomi Pharmaceutical (Osaka, Japan); and in 100 ml of methanol containing, if necessary, 100brotizolam from Nippon Boehringer Ingelheim ng fludiazepam as internal standard.(Kawanishi, Japan).

Oasis HLB extraction cartridges (30 mg) were 2.4. Gas chromatography /mass spectrometry (GC /obtained from Waters (Milford, MA, USA). Di- MS)chloromethane (CH Cl ), methanol and other re-2 2

agents were purchased from Wako (Osaka, Japan). GC/MS analysis was performed using a ShimadzuThe chemical structures of the 19 benzodiazepines QP-5000 (Kyoto, Japan) equipped with a BPX5and two thienodiazepines investigated are illustrated capillary column (15 m30.32 mm I.D., film thick-in Fig. 1. ness 0.25 mm, SGE, Australia). The oven tempera-

H. Inoue et al. / Forensic Science International 113 (2000) 367 –373 369

Fig. 1. Structures of benzodiazepines and thienodiazepines used in this study.

ture was held at 608C for 2 min following injection standard solutions (50 mg/ml in methanol) wereand programmed to 3008C at a rate of 208C/min. injected using the same GC conditions, and the massThe temperature of the injection port and interface spectrometer was operated with a scan mass range ofwas set at 2808C. Helium was used as the carrier gas 40 to 400 atomic mass units.(head pressure 20 kPa, total flow-rate 50 ml /min).The mass spectrometer was operated under electronimpact (EI) mode at an ionization energy of 70 eV. 3. Results and discussionTwo microliters of the samples were injected in thesplitless mode at a column temperature of 608C, and Fig. 2 illustrates the SIM chromatograms of anthe splitter was opened after 1 min. authentic standard solution and extracts from whole

The following masses were monitored by selected blood spiked with working standards A and B.ion monitoring (SIM): m /z 242, 251, 253, 283, 284, Nineteen benzodiazepines and two thienodiazepines289, 294, 302, 305, 308, 310, 313, 315, 387 and 394. were well separated on their SIM chromatogramsTo obtain mass spectra of authentics, 2 ml of and also on the total ion chromatogram (TIC) with

370 H. Inoue et al. / Forensic Science International 113 (2000) 367 –373

Fig. 2. SIM chromatograms of (a) an authentic standard solution containing benzodiazepines and thienodiazepines without extraction(fludiazepam, 2 ng; others, 4 ng each on column), (b) an extract from whole blood spiked with working standard A (200 ng/ml each) and (c)an extract from whole blood spiked with working standard B (200 ng/ml each). 1, medazepam; 2, fludiazepam; 3, diazepam; 4, flutazolam;5, chlordiazepoxide; 6,69, oxazolam; 7, cloxazolam; 8, midazolam; 9, flunitrazepam; 10, bromazepam; 11, nimetazepam; 12, haloxazolam;13, mexazolam; 14, flurazepam; 15, nitrazepam; 16, clonazepam; 17, estazolam; 18, alprazolam; 19, etizolam; 20, triazolam; 21, brotizolam.Chromatographic conditions: column, BPX5 capillary column, 15 m30.32 mm I.D.; film thickness, 0.25 mm; temperature, column 608C (2min) to 3008C (208C/min), injection port 2808C, interface 2808C; carrier gas, He at 20 kPa; total flow rate, 50 ml /min; electron-impactmode at an ionization energy of 70 eV. Two microliters of the samples were injected in the splitless mode at a column temperature of 608C,and the splitter was opened after 1 min.

the exception of oxazolam–cloxazolam separation. peaks that interfered with all benzodiazepines andOxazolam gave two peaks that had the same mass thienodiazepines on the chromatograms. A peak at aspectra. The blank extract from whole blood gave no retention time of 13.9 min appearing on the chro-

H. Inoue et al. / Forensic Science International 113 (2000) 367 –373 371

81matogram obtained from blood extracts was con- Br51:1). The most effective and reliable approachfirmed to be cholesterol by obtaining the EI full scan to identify peaks in the chromatogram from unknownmass spectrum. samples is to obtain full mass spectra of the peaks

Table 1 shows base and other principal ions of the together with coincidence of retention times with thecompounds measured using EI scan mode after GC authentics. In some cases in forensic practice, how-separation. Ions for monitoring each compound were ever, it may be impossible to obtain a full masschosen from among characteristic ions of more than spectrum of a suspected peak due to small con-200 amu, in order to avoid interference derived from centrations. It would be helpful to compare theblood extracts. All oxazolobenzodiazepines (ox- relative intensities of characteristic ions in SIM forazolam, cloxazolam, mexazolam, flutazolam and peak identification.haloxazolam) and chlordiazepoxide gave no molecu- Table 2 shows the retention times and recoverieslar ions under these analytical conditions, while the of benzodiazepines and thienodiazepines used in theothers gave them at varying relative intensities. Most study. The reproducibility of the retention times wasbenzodiazepines, having halogen in their chemical excellent with ,0.03% of the coefficients of vari-structures, displayed characteristic isotope ions along ation. The recoveries of the drugs added to wholewith major fragment peaks. For example, diazepam, blood were satisfactory with the exception of somehaving one chlorine in the structure, the ion of m /z oxazolobenzodiazepines (oxazolam, cloxazolam and286 appeared with that of m /z 284, and the ratio of haloxazolam), which ranged from 11 to 55% in thethe relative intensities was 1:3, which corresponds to samples with 200 ng/ml added.

37 35the natural ratio of the isotopes ( Cl / Cl). Similar- The calibration curves obtained by plotting thely, for bromazepam, having one bromine, the relative peak area ratios of each benzodiazepine to theintensities of m /z 315 and 317 in the spectrum internal standard (fludiazepam) versus the amount of

79obtained from the compound were equal ( Br / each benzodiazepine (medazepam, diazepam, chlor-

Table 1aPrincipal ions of benzodiazepines used in the study

Compound m /z (% intensity)

Monitor Others1Diazepam 284 M (75) 256 (100), 283 (96), 221 (44), 286 (24)1Fludiazepam 302 M (66) 274 (100), 301 (80), 283 (31), 304 (22)1Flurazepam 387 M (1) 86 (100), 58 (16), 99 (11), 87 (11)

1Medazepam 242 (85) 207 (100), 244 (31), 270 M (19), 208 (18)Chlordiazepoxide 284 (44) 282 (100), 283 (77), 241 (14), 253 (12)

1Bromazepam 315 M (62) 236 (100), 77 (99), 317 (60), 288 (57),1Nitrazepam 253 (97) 206 (100), 280 (94), 234 (80), 281 M (44)

1Clonazepam 315 M (52) 280 (100), 314 (77), 234 (60), 286 (57)1Nimetazepam 294 (82) 267(100), 248 (69), 295 M (48), 220 (48)

1Flunitrazepam 313 M (59) 286 (100), 312 (94), 266 (59), 238 (58)Oxazolam 251 (100) 70 (56), 253 (33), 202 (9), 283 (6)Cloxazolam 305 (100) 262 (63), 226 (59), 307 (58), 270 (40)Mexazolam 251 (100) 70 (36), 253 (34),139 (16), 236 (11)Flutazolam 289 (100) 149 (86), 210 (80), 245 (77), 246 (70)Haloxazolam 283 (99) 281 (100), 56 (91), 206 (26), 347 (22)

1Estazolam 294 M (42) 259 (100), 205 (78), 239 (48), 231 (21)1Alprazolam 308 M (50) 279 (100), 204 (96), 273 (64), 281 (34)

1Triazolam 313 (100) 238 (72), 315 (65), 342 M (38), 344 (25)1Midazolam 310 (100) 312 (34), 325 M (21), 311 (20), 297 (11)

1Etizolam 313 (39) 342 M (100), 344 (38), 266 (34), 224 (28)1Brotizolam 394 (100) 245 (87), 392 M (75), 210 (54), 316 (36)

a 1M , molecular ion.

372 H. Inoue et al. / Forensic Science International 113 (2000) 367 –373

Table 2aRetention times and recoveries of benzodiazepines and thienodiazepines

Compound Retention Recovery (%)time (min)

10 ng/ml 200 ng/ml

Medazepam 10.70 (0.03) 71.7 (5.5) 70.4 (9.6)Fludiazepam 11.35 (0.02) – – – –Diazepam 11.60 (0.02) 95.9 (6.0) 68.7 (4.9)Flutazolam 11.67 (0.03) 82.1 (22.8) 120.6 (11.2)Chlordiazepoxide 11.98 (0.02) 128.2 (4.8) 86.4 (4.6)Oxazolam 12.16 (0.02) 61.1 (7.2) 54.5 (8.9)Cloxazolam 12.19 (0.02) n.d. n.d. 11.3 (13.0)Midazolam 12.23 (0.02) 101.7 (9.6) 87.3 (7.5)Flunitrazepam 12.33 (0.02) 118.9 (6.7) 92.0 (2.8)Bromazepam 12.46 (0.04) 138.7 (18.3) 124.4 (5.8)Nimetazepam 12.62 (0.02) 91.4 (9.2) 90.2 (2.8)Haloxazolam 12.74 (0.02) 44.7 (18.4) 34.0 (16.6)Mexazolam 12.84 (0.01) 70.2 (10.6) 78.3 (3.9)Flurazepam 12.88 (0.02) 93.0 (15.0) 89.3 (8.3)Nitrazepam 13.02 (0.02) 124.1 (12.3) 113.6 (1.8)Clonazepam 13.31 (0.02) 104.6 (28.8) 118.5 (7.0)Estazolam 13.53 (0.02) 98.3 (15.9) 95.8 (2.1)Alprazolam 13.65 (0.01) 108.3 (9.9) 97.0 (5.6)Etizolam 13.95 (0.02) 87.8 (16.3) 94.4 (3.8)Triazolam 13.98 (0.02) 115.6 (12.1) 97.0 (5.0)Brotizolam 14.12 (0.02) 123.9 (4.0) 97.5 (3.4)

a Values in parentheses represent coefficients of variation (%) from four determinations. n.d., not detected.

diazepoxide, midazolam, flunitrazepam, Three milliliters of water was sufficient to wash thenimetazepam, mexazolam, estazolam alprazolam and cartridge, and the use of CH Cl for elution of the2 2

triazolam) showed excellent linearity over the con- analytes made it easy to remove the aqueous portion,centration range 5–500 ng/ml blood with correlation which eluted with CH Cl , using a pipette.2 2

coefficients of .0.995. The detection limits were 0.2 The present method detected thermolabile triazolo-ng/ml blood for medazepam and midazolam, 1 ng/ benzodiazepines (estazolam, alprazolam and tri-ml for diazepam, chlordiazepoxide, mexazolam and azolam) and differentiated triazolam from etizolamalprazolam, 2 ng/ml for flunitrazepam, [14]. The method is simple and sensitive for thenimetazepam, estazolam, triazolam and brotizolam, 5 determination of benzodiazepines in whole blood andng/ml for flutazolam, oxazolam, bromazepam, nit- seems to be useful in the practice of forensic science.razepam and etizolam, and 20 ng/ml for cloxazolam,haloxazolam, flurazepam and clonazepam. The sen-sitivity allows for screening and determination ofbenzodiazepines at therapeutic doses [13]. Acknowledgements

According to the manufacturer’s recommendedextraction procedure, 1 ml of 5% methanol in water The authors wish to thank Mr. Takahirois used to wash the HLB cartridges after sample Kobayashi, Nagoya City University Medical School,loading. In the case of whole blood, however, it for his technical assistance and Sumitomo Chemical,seemed to be insufficient to remove heme absorbed Mitsui Pharmaceuticals, Takeda Chemical Ind.,in the cartridge, even if 3 ml of the wash solution Sankyo, Nippon Roche K.K., Upjohn, Yoshitomiwas used. A large quantity of heme coeluted with Pharmaceutical Ind. and Nippon Boehringer Ing-analytes in the subsequent methanol elution step. elheim for donations of drug standards.

H. Inoue et al. / Forensic Science International 113 (2000) 367 –373 373

Positive- and negative-ion mass spectrometry of 24 benzo-Referencesdiazepines, Forensic Sci. Int. 35 (1987) 165–179.

[9] D.A. Black, G.D. Clark,V.M. Haver, J.A. Garbin, A.J. Saxon,[1] S.E. File, S. Pellow, Behavioral pharmacology of minor Analysis of urinary benzodiazepines using solid-phase ex-

tranquilizers, in: D.J.K. Balfour (Ed.), Psychotropic Drugs of traction and gas chromatography–mass spectrometry, J.Abuse, Pergamon Press, New York, 1990, pp. 147–172. Anal. Toxicol. 18 (1994) 185–188.

[2] M.J. Ellenhorn, Benzodiazepines, in: Ellenhorn’s Medical [10] C.M. Moore, K. Sato, Y. Katsumata, Rapid monitoring ofToxicology, 2nd Edition, Williams & Wilkins, Baltimore, benzodiazepines in clinical samples by using on-line column1997, pp. 687–695. switching HPLC, Clin. Chem. 37 (1991) 804–808.

[3] M. Mazhar, S.R. Binder, Analysis of benzodiazepines and [11] M. Kleinschnitz, M. Herderich, P. Schreier, Determination oftricyclic antidepressants in serum using a common solid- 1,4-benzodiazepines by high-performance liquid chromatog-phase clean-up and a common mobile phase, J. Chromatogr. raphy–electrospray tandem mass spectrometry, J. Chroma-497 (1989) 201–212. togr. B 676 (1996) 61–67.

[4] A. Mori, E. Takeda, H. Suzuki, I. Ishiyama, Case reports of [12] R.H. Liu, D.E. Gadzala, Sample pretreatment, in: Handbookmurder involving benzodiazepine drugs, Act. Crim. Japon. of Drug Analysis, American Chemical Society, Washington,59 (1993) 200–206. DC, 1997, pp. 35–53.

[5] K. Uemura, S. Komura, Death caused by triazolam and [13] M.R. Repetto, M. Repetto, Habitual, toxic, and lethal con-ethanol intoxication, J. Forensic Med. Pathol. 16 (1995) centrations of 103 drugs of abuse in humans, Clin. Toxicol.66–68. 35 (1997) 1–9.

[6] K. Kudo, T. Nagata, T. Imamura, S. Kage, Y. Hida, Forensic [14] H. Inoue, Y. Maeno, M. Iwasa, J. Monma, R. Matoba, Use ofanalysis of triazolam in human tissues using capillary gas high-performance liquid chromatography as an extractionchromatography, Int. J. Leg. Med. 104 (1991) 67–69. procedure for analysis of triazolam in decomposed human

[7] T. Edeki, D.W. Robin, C. Prakash, I.A. Blair, A.J.J. Wood, muscle by gas chromatography–mass spectrometry, J. Chro-Sensitive assay for triazolam in plasma following low oral matogr. B 701 (1997) 47–52.doses, J. Chromatogr. 577 (1992) 190–194.

[8] H. Hattori, O. Suzuki, K. Sato, Y. Mizutani, T. Yamada,