Simultaneous Analysis of 14 Benzodiazepines in Oral Fluid by Solid-Phase Extraction and LC-MS-MS

Simultaneous Analysis of 14 Benzodiazepines in Oral Fluid by Solid-Phase Extraction and LC-MS-MS George Ngwa Chemistry, Lehigh University, 6 E. Packer Ave., 5eeley G. Mudd Building, Bethlehem, Pennsylvania 18015 Dean Fritch and Kristen Blum OraSure Technologies, Research and Development, Bethlehem, Pennsylvania Gregory Newland Applied Biosystems, Applied Markets/Small Molecules, 21 S. 8th St., Bangor, Pennsylvania 18013 Abstract A simple and rapid procedure for the simultaneous screening of 14 benzodiazepines in oral fluid is presented. The procedure involves a solid-phase extraction followed by liquid chromatography-tandem mass spectrometry (LC-MS-MS). The target compounds include diazepam, oxazepam, temazepam, nordiazepam, Iorazepam, chlordiazepoxide, alprazolam, cc-hydroxyalprazolam, desalkylflurazepam, hydroxyethylflurazepam, clonazepam, 7-aminoclonazepam, flunitrazepam, and 7-aminoflunitrazepam. Oral fluid was obtained using a simple device that collects approximately 0.4 ml of oral fluid and dilutes it with 0.8 ml of preservative. The oral fluid sample preparation involves a solid-phase extraction on a Varian Bond Elut cartridge. Quantitation was performed by LC-MS-MS using nordiazepam-d 5 as the internal standard. The extraction efficiency exceeded 83% for all compounds except for 7-aminoclonazepam, which had a recovery of 55%. The limits of quantitation ranged from 0.1 ng/ml to 1.0 ng/ml in the multiple reaction monitoring mode. This method was used to confirm 41 patients that screened positive using the OraSure Technologies Benzodiazepine Intercept MICRO-PLATE Enzyme Immunoassay kit. All screened-positive patients were positive for at least one of the analyzed benzodiazepines, thus showing that this method is suitable for confirmation of the Intercept Benzodiazepine assay. Introduction Benzodiazepines are widely prescribed drugs used as anxiolytics, sedative hypnotics, anticonvulsants, and muscle relaxants. Apart from their therapeutic applications, benzodiazepines are often abused by drug addicts (1). As a consequence, this class of drugs and their metabolites are frequently present in both clinical and forensic cases. Although the first benzodiazepines were first marketed over 40 years ago, the screening for highly potent compounds remains an ana- lytical challenge. Indeed, low nanogram levels of benzodiazepines and their metabolites often remain undetected by traditional enzyme-multiplied immunoassay (EMIT) or cannot be confirmed by high-performance liquid chromatographic (HPLC) or gas chromatographic (GC) screening techniques (2-4). Several factors complicate the analysis of benzodiazepines. A first complicating factor is the wide range of therapeutic doses used for these compounds. An overview of some types of benzodiazepines, their therapeutic dose and corresponding plasma concentration has been presented by different authors (5,6). h second complicating factor is the chemical heterogeneity of benzodiazepines. All benzodiazepines display some structural similarity, but they show varied chromatographic behavior because of wide polarity differences. These differences limit the number of compounds that can be successfully chromatographed simultaneously. Although numerous methods for the determination of benzodiazepines exist, few of them can be applied for the simultaneous analysis of both parent compounds and their corresponding metabolites (7-11). Even though urine, blood, hair, sweat, and oral fluid have been used as matrices for monitoring drugs of abuse, there is a lack of information on the use of oral fluid for the testing and analysis of benzodiazepines in the literature (12-17). Much of the work that has been reported on the analyses of benzodiazepines involves urine and plasma (18-21). Analysis of oral fluid for benzodiazepines has been limited primarily because of small specimen size and low drug concentrations (22). Analysis of oral fluid for benzodiazepines will be useful, however, because of the presence of parent drugs rather than metabolites as frequently found in urine. Detection of parent drugs in oral fluid would be a better indication of recent use and establish a higher probability that the subject (in the case of impaired drivers) is experiencing pharmacological effects at the time of sampling. Moreover, the noninvasive collection of R,..pr~du(tion (phot(~copying) o( edit()rial (ontent ot this journal is prohibited without publisher's permission. 369

oral fluid sample, which is relatively easy to perform, can be achieved under close supervision and without the invasion of privacy. Despite the advantages of oral fluid as a screening matrix for benzodiazepines, the literature appears to be sparse on the use of this matrix. The works of Kintz et al. (16) and Quintela et al. (17) include a few of the papers that describe an LC-MS-MS method for the screening of benzodiazepines in oral fluid. The run time for the analysis by Kintz et al. (16) was approximately 18 min. The purpose of this work was to establish a simple, rapid (shorter run times), sensitive, and novel LC-MS-MS confirmation procedure for the simultaneous analysis of 14 commonly abused benzodiazepines in oral fluid. This assay would provide a valid method for confirmation of screening results in conjunction with the use of OTI Benzodiazepine Intercept MICRO-PLATE enzyme immunoassay (EIA). The confirmation assay was based on a combination of retention time, presence of quantifier and qualifier ions as well as a quantifier/qualifier ion ratio for each of the analytes. Also a fast and automated solid-phase extraction with high recoveries for these drugs is presented. The method was developed for those compounds that are most representative of benzodiazepines present in the United States, although many other benzodiazepines could be analyzed under the same conditions. Experimental Solvents, reagents, and columns Ethyl acetate was purchased from J.T. Baker (Phillipsburg, N J). Acetonitrile and hexane were purchased from G.J. Chemical (Newark, NJ). Methanol and glacial acetic acid were purchased from VWR International (West Chester, PA). Potassium phosphate (monobasic) and sodium acetate were purchased from Sigma Aldrich (St. Louis, MO). One-milliliter ampoules of 1 mg/ml of diazepam (in methanol), oxazepam (in methanol), temazepam (in methanol) nordiazepam (in methanol), lorazepam (in acetonitrile), chlordiazepoxide (in methanol), alprazolam (in methanol), hydroxyalprazolam (in methanol), clonazepam (in methanol), flunitrazepam (in methanol), hydroxyethylflurazepam (in methanol), desalkylflurazepam (in methanol), 7-aminoclonazepam (in acetonitrile), 7-aminoflunitrazepam (in methanol), nordiazepam-d5 (in methanol and used as internal standard), S(+)-amphetamine (in methanol), S(+)-methamphetamine (in methanol), benzoylecgonine (in methanol), (-)-A~-THC (in methanol), morphine (in methanol), codeine (in methanol), (_+)-MDA (in methanol), (_+)-MDEA (in methanol), and (+)- MDMA were purchased from Cerilliant (Round Rock, TX). The Varian Certify Bond Elut (3 ml, 130 rag) and Zorbax Bonus-RP columns were obtained from Varian Technologies (Palo Alto, CA) and Agilent Technologies (Santa Clara, CA), respectively. Preparation of standards and buffers A working solution containing 200 ng/ml of all the drugs was prepared by diluting the 1 mg/ml stock in a methanol/water mixture (50:50 v/v). This standard mixture was further diluted in methanol/water (50:50, v/v) to yield working solutions of 50 ng/ml and 10 ng/ml to be used for spiking the calibrators. All standards and working solutions were stored at +4~ Sodium acetate buffer (0.1M, ph 5.0) for extraction was prepared using sodium acetate. The solution was adjusted to ph 5.0 with 1M acetic acid. Phosphate buffer (0.1M, ph 6.0) for extraction was prepared using KH2PO4. The solution was adjusted to ph 6.0 with 1M KOH. Phosphate buffer (0.1M, ph 8.4) was prepared as described except that the ph was brought to 8.0 with 1M KOH. Specimens Oral fluid diluent (a synthetic oral fluid) prepared by Ora- Sure Technologies (Bethlehem, PA) was used to prepare calibrators for the analysis. Calibrators at 0.1, 0.5, 1.0, 5.0, 10.0, and 20.0 ng/ml were prepared by spiking blank oral fluid diluent with the appropriate working solutions. To evaluate for matrix effect and ion suppression, blank oral fluids were collected from six laboratory personnel. Using the OraSure Intercept device, these volunteers held one pad in each side of their mouth for three minutes after which the pads were placed into the blue OraSure diluent in a plastic tube. The pads along with the diluent were centrifuged for 10 min at 3000 rpm before use. It should be noted that the Intercept device collects an average of 0.38 + 0.19 (SD)with a range of 0.05 to 0.8 ml of oral fluid. Given the fact that the collected oral fluid is diluted with 0.8 ml of preservative solution before analysis, a dilution factor of 1:3 is arbitrarily accepted (23). Positive specimens were obtained from subjects in a drug treatment center. The specimens were analyzed with the OTI Intercept MICRO-PLATE enzyme immunoassay kit following manufacturer's procedures. Each specimen was analyzed in duplicate. Quality control samples (below cutoff control and above cutoff control) were utilized in all cases. Specimens with mean absorbance less than or equal to the calibrator were considered positive and specimens with responses greater than the calibrator were considered negative. The samples were screened at a drug treatment center, then stored at room temperature for three days and ready to be discarded. At this point, the samples that screened positive for benzodiazepines by EIA were transferred to tubes with no identification other than the absorbance values of the screen. These positive samples were stored frozen for one year before analysis by the LC-MS-MS method. Extraction Oral fluid diluent or patient sample (0.4 ml) was added to 3.6 ml of 0.1M acetate buffer (ph 5). Twenty-five microliters of 200 ng/ml nordiazepam-d5 as internal standard was added to all calibrators and patient samples. The Zymark RapidTrace extraction instrument from Caliper (Princeton, N J) was used for the automation of the rest of the extraction process as follows: condition the Varian Certify Bond Elut columns with methanol followed by distilled water; load sample at a rate of 2 ml/min; rinse with water followed by 1.3 ml of a mixture of 370

acetonitrile/0.1m phosphate buffer (ph 6, 20:80, v/v); dry for 5 rain and elute with 2.6 ml of ethyl acetate; dry down extracts under a stream of nitrogen; and reconstitute samples in 100 IJL of a methanol/water mixture (50:50, v/v). LC-MS-MS procedure Chromatography. The HPLC system was an Agilent 1100. Analytes were separated on a Zorbax Bonus-RP (2.1 x 100 ram, 3.5 micron, C14) column. Thirty microliters of the extract was injected onto the column. The mobile phase, delivered at a flow rate of 200 IJL/min at 22~ was a gradient of methanol in 0.1% formic acid programmed as follows: 50% methanol for 0.50 min, increased linearly to 95% methanol over the next 3.0 rain where it was held constant for 1.30 rain before returning to the initial conditions within 1.0 min and equilibrating for 2.20 min, which resulted in a total run time of 8.0 rain. MS. Detection was performed on an API 3200 triplequadrupole instrument from Applied Biosystems MDS SCIEX. Ionization was achieved using APCI in positive mode. The optimum conditions were curtain gas, 50.00 psi; collisionally activated dissociation (CAD), 5.00 psi; heated nebulizer tem- perature (TEM), 650.00~ nebulizing gas (GS1), 35.00 psi; and heater gas (GS2), 60.00 psi. In order to establish the appropriate multiple reaction monitoring (mrm) conditions for the individual compounds, solutions of standards at 500 ng/ml in methanol/water (50:50, v/v) were infused into the MS, and the following parameters were optimized for the different ions: declustering potential (DP), entrance potential (EP), collision energy (CE), and collision exit potential (CXP). Data acquisition, peak integration, and calculation were interfaced to a computer workstation running on the Analyst 4.1 software. The parent ions, the corresponding daughter ions, the DP, EP, CE, and CXP for the benzodiazepines and internal standard are presented in Table I. Method Validation Standard calibration curves were obtained using oral fluid diluent spiked with the 14 compounds at concentration levels of 0.1, 0.5, 1.0, 5.0, 10.0, and 20.0 ng/ml. In order to establish that the use of the collection device did not affect the obtained resuits, three short experiments were conducted. In the first experiment 0.4 ml of oral fluid diluent (synthetic oral fluid) was spiked with 3 ng/ml of all the analyzed drugs. In the second and third experiments, 5 ml of saliva (collected by spitting) was spiked with 3 ng/ml of all the analyzed benzodiazepines. A Drug 0.4-mL portion of this was removed for analysis, while another 0.4-mL portion was pipetted onto the collection pad before placing in 0.8 ml of the blue solution in the collection vial. All three portions were stored at room temperature for three days, after which they were extracted and analyzed by the developed method. The LOQ was taken as the concentration of the drug resulting in a signal 10 times the standard deviation of the blank. The limit of detection was determined as the lowest concentration of an analyte that the method could reliably differentiate from background noise and corresponded to the lowest concentration of the drug resulting in a signal-to-noise ratio equal or greater than 3:1 for both quantifier and qualifier ions. Within-batch precisions and accuracies (n = 6) were determined using the oral fluid diluent (synthetic oral fluid) spiked with the 14 compounds at 0.1 and 10.0 ng/ml. Recovery of the compounds from oral fluid diluent was determined in triplicate at two different concentrations (low and high) by comparing values obtained with unextracted and extracted standards. For each concentration, three blank samples were spiked with the internal standard and the appropriate Table I. MRM Transitions and Conditions for All Compounds and the Internal Standard* Precursor Product DP ~ EP CEP CE CXP Dwell Ion (m/z) Ion (m/z) (V) (V) (V) (V) (V) (ms) Flunitrazepam 314 268 60 10 15 34 20 50 239 60 10 15 44 20 50 Oxazepam 287 241 38 5 15 31 5 50 269 38 5 15 20 5 50 Diazepam 285 193 55 6 15 42 14 50 154 55 6 15 36 14 50 Lorazepam 321 275 42 6 15 26 10 50 229 42 6 15 26 10 50 7-Aminoclonazepam 286 121 38 6 15 31 7 50 222 32 6 15 20 7 50 Desalkylflurazepam 289 140 59 5 14 40 11 50 226 59 5 14 40 11 50 Clonazepam 316 270 60 5 15 34 22 50 214 60 5 15 50 22 50 ~-Hydroxyalprazolam 326 297 60 5 15 34 25 50 216 60 5 15 55 25 50 Alprazolam 309 381 64 5 15 35 20 50 205 64 5 15 52 20 50 Temazepam 301 255 37 6 14 30 23 50 283 37 6 14 17 23 50 Chlordiazepoxide 300 283 39 6 14 19 20 50 227 36 6 14 34 20 50 7-Aminoflunitrazepam 284 135 55 4 14 39 14 50 226 55 4 14 39 14 50 Hydroxyethylflurazepam 333 211 55 7 15 51 15 50 305 55 7 15 28 15 50 Nordiazepam 271 140 55 7 14 41 6 50 91 55 7 14 51 4 50 Nordiazepam-ds 276 140 55 7 14 51 4 50 * Quantifier ions are shown in bold. t Abbreviations: DP, declustering potential; EP, entrance potential; CEP, collision entrance potential; CE, collision energy; and CXP, collision exit potential. 371

amount of each drug (0.1 ng/ml and 10 ng/ml) and three others were spiked only with the internal standards. They were extracted as previously described. The dried extracts of the samples with the drugs were reconstituted with 100 pl of reconstitution solvent [methanol/water, 50:50 (v/v)], while the others were reconstituted with 100 pl of the reconstitution solvent containing the appropriate amount of each drug (0.1 ng/ml and 10.0 ng/ml). The latter were used as unextracted standards. The specificity of the method was determined by analyzing oral fluid from six non-drug consuming subjects. Oral fluid collections were performed with the OraSure Intercept device. Additionally, 1000 ng/ml each of a panel of nine drugs [S(+)-amphetamine (in methanol), S(+)-methamphetamine (in methanol), benzoylecgonine (in methanol), (-)-Ag-THC (in methanol), morphine (in methanol), codeine (in methanol), ( (in methanol), ( (in methanol), and ( MDMA] commonly found in drug addicts was spiked in a neat sample of the benzodiazepines (1.0 ng/ml) and analyzed by this method. To evaluate for matrix effect, a comparison of the instrumental response for a calibrator at 5 ng/ml was performed with two sets of samples (n = 3). In the first set of samples, the drugs were spiked directly into the mobile phase (50:50 methanol/water) providing a relative 100% value. In the second set of samples the drugs were spiked into pre-extracted blank oral fluid diluent samples. Comparison of sample set one and sample set two permits the calculation of the matrix effect (ME %) from the ratios of corresponding responses from the two sample sets. To evaluate for matrix variability, we collected oral fluid from six volunteers (non-drug users) in our laboratory using the OraSure Intercept device. The collected oral fluid from each volunteer was divided into three equal portions of 500 pl each and spiked with the 14 drugs at concentrations of 0.1, 5.0, and 10.0 ng/ml. These were extracted and the resulting curves overlaid. Results and Discussion Selectivity of the method was achieved by a combination of retention time, precursor and transition ions. The use of a quantifier, qualifier and qualifier/quantifier area ratios ensured identification of the drug. In this method a relative quali- Table II. Summary of Validation Parameters Within-Batch Within-Batch Compound Conc. Extraction Precision Accuracy LOQ LOD Name Linearity (ng/ml) Rec. (%) (CV, n = 6) (n = 6) (ng/ml) (ng/ml) Flunitrazepam 0.1-20 ng/ml, r = 1.0000 0.1 95 8.4 9.2 0.1 0.02 10 95 3.2 9.7 Oxazepam 0.1-20 ng/ml, r = 0.9998 0.1 106 19.1 6.3 0.1 0.05 10 92 3.2 5.5 Diazepam 0.1-20 ng/ml, r = 1.0000 0.1 95 10.3 8 0.1 0.05 10 94 5.9 6 Lorazepam 0.1-20 ng/ml, r= 1.0000 0.1 73 14.3-28.2 0.1 0.05 10 93 3.1-1.6 7-Aminoclonazepam 1.0-20 ng/ml, r = 1.0000 1 54 17 5.4 1 0.5 10 55 16 17 Desalkylflurazepam 0.1-20 ng/ml, r= 0.9998 0.1 103 9.9 12 0.1 0.05 10 93 3.9 12.6 Clonazepam 0.1-20 ng/ml, r = 0.9999 0.1 99 9.2 17.3 0.1 0.05 10 94 4.3 11 ct-hydroxyalprazolam 0.5-20 ng/ml, r = 0.9991 0.5 107 15 0.6 0.5 0.2 10 95 5.6 22.5 Alprazolam 0.1-20 ng/ml, r = 1.0000 0.1 93 14.3 4.9 0.1 0.02 10 83 7.7 10.2 Temazepam 0.1-20 ng/rnl, r = 0.9999 0.1 89 13.7 9.1 0.1 0.05 10 92 3.4 14.1 Chlordiazepoxide 1.0-20 ng/ml, r = 0.9998 1 71 8.4 3.9 1 0.5 10 93 3.2 26.3 7-Aminoflunitrazepam 0.5-20 ng/ml, r = 0.9987 0.5 72 11-25.2 0.5 0.1 10 83 17-13.5 Hydroxyethylflurazepam 0.1-20 ng/ml, r= 0.9995 0.1 96 17.3 4.5 0.1 0.05 10 93 4.5 7.5 Nordiazepam 0.1-20 ng/ml, r= 0.9999 0.1 91 15.1 1.3 0.1 0.05 10 90 3.4 1.5 Nordiazepam-d5 N/A 10 95 5.7 N/A N/A N/A 10 96 5.7 N/A 372

fier/quantifier ion ratio of 20% was used. Oral fluid analyzed from six non-drug consuming subjects from our laboratory showed no peaks indicating that this method is selective and specific for the analyzed benzodiazepines. This was also true for the panel of 10 drugs analyzed. No peaks were observed. The method was validated for linearity, LOD, LOQ, precision, accuracy, and analytical recovery by the analyses of spiked oral fluid samples. The obtained LOQ was 0.1 ng/ml for most analytes except for hydroxyalprazolam and 7-aminoflunitrazepam with an LOQ of 0.5 ng/ml and 7-aminoclonazepam and chlordiazepoxide with an LOQ of 1.0 ng/ml. These LOQs with a few exceptions are comparable with previous reports dealing with 14 benzodiazepines in oral fluid (16). Table II shows a summary of the validation parameters for the drugs. 7-Aminoclonazepam, chlordiazepoxide, and 7-aminoflunitrazepam showed low recoveries, 54%, 71%, and 72%, respectively, at 0.1 ng/ml, using the established extraction method. To improve on the recoveries of these drugs three different extraction methods (two liquid-liquid and one solid phase) were performed as described in the literature. The liquid-liquid extraction method by Kintz et al. (16) using 3 ml of diethylether/methylene chloride (50:50) provided a better recovery (75%) for 7-aminoclonazepam. However, the recoveries for the other drugs were either substantially reduced or enhanced. For example, the recovery for diazepam dropped to 52%, and alprazolam, hydmxyalprazolam, and hydroxyethyl-alprazolam had increased recoveries of 211%, 233%, and 172%, respectively. Recoveries greater than 100% generally indicate Table III. EIA and LC-MS-MS Results of Analyses of Authentic Samples from 41 Subjects Patient EIA Oxazepam Diazepam 7.Aminoclonazepam Clonazepam Hydroxyalprazolam Alprazolam Chlordiazepoxide Nordiazepam ID Ratio (ng/ml) (ng/ml) (ng/ml) (n~/ml) (n~/ml) (ng/ml) (ng/ml) (ng/ml) I 0.002 2 0.002 3 0.003 4 0.028 5 0.039 6 0.051 7 0.056 8 0.077 0.39 1.25 9 0.091 0.44 10 0.109 11 0.128 12 0.142 13 0.146 14 0.155 15 0.183 16 0.251 17 0.299 0.11 0.14 18 0.301 0.35 19 0.309 0.23 20 0,315 0.75 21 0.339 0.14 22 0,373 23 0.376 24 0.384 0.90 25 0.419 26 0.487 27 0.539 28 0.560 0.18 0.28 29 0.591 30 0,606 31 0.613 32 0.625 33 0.626 0.34 34 0.681 0.50 35 0.714 36 0.825 0.12 37 0.830 38 0.913 0.18 0.47 39 0.994 0.39 40 1,002 0.35 41 1.460 121.00 0.32 0.64 19.50 1.14 39.13 110.25 1.44 21.38 20.70 7.04 1.01 20.70 2.45 0.60 1.10 13.10 0.13 16.90 4.93 1.18 9.15 1.83 1.85 1.98 6.79 0.31 1.32 0.12 0.11 7.88 5.55 4.60 5.77 1.50 4.12 0.86 2.42 2.43 2.04 29.30 2.24 2.94 1.97 1.66 0.97 0.62 1.54 1.70 1.21 3.12 0.89 1.06 1.11 7.21 1.07 6.85 3.76 2.00 0.35 12.76 0.14 0.99 4.22 0.40 1.20 373

that ion enhancement was occurring for these drugs by this method. Given the fact that this liquid-liquid method was laborious, could not be automated, and did not provide any significant advantage(s) over the solid-phase extraction method, its use was discontinued and the reason(s) for the ion enhancement was not investigated any further. The same was true for the other liquid-liquid (using methylene chloride/isopropanol) and solid-phase extraction [using the strata columns from Phenomenex along with the recommended extraction method (23)] methods. These gave better recoveries for a few drugs but poor recoveries for the others. Overall the solid-phase extraction procedure involving the use of the Varian Certify Bond Elut column, which is an adaptation with slight modifications of that used by Moore et al. (24), gave excellent recoveries for a broad range of benzodiazepines and hence lower limits of detection, thus rendering this method suitable for the analysis of benzodiazepines in oral fluid. Upon extraction and running of a three-point curve (0.1, 5.0, and 10.0 ng/ml) from the six volunteers (non-drug users), it was observed that the curves overlaid directly on each other indicating that there was little or no matrix effect from personto-person. There was little or no interference with the analytes by any extractable endogenous materials present in oral fluid as indicated by the matrix effect (ME%), which was + 11% or lower for all drugs except for 7-aminoclonazepam and 7-aminoflunitrazepam, which were 66% and 77%, respectively. But their consistent recoveries, good chromatograms, and linearity rendered the quantitation of these analytes possible. The applied gradient in the LC ensured the elution of all the drugs examined within 6 rain and produced chromatographic peaks of accepted symmetry as illustrated in Figure 1. Only chlordiazepoxide showed poor chromatography with this elution. However, the chromatography of chlordiazepoxide improves significantly as the concentration increases. This accounts for the reason why the LOQ for chlordiazepoxide is much higher (1.0 ng/ml) than those of the other drugs. The short analysis time along with the automation (using the Zymark RapidTrace instrument) of the extraction procedure leads to a high throughput for the analysis. The use of the collection device did not substantially affect the results obtained. A comparison of area counts for the same concentration of drugs spiked in oral fluid diluent (synthetic saliva), oral saliva (collected by spitting), and pad saliva (obtained by spitting before spiking on the collection pad) was done and was found to be less than 15% for all drugs. The recovery of the drugs from the collection pad was greater than 85% for all drugs. Thus, neither the use of the oral m :'J j:::: ]' lgalk for CIonazepam 'Z NorWazepaw,-D s inlel~d standard ]:! i 84 :U Figure 1. Extracted ion chromatogram showing the separation of the analyzed benzodiazepines. Peak identification: 1, 7-aminoclonazepam; 2, 7-aminoflunitrazepam; 3, chlordiazepoxide; 4, c~-hydroxyalprazolam; 5, alprazolam; 6, flunitrazepam; 7, clonazepam; 8, Iorazepam; 9, oxazepam; 10, nordiazepam; 11, temazepam; 12, nordiazepam-ds; 13, desalkylflurazepam; and 14, diazepam. 4!:! -, '-~--~----J Figure 3. Chromatogram for patient #6 positive for clonazepam at a concentration of 0.60 ng/ml. [ i s Z~ Peak for alpra~hmt ~Z Nordiazcpam-I)~ Internal standard :Z Z Peak for 7.Amlnoclonazclnml Nordiaz~m.I)~ Internal ~zal~lard ~E... Z', 7~ -~,i~, L-84 ~... i -I Figure 2. Chromatogram for patient #6 positive for alprazolam at a concentration of 13.10 ng/ml. Figure 4. Chromatogram for patient #6 positive for 7-aminoclonazepam at a concentration of 2.45 ng/ml. 374

fluid diluent nor the collection device substantially affected the test results. To demonstrate the applicability of this method, authentic oral fluid samples collected from patients in a drug facility center (and screened positive by EIA) were analyzed by this method. Each of the 41 subjects was positive for at least one of the benzodiazepines included in this study. Table III summarizes the results of the subjects. It should be noted that the concentrations shown in Table III are estimates, which are about one-third of the concentrations in the patients sample given that the oral fluid (about 0.4 ml) collected from the patients with the Intercept collection device is diluted with a preservative solution (0.8 ml) before analysis. Figures 2-4 represent the chromatograms obtained for one of the subjects positive for more than one drug. The following observations are made on the basis of these resuits. Of the 41 patient samples analyzed, the most common benzodiazepine detected was alprazolam. Thirty-four patients were positive for alprazolam with concentrations ranging from 0.86 to 121.00 ng/ml. Of these 34, 9 were positive for hydroxyalprazolam with concentrations ranging from 0.11 to 1.44 ng/ml. Eight of the patients positive for alprazolam were also positive for clonazepam or its metabolite 7-aminoclonazepam. Another 9 of the 34 alprazolam patients were positive for diazepam, oxazepam, and/or nordiazepam. Although no formal stability studies were done, it can be assumed that the benzodiazepines were all stable under the storage condition because all 41 subjects (previously screened positive by EIA) were still positive by the LC-MS-MS method one year after storage in frozen condition. Conclusions A simple, rapid, sensitive, and novel LC-MS method was developed and validated that is suitable for the confirmation of 14 benzodiazepines present in oral fluid. The use of a combination of retention time, presence of quantifier and qualifier ions as well as a quantifier/qualifier ion ratio allowed this method to be used as a confirmation method for the OTI Benzodiazepine Intercept MICRO-PLATE EIA. The method was applied for the confirmation of authentic samples from 41 patients from a drug facility center. The assay should be useful for confirmation and quantification of low doses of these compounds in authentic oral fluid samples collected in forensic toxicological cases. Acknowledgments The authors would like to acknowledge the help of OraSure Technologies, whose facilities and instruments were used for this work. The EIA kit from OraSure Technologies and their personnel were very helpful in the screening of the 41 patients whose results were later confirmed by our LC-MS-MS method. References I. S.J. Mule and G.A. Casella. Quantitation and confirmation of the diazolo- and triazolobenzodiazepines in human urine by gas chromatography/mass spectrometry. J. Anal. Toxicol. 13:179-I 84 (1989). 2. M. Augsburger, L. Rivier, and P. Mangin. Comparison of different immunoassays and GC-MS screening of benzodiazepines in urine. J. Pharm. Biomed. Anal. 18:681-687 (1998). 3. C. Drouet-Coassolo, C. Aubert, P. Caossolo, and J.P. Cano. Capillary gas chromatographic-mass spectrometric method for the identification and quantification of some benzodiazepines and their unconjugated metabolites in plasma. J. Chromatogr. 487: 295-299 (1989). 4. A. Sioufi and J.P. Dubois. Chromatography of benzodiazepines. ]. Chromatogr. 531:459-480 (1990). 5. M. Schulz and A. Schmoldt. Therapeutic and toxic blood concentrations of more than 500 drugs. Die Pharmazie 52:895-911 (1997). 6. H. Ashton. Guidelines for the rational use of benzodiazepines. When and what to use. Drugs 48:25-40 (1994). 7. H. Gjerde, E. Dahlin, and A.S. Christophersen. Simultaneous determination of common benzodiazepines in blood using capillary gas chromatography. J. Pharm. Biomed. Anal. 10:317-322 (1992). 8. R.C. Baselt, C.B. Stewart, and S.J. Franch. Toxicological determination of benzodiazepines in biological fluids and tissues by flame-ionization gas chromatography. J. Anal. Toxicol. 1: 10-13 (1977). 9. C. Kratzsch, O. Tenberken, F.T. Peters, A.A. Weber, T. Kraemer, and H.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(8): 856-872 (2004). 10. 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). 11. M. Laloup, R. Fernandez, M. del Mar, W. Michelle; De Boeck Gert, 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(7): 616-626 (2005). 12. P. Kintz. Drug testing in addicts: a comparison between urine, sweat, and hair. Ther. Drug Monit. 18:450-455 (1996). 13. D.A. Kidwell, J.C. Holland, and S. Athanaselis. Testing for drugs of abuse in saliva and sweat. J. Chromatogr. 713:111-135 (1998). 14. R.E. West and D.P. Ritz. GC/MS analysis of five common benzodiazepine metabolites in urine as tert-butytdimethylsilyl derivatives. J. Anal Toxicol. 17:50-58 (1993). 15. A.A. Elian. Detection of low levels of flunitrazepam and its metabolites in blood and bloodstains. Forensic Sci. Int. 101: 107-111 (1999). 16. P. Kintz, M. Villain, M. Concheiro, and V. Cirimete. Screening and confirmatory method for benzodiazepines and hypnotics in oral fluid by LC-MS/MS. Forensic 5ci. Int. 150:213-220 (2005). 17. O. Quinteta, A. Cruz, A. de Castro, M. Concheiro, and M. Lopez- Rivadulla. Liquid chromatography-electrospray ionisation mass spectrometry for the determination of nine selected benzodiazepines in human plasma and oral fluid. J. Chromatogr. B 825: 63-71 (2005). 18. C.E. Jones, F.H. Wians, Jr., L.A. Martinez, and G.J. Merritt. Benzodiazepines identified by capillary gas chromatography-mass spectrometry, with specific ion screening used to detect benzophenone derivatives. Clin. Chem. 35:1394-1398 (1989). 19. A.D. Fraser. Alprazolam abuse and methadone maintenance. 375

J. Am. Med. Assoc. 258:2061-2062 (1987). 20. A.D. Fraser, W. Bryan, and A.F. Isner. Urinary screening for alprazolam and its major metabolites by the Abbott ADx and TDx analyzers with confirmation by GC/MS. J. Anal. ToxicoL 15: 25-29 (1991). 21. P.H. Dickson, W. Markus, J. McKernan, and H.C. Nipper. Urinalysis of c~-hydroxyalprazolam, c~-hydroxytriazolam, and other benzodiazepine compounds by GC/EIMS. J. Anal. Toxicol. 16: 67-77 (1992). 22. R Kintz, V. Cirimele, and B. Ludes. Detection of cannabis in oral fluid (saliva) and forehead wipes (sweat) from impaired drivers. J. Anal. Toxicol. 24:25-29 (2000). 23. H. Shahana, K. Krishna, and G. Michael. Single-step elution and analysis of acidic and basic benzodiazepines from urine by solidphase extraction and GC-MS. LC-GC North America, June 2005. 24. C. Moore, G. Long, and M. Marr. Confirmation of benzodiazepines in urine as trimethylsilyl derivatives using gas chromatography-mass spectrometry. J. Chromatogr. B 655:132-137 (1994). Manuscript received January 12, 2007; revision received April 25, 2007. 376