Technical Note] Introduction

Transcription

1 Technical Note] Urine Benzodiazepine Screening using Roche Online KIMS Immunoassay with p-glucuronidase Hydrolysis and Confirmation by Gas Chromatography- Mass Spectrometry* Kevin L. Klette 1,*, Russell F. Wiegand 2, Carl K. Horn 2, Peter R. Stout 3, and Joseph Magluilo, Jr. 4 1Navy Drug Screening Laboratory, 320 B Street, Great Lakes, Illinois ; 2Navy Drug Screening Laboratory, P.O. Box 113, Bldg. H-2033, Naval Air Station, Jacksonville, Florida 32212; 3Aegis Sciences Corporation, 345 Hill Avenue, Nashville, Tennessee 37120; and 4Office of the Armed Forces Medical Examiner, Division of Forensic Toxicology, AFIP, Rockvifte, Maryland 'Abstract Performance of the Roche Online KIMS (kinetic interaction of microparticles in solution) benzodiazepine (BZD) immunoassay (IA) with and without p-glucuronidase treatment was evaluated on a Hitachi Modular automated IA analyzer calibrated using nordiazepam at 100 ng/ml. Reproducibility, linearity, accuracy, sensitivity, and interferences were evaluated. Precision of the assay (percent coefficient of variation (%CV)) with and without addition of the enzyme was less than 6% and 9%, respectively, with linearity (r 2 value of and ), respectively. Betweenrun precision of a 125 ng/ml nordiazepam control (n = 287) over 67 days, produced a %CV of 13.6% for the hydrolytic assay. Modification of the BZD assay to include automated hydrolysis of urinary BZD glucuronide conjugates was evaluated using three glucuronidated BZD standards prepared at concentrations ranging from 250 to 10,000 ng/ml. With hydrolysis, temazepam, oxazepam, and Iorazepam glucuronides, produced crossreactivities of 25%, 15%, and 20%, respectively. Without hydrolysis, the glucuronidated BZD standards produced less than 1% cross-reactivity in the assay. The ability of the assay to differentiate between positive and negative samples was evaluated by assaying 20 negative urine samples and serial dilutions of certified drug-free urine fortified with 28 different BZDs. All of the negative and positive urine samples produced the appropriate screening result. Cross-reactivities of 27 different BZDs, calculated as the normalized IA response divided by the BZD concentration that produced a response approximately equivalent to the response of a 100 ng/ml nordiazepam standard and multiplied by 100, ranged from 15% to 149%. Human urine samples (n -- 28) that were previously found to contain BZDs by gas chromatography-mass spectrometry (GC-MS) also produced a positive BZD IA result. The IA was challenged with 78 potentially interfering compounds, and none produced a positive BZD * Disclaimer: The opinions contained in the publication are not to be construed as official or as reflecting the views of the Department of Defense or the Department of the Navy. This work was supported in part by the American Registry of Pathology. t The author to whom correspondence should be addressed. E-maih klklelte@nhgl.med.navy.mil. response. As a part of the validation, a large number of human urine samples (29,500) were assayed using the modified Online BZD IA method to evaluate the performance of the method in production. Of the 29,500 samples tested, 80 produced a positive IA result. Analysis by GC-MS confirmed the presence of at least 1 BZD compound in 61 of the samples corresponding to a confirmation rate of 76%. The Online BZD IA modified by the automatic addition of I~-glucuronidase appears well adapted for the rapid detection of BZDs and their metabolites in human urine. Introduction Since the introduction of chlordiazepoxide (Librium in 1961, benzodiazepines (BZDs) have been used for general anesthesia and the treatment of sleep disorders, anxiety-induced depression, stress, muscle spasms, seizures, alcohol withdrawal, and other problems (1-3). BZDs are considered to be safer, more effective, and have largely replaced the use of barbiturates as sedative-hypnotics. Although BZDs are valuable therapeutic agents that are well tolerated and exhibit a low order of acute and chronic toxicity, chronic use can produce dependence and abuse (4,5). BZDs are used recreationally to augment the effects of heroin and cocaine, mitigate the effects of withdrawal symptoms of heroin, and increase the seizure threshold of cocaine users (6). BZD use has been shown, experimentally, to impair driving skills (7-12), and it is related to an increased risk of motor vehicle crashes (13,14) and driving under the influence of drug cases (15). BZDs are also misused, particularly in cases of date rape, robbery (16), and suicidal poisonings (17-21). Clinical and forensic toxicology laboratories require immunoassay (IA) applications that provide rapid qualitative detection of BZDs in urine with high sensitivity and specificity to include both abusive use of these drugs and low-totherapeutic doses of the newer, more potent BZDs. The ability to detect BZDs in human urine using commer- Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 193

2 cially available immunochemical assays is problematic because of several factors related to their wide variation in potency, structure, metabolism, and elimination. Many of the BZDs utilize common metabolic pathways resulting in several pharmacologically active metabolites excreted primarily as glucuronide conjugates with only low levels of the parent drug. In contrast, the newer more potent generation of BZDs are effective therapeutically at much lower doses and are detected unchanged in urine at low concentrations. Because of their wide range of therapeutic efficacy, the concentration range of BZDs and their metabolites varies from nanogram-per-milliliter to microgram-per-milliliter quantities in human urine. The lack of immunochemical assay sensitivity to low-dose (< 10 mg/day) BZDs (e.g., triazolam, alprazolam, flunitrazepam, and nitrazopam) or BZDs excreted primarily as glucuronide conjugates has been cited by several investigators (22-25). Analytical identification by immunochemical methods is further complicated because of the variable immunoreactivity of the antibodies utilized in the assay system to the diverse structural differences found in the BZD class of drugs. Several published studies have reported increased BZD sensitivities by utilizing manual enzymatic pretreatment steps prior to analysis to cleave glucuronide conjugates with practical hydrolysis times ranging from 30 min to 2 h (16,22,24,26,27). The excessive time requirement of manual hydrolysis hinders the use of automated IA systems for the rapid detection of BZDs required for emergency, clinical, and forensic investigations. The purpose of this study was to validate the performance of the Roche Online KIMS (kinetic interaction of microparticles in solution) BZD IA modified to include the automated addition of 13-glucuronidase for high-speed analysis of BZDs and their metabolites in human urine. The assay was validated regarding reproducibility, linearity, accuracy, sensitivity, and interferences from common over-the-counter products, prescription drugs and some of their metabolites, and other drugs of abuse. The cross-reactivity of 27 BZDs and their metabolites and the ability of the assay to discriminate samples previously found to contain BZDs by gas chromatography-mass spectrometry (GC-MS) from known negative samples was also investigated. A large sampling of randomly collected human urine samples was initially screened using the modified Roche Online BZD IA followed by analysis of the presumptive positive samples by GC-MS to further document assay sensitivity, specificity, and performance and to characterize the distribution of BZDs in authentic human urine samples. Materials and Methods Table I. Quantitation and Identity Ions (m/z) Monitored of BZDs for the GC-MS Method Immunoassay BZD IA screening was accomplished on the P module of a Roche/Hitachi modular automated IA analyzer (Indianapolis, IN). The Online KIMS BZD assay and a qualitative calibrator Compound Internal Standard Quantitating Ion Qualifier #1 Qualifier #2 Nordiazepam Nordiazepam-d Desalkylflurazepam Nordiazepam-ds Nordiazepam-ds* Oxazepam Diazepam-ds Lorazepam Diazepam-d Diazepam Diazepam-d s ] Diazepam-d5 * Temazepam Temazepam-d s Midazolam Temazepam-d Temazepam-ds* (z-oh Midazolam (Z-OH Alprazolam-d s (Z-OH Alprazolam (Z-OH Alprazolam-ds Alprazolam (Z-OH Alprazolam-d (Z-OH Triazolam (Z-OH Alprazolam-d~ (z-oh Alprazolam-d~* OH-Hydroxyethylflurazepam 7-Aminoclonazepam-d Aminoflunitrazepam 7-Aminoclonazepam-d Aminoclonazepam 7-Aminoclonazepam-d Flurazepam 7-Aminoclonazepam d Prazepam 7-Aminoclonazepam d Aminoclonazepam-d4* Aminonitrazepam Oxazepam-d Oxazepam-ds* * Internal standard. 194

3 (C.f.a.s. DAT CAL 3) containing nordiazepam at a concentration of 100 ng/ml was purchased from Roche (Indianapolis, IN). Escherichia coli (E. coli) [3-glucuronidase with an enzyme activity rated at a minimum of 200 units/rag at 37~ was obtained from Roche Molecular Biochemicals (Mannheim, Germany) and added to the reagent I (R1) in a ratio of 28 lal enzyme solution per milliliter. The resulting enzyme activity of R1 was 5.6 units/ml. Except for the addition of [3-glucuronidase to the R1 reagent, all IA reagents, standards, and controls were utilized in accordance with the manufacturer's instructions for the analysis of urine samples. Immunoassay standards and controls A nordiazepam solution purchased from Cerilliant (Austin, TX) was added to certified drug-free urine from Roche (Indianapolis, IN) to prepare the 125 ng/ml and 75 ng/ml interassay precision controls. The Hitachi P module was calibrated daily prior to running samples using a single-point calibration per manufacturer specifications and the resulting absorbance normalized to equal 100 arbitrary units. In order to assess the precision, accuracy, and linearity of the assay, 10 replicates each of the nordiazepam standards prepared at 0 ng/ml, 50 ng/ml, 75 ng/ml, 100 ng/ml, 125 ng/ml, 150 ng/ml, and 200 ng/ml were analyzed. The resulting mean and standard deviation (SD) were calculated using absorbances for each concentration tested. A one-way, crossed, fixed-effect model analysis-of-variance (ANOVA) test was used to determine significant differences between concentration levels using Microsoft Excel 2000 (Redmond, WA). Linear regres- sion analysis was also performed using Microsoft Excel Accuracy was assessed by analyzing 28 human urine samples that were donated by the Armed Forces Institute of Pathology (AFIP, Rockville, MD) and Aegis Sciences Corporation (Nashville, TN) and previously determined to contain one or multiple BZDs by GC-MS. Seventy-eight over-the-counter (OTC) drug/drug metabolite and pharmaceutical drugs prepared as multi-constituent standards consisting of the following drug categories: barbiturates, cocaine and cocaine metabolites, tricyclic antidepressants, stimulants, hallucinogens, and opiates, and 25 single-constituent BZD standards were utilized to assess specificity and/or cross-reactivity of the assay. These standards were donated by the AFIP or purchased from Alltech Associates, Inc. (Deerfield, IL) or Cerilliant (Austin, TX). In addition, three glucuronidated standards of temazepam, oxazepam, and lorazepam were purchased from Alltech Associates to determine their cross-reactivity to the immunoassay and to evaluate the efficiency of hydrolysis. GC-MS standards and controls Methanolic standards of nordiazepam, desalkylflurazepam, oxazepam, lorazepam, diazepam, temazepam, (z-hydroxymidazolam, c~-hydroxyalprazolam, alprazolam, (~-hydroxytriazolam, midazolam, 2-OH hydroxyethylflurazepam, 7-aminoflunitrazepam, 7-aminonitrazepam, 7-aminoclonazepam, flurazepam, and prazepam purchased from Cerilliant in either 1.0 mg/ml or 100 lag/ml vials were spiked into certified drugfree urine from Roche (Indianapolis, IN) at a final concentration of 100 ng/ml. Deuterated stock solutions of nordiazepare-d5, oxazepam-d5, diazepam-d5, temazepam-d5, (x-oh alprazolam-d5, and 7-aminoclonazepam-d4 purchased from Cerilliant (Austin, TX) and spiked into certified drug-free urine at a final concentration of 166 ng/ml were used as the internal standards. Human urine sample analysis Human urine samples (29,500) collected randomly under forensic conditions were IA screened using the modified Online BZD assay at a cutoff concentration of 100 ng/ml (nordiazepam). All analyses included 4% open quality controls consisting of nordiazepam prepared at 75 and 125 ng/ml and a hydrolysis control (temazepam glucuronide) prepared at 500 ng/ml. Hydrolysis and extraction E. coli [Lglucuronidase (2,000,000 units) purchased from Sigma Chemical (St. Louis, M0) was reconstituted in 10 ml ph 6.0 phosphate buffer to 200,000 units/ml. Sample aliquots Table II. Mean, Standard Deviation, and %CV of Nordiazepam Absorbance Values using the Online IA Tested With and Without p-glucuronidase Without enzyme O O Mean , t079.8 Standard deviation O %CV , With enzyme Mean Standard deviation %CV i o 120{)0 L ~ No Enzyme '~ o Enzyme ~ ~~ ~ ~ = ooo 4 R e ~ Nordiazepam concentration (ng/ml) Figure 1. Linear regression for Roche Online KIMS BZD reagent with and without ~-glucuronidase

4 were enzymatically hydrolyzed by combining 100 IJL phosphate buffer/[3-glucuronidase solution with 5 mls of urine for a final concentration of 4000 units/ml. The samples were incubated at 50~ for I h, cooled to ambient temperature, and centrifuged at 3500 rpm for 5 rain. The hydrolyzed samples were poured into ZSDAU020 preconditioned solid-phase extraction columns from United Chemical Technologies (Bristol, PA) and allowed to flow at a rate not to exceed 1 ml/min. Columns were washed with 2 ml of deionized water followed by 2 ml of 20% acetonitrile in 0.1M phosphate buffer. Columns were then dried and rinsed with 2 ml of hexane. Analytes were eluted with 3 ml of ethyl acetate/ammonium hydroxide (98:2, v/v) and evaporated to dryness at a temperature of 50~ Samples were derivatized by the addition of 50 IJL ofn, O-bis (Trimethylsilyl) trifluoroacetamide containing 1% trimethylchlorosilane purchased from Pierce Chromatography (Rockford, IL) and 50 IlL of ethyl acetate. The tubes were vortex mixed, capped, and incubated at 70~ for 30 min. Samples were cooled to ambient temperature and transferred to GC-MS glass conical vials for analysis. All solvents were HPLC grade. GC-MS analysis was performed using Agilent (Palo Alto, CA) 6890/5973 GC-MS detectors equipped with a G2614A autosampler and Agilient ChemStation software (version G1701CA). An RTX-200 capillary column (15 m x 0.25 mm x 0.251Jm) from Restek Corporation (Bellefonte, PA) provided analytical separation. The carrier gas was helium with a purity level of , at a constant flow of 1 ml/min. Samples (2 IJL) were injected in a pulsed split mode with a split ratio of 10:1. Pulse pressure was 30.0 psi for 0.75 rain. Injector temperature was 250~ with a transfer line temperature of 280~ GC temperature program conditions was 125~ for 1.0 min, increased at 15~ to 235~ increased 8~ to 275~ and finally increased 25~ to 300~ for a hold time of 3 rain. Total GC run time was rain. The mass selective detector was tuned in the standard autotune mode, and data collection was accomplished at 200 electron volts above the tune value. All GC-MS analyses were performed in the electron-impact selected ion monitoring (SIM) mode using the quantitation and qualifier (identity) ions indicated in Table I. Quantitation of BZDs in controls and samples was accomplished by single-point calibration against a multi-constituent calibrator containing 100 ng/ml of the BZDs listed in Table I. Identification of the target analyte was considered acceptable if the specimens and controls exhibited retention times within +_ 2% and identity ion abundance ratios within + 20% of the calibration standard. Additionally, all quality control samples within each analytical set were re- Table III. Effect of Using ~-Glucuronidase Enzyme as an Additive to the Immunoassay* Conc.(ng/mt) With ~-Glucuronidase Without ~-Glucuronidase Lorazepam Oxazepam Temazepam torazepam Oxazepam Temazepam 10, * Normalized IA responses of three glucuronidated BZD standards prepared at 250, 500, 1000, and 10,000 ng/ml in certified drug-free urine. The screening cutoff calibrator absorbance was normalized to equal 100 arbitrary units (i.e., a number greater than or equal to I00 is positive), Table IV. Immunoassay Response and GC-MS Concentration of BZDs in Human Urine Samples Previously Determined to Contain BZDs Normalized IA Response* Drug Result (conc. ng/ml) Nordiazepam (150), Oxazepam (5660), Temazepam (330) Alprazolam (270) c~-oh-alprazolam (150), Alprazolam (260) Nordiazepam (350), Alprazolam (1420), Temazepam (860) Nordiazepam (130), Diazepam (50), Temazepam (1050! Nordiazepam (220), oc-oh-alprazolam (230/, Temazepam (130), Alprazolam (50) Nordiazepam (2940) Temazepam (630) Temazepam (2860) ~-OH-Alprazolam (t 60), Alprazolam (250) ~-OH-Alprazolam (70), Alprazolam (180) o~-oh-alprazolam (460), Alprazolam (2480) o~-oh-alprazolam (540), Alprazolam (2480) c~-oh-alprazolam (160), Alprazolam (500) c~-oh-alprazolam (590) Temazepam (1830) Alprazolam (180) Nordiazepam (70) Nordiazepam (80), Diazepam (160) Lorazepam (1000) c~-oh-alprazolam (264) Nordiazepam (2000) Oxazepam (1100), Temazepam (521) Oxazepam (421) Nordiazepam (399), Temazepam (2000) Oxazepam (981) Nordiazepam (442) * The screening cutoff calibrator absorbance was normalized to equal 100 arbitrary units (i.e., a number greater than or equal to 100 is positive). 196

5 quired to quantitate within + 20% of the theoretical concentration. Mass spectrum analysis was linear to 1500 ng/ml with an r 2 = The limit of detection was established at 10 ng/ml with a limit of quantitation of 25 ng/ml. The assay's upper limit of linearity was 1.5 lag/ml. Results and Discussion ANOVA analysis confirmed that the seven concentrations of nordiazepam standards prepared at 0 ng/ml, 50 ng/ml, 75 ng/ml, 100 ng/ml, 125 ng/ml, 150 ng/ml, and 200 ng/ml were significantly resolved at a 95% confidence level with F-ratios of 4040 with the enzyme (13-glucuronidase) and 2040 without the enzyme. Using the absorbance data from these 10 replicates, the percent coefficient of variation (%CV) calculated was less than 10% for each concentration (Table II). The mean of the 10 absorbances at the 100% cutoff level varied from the calibration absorbance by only 1.3%. The assay was linear with r 2 values of with enzyme and without enzyme (Figure 1). The addition of [3-glucuronidase did not significantly affect the performance of the assay's ability to detect nordiazepam. There was no significant difference (cz = 0.05) between the linearity data sets with and without the enzyme by a two-way ANOVA (performed using Microsoft Excel) of the absorbance data across the calibration range. The effect of enzymatic hydrolysis on the response of glucuronide conjugated BZDs in urine was determined by analyzing single constituent standards of temazepam glucuronide, oxazepam glucuronide, and lorazepam glucuronide prepared at 250 ng/ml, 500 ng/ml, 1000 ng/ml, and 10,000 ng/ml. The purity of these controls was assessed using LC-MS analysis in the full scan and SIM modes. There was no trace of free temazepam or lorazepam in the glucuronidated controls. The glucuronidated oxazepam control contained less than 0.5% oxazepam. These glucuronidated BZD controls were analyzed both with and without the addition of 13-glucuronidase (Table III). Without the enzyme, only oxazepam glucuronide displayed a positive result at 10,000 ng/ml. With the enzyme present, however, positive results were produced at concentrations of 400 ng/ml for temazepam glucuronide, 500 ng/ml for lorazepam glucuronide, and 650 ng/ml for oxazepam glucuronide. The ability of the Online BZD IA to differentiate between positive and negative samples was evaluated by assaying 20 human urine samples that previously screened negative in a drug test panel for marijuana, cocaine, amphetamines/amphetamine designer drugs, methamphetamine, barbiturates, phencyclidine, and opiates, and 28 human urine samples that had previously been found to contain one or more BZDs and their metabolites by GC-MS. All of the negative and positive human urine samples produced the appropriate screening result relative to the 100 ng/ml BZD cutoff. The identity and GC-MS concentrations of the positive BZDs are listed in Table IV. The cross-reactivity of 27 BZDs or their metabolites was determined by analyzing single standard solutions at various concentrations. The concentration of BZD that produced a response approximately equivalent to the normalized response (100 arbitrary units) of the nordiazepam calibrator is shown in Table V. The percentage of cross-reactivity was calculated as the normalized IA response divided by the concentration and multiplied by 100. Although most of the BZDs exhibited crossreactivities of less than the nordiazepam calibrator (52-94%), seven of the BZDs (norfludiazepam, (z-oh-midazolam, triazolam, alprazolam, diazepam, ~-OH-triazolam, and ~-OH-alprazolam produced cross-reactivities that equaled or exceeded (94-149%) the reactivity of the nordiazepam calibrator. In contrast, halazepam, chlordiazepoxide, flunitrazepam, and the glucuronidated BZDs (temazepam, oxazepam, and lorazepam) produced lower cross-reactivities that ranged from 15 to 28%. Further improvement in assay sensitivity to glucuronide conjugated BZDs may be realized by increasing the time of reaction. Interference of the assay was evaluated using standards containing 78 common OTC products, prescription drugs and some of their metabolites, and other drugs of abuse. These compounds were tested in both single- and multi-constituent Table V. Cross-Reactivities of BZDs and their Metabolites* BZD Normalized Percentage Concentration Response Cross-Reactivity t Temazepam Temazepam glucuronide Oxazepam Oxazepam glucuronide Clonazepam Lorazepam Lorazepam glucuronide Noffludiazepam oc-oh-midazolam Flunitrazepam Aminoflunitrazepam Triazolam Alprazolam Nordiazepam I Diazepam Hydroxyethylflurazepam c-OH-Triazolam 105 "t ~-OH-Alprazolam I Bromazepam Clorazepate Nitrazepam Chlordiazepoxide Prazepam Chlobazam Halazepam Midazolam Aminoclonazepam Flurazepam * Concentration of BZDs that produced an IA response approximately equivalent to the calibrator (I00 ng/ml nordiazepam). "~ Percentage of cross-reactivity calculated as the normalized IA response divided by the concentration multiplied by

6 mixtures, and the concentrations were chosen as easily achieved physiological concentrations. Of the wide range of compounds studied, none produced a positive response relative to the nordiazepam calibrator (Table VI). A total of 29,500 human urine samples were assayed using the modified Online BZD IA method to evaluate the performance of the method in production. A temazepam glucuronide control prepared at 400 ng/ml was included in each ana]ytica] batch to control for the hydrolytic performance of the enzyme during the course of running these samples. Between-run precisions of the 125 ng/ml nordiazepam control (n = 287) and the temazepam glucuronide control (n = 282) over 67 days during the testing of actual human urine samples, produced %CVs of 13.6% and 14.8%, respectively. The assay produced a Table Vl. Interference Compounds Tested in Drug-Free Urine* positive result in 80 of the samples. Further analysis on two of the presumptively positive samples was not possible as one was an evidentiary sample and the other sample contained an insufficient volume for further analysis by GC-MS. Table VII lists the normalized IA data, drug result, and concentration of the samples analyzed by GC-MS using the panel of BZDs indicated in Table I. At least one BZD or metabolite was identified in 61 of the 80 samples corresponding to a confirmation rate of 76%. It should be noted, however, that the 19 samples that failed to confirm positive for a BZD by GC-MS might have contained other cross-reacting BZDs or their metabolites not examined by the GC-MS procedure. The IA results for the non-confirmed samples ranged from 100 to 556 ng/ml. It is also possible that high potency benzodiazepines Drug pg/ml Drug pg/ml I Acetaminophen I ,11-Dihydrocarbamazepine I Methylprimidone I00 4 5,5-Diphenylhydantoin ~-Methyl-~-propylsuccinimide Amitriptyline 10 7 Amobarbital 5 8 Amoxapine 2 9 Azacyclonol 2 10 Barbital Brompheniramine maleate Butabarbital 5 ] 3 Caffeine Carbamazepine Chloroquine Chlorpheniramine Chlorpromazine Cimetidine Citalapram I0 20 Cocaine I0 2] Codeine Cotinine d,/-methamphetamine d-amphetamine d-ephedrine HCI Desipramine Dextromethorphan hydrobromide Dextropropoxyphene HCI ] Diacetylmorphine ] 0 30 Diphenhydramine Doxepin 5 32 Doxylamine succinate d-pseudoephedrine HCI EDDP Ethosuximide Ethotoin Fluoxetine Glutethimide Hydromorphone Hydroxyzine 2 41 Imipramine HCI Lidocaine HCI monohydrate Mefloquine Meperidine Mephenytoin Mephobarbital Methadone I0 48 Methaqualone HCI I00 49 Metharbital I00 50 Methsuximide I00 51 Methyl PEMA Methylphenidate HCI Morphine Nalorphine Naproxen Nicotine n-normethsuximide Nortriptyline Orphenadrine Oxycodone PEMA Pentazocine 5 63 Pentobarbital 5 64 Phencyclidine 5 65 Phenethylamine Phenobarbital Phensuximide Phenylpropanolamine HCI PMA Primidone Proadifen 2 72 Promethazine Quinidine 2 74 Ranitidine HCI Salicylic acid Secobarbital 5 77 Sertraline Verapamil 10 * Note: Compounds found in commercially available or in-house prepared multi- or single-constituent standards. 198

7 Table VII. Immunoassay Response and GC-MS Concentration of BZDs Detected from 29,500 Authentic Human Urine Samples Normalized Drug Result Normalized Drug Result IA Response (Conc. ng/ml ) IA Response (Conc, ng/ml) Temaza )am (50) Temaza )am (113) Temaza )am (114) Temaza )am (150) Temaza )am (156) Temaza )am (166) Temaza )am (192) Temaza )am (284) Temaza )am (304) Temaza )am (720) Temaza )am (1061) Nordiazepam (151) Oxazepam (71) Nordiazepam (76), Oxazepam (185) Nordiazepam (82) Oxazepam (89) Nordiazepam 75, Temazapam (110) Nordiazepam (76), Temazapam (189) Nordiazepam (200), Temazapam (205) Nordiazepam present, Temazepam (1890) Nordiazepam present, Temazepam (4413) Nordiazepam present, Temazepam (23556) Nordiazepam present, Temazepam (443) Nordiazepam present, Temazepam (38) Traces Diazepam, Nordiazepam, Temazepam Nordiazepam (102), Oxazepam (220), Temazepam (205) Nordiazepam (51), Oxazepam (126), Temazepam (226) Nordiazepam (232), Oxazepam (422), Temazepam (780) Nordiazepam (200) Nordiazepam (150) Nordiazepam (268), Lorazepam (147) Nordiazepam (43.7), Oxazepam (109), Lorazepam (109) Nordiazepam (74), Oxazepam (135), Lorazepam (135) may not have great enough cross-reactivity and concentrations to trigger a positive screening assay and thus were not subjected to confirmatory analysis. In all, 109 BZDs [parent compound(s) an~or metabolites] were identified in the 61 positive samples. The 109 identified compounds were representative of 11 different BZDs (Figure 2). The most frequently identified BZDs were temazepam (24%), nordiazepam (23%), oxazepam (17%), ~-OH-alprazolam (12%), and alprazolam (10%). Of the 61 GC-MS positive samples 11 contained temazepam alone and 3 samples contained both temazepam and oxazepam, indicating the use of temazepam. Fifteen of the samples contained a combination of nordiazepam and temazepam or oxazepam (or both), indicating the use of diazepam. Nordiazepam S Nordiazepam (57)Oxazepam(109) Lorazepam (109) Nordiazepam (57)Oxazepam(109) Lorazepam (109) Lorazepam (123) Alprazolam (200), e~-oh-alprazolam (40) Alprazolam (21), o~-oh-alprazolam (85) Alprazolam (50), ec-oh-alprazolam (310) Alprazolam (65), ec-oh-alprazolam (1434) e~-oh-alprazolam (268), Alprazolam (122) ec-oh-alprazolam (78), Alprazolam (19) o~-oh-alprazolam (305), Alprazolam (247) ec-oh-alprazolam (113), Alprazolam (46) o~-oh-alprazolam (40) o~-oh-alprazolam (157), Alprazolam (55) ~-OH-Alprazolam (122), Alprazolam (14) ec-oh-aiprazolam (40), Alprazolam Nordiazepam (50) Oxazepam (109), Temazepam (63) Oxazepam (389), Temazepam (263) Oxazepam (21), Temazepam(151) Oxazepam (143) Flurazepam (143) Nordiazepam (60), Oxazepam (114), Flurazepam (114) Nordiazepam (110), Oxazepam (75), Flurazepam (75) Oxazepam (48) Oxazepam (162) 7-Aminoclonazepam (142) 7-Aminodonazepam (267) oc-oh-midazolam (215) r (1083) 7-Aminonitrazepam(503) 7-Aminonitrazepam (498) ~ J / / / / / /,,.. O".~ / Figure 2. Bar graph representing the distribution of the 109 BZDs identified by GC-MS from 29,500 human urine samples. 199

8 alone was identified in two of the samples, and five samples contained both nordiazepam (with or without additional diazepam metabolites) and lorazepam. In total, the use of diazepam was indicated in 22 of the 61 samples (36%). Either alprazolam or cr and alprazolam were present in 12 samples. Conclusions This report documents the performance of the Roche Online KIMS BZD IA modified to include the automated addition of [3-glucuronidase by the IA analyzer. The assay demonstrated increased sensitivities to glucuronidated BZDs while maintaining sensitive detection of other BZDs and their metabolites in human urine in a production environment. This is supported by the pattern of BZDs detected in the large sampling of randomly collected human urine samples. Addition of [3-glucuronidase to the R1 reagent and automated pipetting by the analyzer provides for the rapid detection of the majority of BZDs in urine. The ability to quickly detect the presence of BZDs in human urine is an integral component of drug screening programs in forensic and clinical toxicology. Acknowledgments The authors acknowledge with thanks Karoline K. Shannon of the Office of the Armed Forces Medical Examiner, Division of Forensic Toxicology, Rockville, Maryland for her assistance in the efficient extraction, processing and analysis of samples, and John Dubay, Warren Taylor, Mike Brown, Mary K. Carter, and Mr. Erasmo Villaluz of the Navy Drug Screening Laboratory, Jacksonville, FL in the processing of IA samples for this study. References 1. K. Farrell. Benzodiazepines in the treatment of children with epilepsy. Epilepsia 27:45-51 (1986). 2. T.R. Browne. Therapy of status epilepticus. Compr. Ther. 8:28-36 (1982). 3. Goodman and Gilman's The Phan~acological Basis of Therapeutics, 10th ed., J.G. Hardman, L.E. Limbird, and A.G. Gilman, Eds. McGraw-Hill, New York, NY, Research monograph series 73. National Institute on Drug Abuse. Benzodiazepines, 1986, p D.J. Garretty, K. Wolff, A.W.M. Hay, and D. Raistrick. Benzodiazepine misuse by drug addicts. Ann. Clin. Biochem. 34:68-73 (1997). 6. L.P. Longo. Non-benzodiazepine pharmacotherapy of anxiety and panic in substance abusing patients. Psychiatr. Ann. 28: (1998). 7. M. Linnoila. Drug interaction on psychomotor skills related to driving: hypnotics and alcohol. Ann. Med. Exp. Biol. Fenn. 51: (1973). 8. M. Linnoila. Effects of diazepam, chlordiazepoxide, thioridazine, 200 haloperidole, flupenthixole and alcohol on psychomotor skills related to driving. Ann. Med. Exp. Biol. Fenn. 51: (1973). 9. M. Linnoila and S. Hakkinen. Effects of diazepam and codeine, alone and in combination with alcohol, on simulated driving. Clin. Pharmacol. Ther. 15: (1974). 10. M. Linnoila and M.J. Mattila. Drug interaction on psychomotor skills related to driving: diazepam and alcohol. Eur. J. Clin. Pharmacol. 5: (1973). 11. M. Linnoila, I. Saario, and M. Maki. Effect of treatment with diazepam or lithium and alcohol on psychomotor skills related to driving. Eur. J. Clin. Pharmacol. 7: (1974). 12. M. Linnoila, I. Saario, J. Olkoniemi, R. Liljequist, J.J. Mimberg, and M. Maki. Effect of two weeks' treatment with chlordiazepoxide or flupenthixole, alone or in combination with alcohol, on psychomotor skills related to driving. Arzneimittelforschung 25: (1975). 13. F. Barbone, A.D. McMahon, P.G. Davey, A.D. Morris, I.C. Reid, D.G. McDevitt, and T.M. MacDonald. Association of road-traffic accidents with benzodiazepine use. Lancet 352" (1998). 14. B. Hemmelgarn, S. Suissa, A. Huang, J.F. Boivin, and G. Pinard, Benzodiazepine use and the risk of motor vehicle crash in the elderly. J. Am. Med. Assoc. 278:27-31 (1997). 15. L.P. Raymond, B.W. Steele, and H.C. Walls. Benzodiazepines in Miami-Dade County, Florida driving under the influence (DUI) cases ( ) with emphasis on Rohypnol confirmation, patterns of use, psychomotor impairment, and results of Florida legislation. J. Anal. Toxicol. 23: (1999). 16. A. Boussairi, J.P. Dupeyron, B. Hernandez, D. Delaitre, L. Beugnet, P. Espinoza, and O. Diamant-Berger. Urine benzodiazepines screening of involuntary drugged and robbed or raped patients. J. Toxicol. Clin. Toxicol. 34: (1996). 17. J.E. Ray, D.K. Reilly, and R.O. Day. Drugs involved in self-poisoning: verification by toxicological analysis. Med. J. Aust. 144: (1986). 18. T. Rygnestad, K. Aarstad, K. Gustafsson, and U. Jenssen. The clinical value of drug analysis in deliberate self-poisoning. Hum. Exp. Toxicol. 9: (1990). 19. K. Michel, C. Knecht, I. Kohler, and M. Sturzenegger. Attempted suicide in the Bern region. Schweiz. Med. Wochenschr. 121: (1991)(in German). 20. J.G. Harvey and M.S. Cbristian. Deliberate overdosage in a hospital catchment area: preliminary results of a 7-year old study. Br. J. Acc. Emerg. Med. 1:24-25 (1986). 21. R.H. Adams. An accident and emergency department's view of self-poisoning: a retrospective study from the United Norwich Hospitals Hum. Toxicol. 5:5-10 (1986). 22. P.T. Mearrick and J.S. Chahl. Screening for benzodiazepines in urine after hydrolysis of glucuronide conjugates. Clin. Chem. 31: 153 (1985). 23. O. Beck, P. Lafolie, G. Odelius, and L. Boreus. Immunological screening of benzodiazepines in urine: improved detection of oxazepam intake. Toxicol. Lett. 52:7-14 (1990). 24. O. Beck, P. Lafolie, P. Hjemdahl, S. Borg, G. Odelius, and P. Wirbing. Detection of benzodiazepine intake in therapeutic doses by immunoanalysis for urine: two techniques evaluated and modified for improved performance. Clin. Chem. 38: (1992). 25. R. Meatheralh Benzodiazepine screening using EMIT II and TDx: urine hydrolysis pretreatment required. J. Anal. Toxicol. 18: (1994) 26. P. Simonsson, A. Liden, and S. Lindberg. Effect of 13-glucuronidase on urinary benzodiazepine concentration determined by fluorescence polarization immunoassay. Clin. Chem. 41.' (1995). 27. R. Meatherall. Optimal enzymatic hydrolysis of urinary benzodiazepine conjugates. J. Anal. Toxicol. 18: (1994). Manuscript received December 5, 2003; revision received June 2, 2004.