Introduction. Norlida Harun*, Robert A. Anderson, and Eleanor I. Miller. Abstract

Transcription

1 Validation of an Enzyme-Linked Immunosorbent Assay Screening Method and a Liquid Chromatography Tandem Mass Spectrometry Confirmation Method for the Identification and Quantification of Ketamine and Norketamine in Urine Samples from Malaysia Norlida Harun*, Robert A. Anderson, and Eleanor I. Miller Forensic Medicine and Science, Division of Cancer Sciences and Molecular Pathology, Faculty of Medicine, University of Glasgow, Scotland, G12 8QQ United Kingdom Abstract An ELISA and a liquid chromatography tandem mass spectrometry (LC MS MS) confirmation method were developed and validated for the identification and quantitation of ketamine and its major metabolite norketamine in urine samples. The Neogen ketamine microplate ELISA was optimized with respect to sample and enzyme conjugate volumes and the sample preincubation time before addition of the enzyme conjugate. The ELISA kit was validated to include an assessment of the dose-response curve, intra- and interday precision, limit of detection (LOD), and cross-reactivity. The sensitivity and specificity were calculated by comparison to the results from the validated LC MS MS confirmation method. An LC MS MS method was developed and validated with respect to LOD, lower limit of quantitation (LLOQ), linearity, recovery, intra- and interday precision, and matrix effects. The ELISA dose-response curve was a typical S-shaped binding curve, with a linear portion of the graph observed between 25 and 500 ng/ml for ketamine. The cross-reactivity of 200 ng/ml norketamine to ketamine was 2.1%, and no cross-reactivity was detected with 13 common drugs tested at 10,000 ng/ml. The ELISA LOD was calculated to be 5 ng/ml. Both intra- (n = 10) and interday (n = 50) precisions were below 5.0% at 25 ng/ml. The LOD for ketamine and norketamine was calculated statistically to be 0.6 ng/ml. The LLOQ values were also calculated statistically and were 1.9 ng/ml and 2.1 ng/ml for ketamine and norketamine, respectively. The test linearity was ng/ml with correlation coefficient (R 2 ) > 0.99 for both analytes. Recoveries at 50, 500, and 1000 ng/ml range from 97.9% to 113.3%. Intra- (n = 5) and interday (n = 25) precisions between extracts for ketamine and norketamine were excellent (< 10%). Matrix effects analysis showed an average ion suppression of 5.7% for ketamine and an average ion enhancement of 13.0% for norketamine for urine samples collected from six individuals. A comparison of ELISA and LC MS MS results demonstrated a sensitivity, specificity, and efficiency of 100%. These results indicated that a cutoff value of 25 ng/ml ketamine in the ELISA screen is particularly suitable and reliable for urine testing in a forensic toxicology setting. Furthermore, both ketamine and norketamine were detected in all 34 urine samples collected from individuals socializing in pubs by the Royal Malaysian Police. Ketamine concentrations detected by LC MS MS ranged from 22 to 31,670 ng/ml, and norketamine concentrations ranged from 25 to 10,990 ng/ml. The concentrations of ketamine and norketamine detected in the samples are most ikely indicative of ketamine abuse. Introduction Ketamine was first synthesized in 1962 by Calvin Stevens at the Parke Davis Laboratory in Detroit, Michigan. It was first marketed under the tradename Ketalar in 1970 following Food and Drug Administration (FDA) approval for human use as a short-acting general anesthetic drug that was also used for animals (1,2). Ketamine produces narcotic effects similar to phencyclidine (PCP) and hallucinogenic effects similar to lysergic acid diethylamide (LSD), making it popular with users who sought a dissociative experience (3). The recreational use of ketamine as a rave, party, and nightclub drug has widened, thus increasing public concerns about the potential hazards of this drug (2,4). Ketamine abuse was reported in the United Kingdom and Europe in the 1990s (5). In the United Kingdom, ketamine was classified as a Class C drug under the Misuse of Drugs Act 1971 and therefore carries up to 2 years imprisonment and/or an unlimited fine for possession and up to 14 years imprisonment and/or an unlimited fine for possession 310 Reproduction (photocopying) of editorial content of this journal is prohibited without publisher s permission.

2 to supply or supplying (6). Its misuse later spread to Asian countries, in particular, China, Malaysia, Taiwan, and Singapore (1). In the United States, ketamine has been classified as a club drug by the National Institute on Drug Abuse (NIDA) (7). The rapid growth of ketamine misuse worldwide has led to the development of various methods such as gas chromatography mass spectrometry (GC MS) (8 16), liquid chromatography (LC) MS (17 20), a combination of GC MS and LC MS (22), and emzyme-linked immunosorbent assay (ELISA) (10,15,20). Ketamine undergoes N-demethylation by liver microsomal cytochrome P450 enzymes CYP 3A4, CYP 2B6, and CYP 2C9 to form its primary metabolite, norketamine (22). Norketamine undergoes dehydrogenation to form dehydronorketamine, which is then conjugated with glucuronic acid before excretion in urine. The basic ketamine metabolic pathway is shown in Figure 1 (23). Urine continues to be a widely used specimen for the analysis of drugs of abuse in some situations, such as workplace testing, for a number of reasons including the non-invasive method of collection, simple sample pre-treatment and analysis, the large volume of sample that can be obtained for analysis, the wide drug detection window compared to blood, and the presence of the parent drug and metabolites in high concentrations compared to other matrices (24). Ketamine and norketamine can be detected in urine up to 5 and 6 days, respectively, after administration, and dehydronorketamine can be detected in urine for up to 10 days (25). Currently, ELISA is popular within the forensic toxicology community (26) and is, at the time of writing, the only immunoassay system available for the rapid, qualitative screening of ketamine and its metabolites (20). The test is fast, simple, can be automated to screen a large number of samples simultaneously, requires a small sample volume, and can be applied in the analysis of a range of biological matrices. LC MS MS has become a useful tool in the analysis of drugs of abuse in urine. In this study, electrospray ionization (ESI), a soft ionization technique in LC MS MS analysis, was applied in the selected reaction monitoring (SRM) mode to obtain better sensitivity and satisfy identification criteria requirements (24). ketamine N-demethylation norketamine hydroxylation and conjugation cyclohexanone glucuronide derivatives * the chiral point Figure 1. Basic ketamine metabolic pathway (23). dehydrogenation Although kg ketamine was seized by the Malaysian Royal Police in 2007 (27), ketamine was not included in routine investigations when drug abuse suspects were tested by the National Agency of Drug Abuse in Malaysia. Consequently, there are no data currently available on the extent of ketamine abuse in Malaysia. The aim of this current study was to develop and validate an ELISA screening method and an LC MS MS confirmation method to be used in tandem for the respective identification and quantitation of ketamine and norketamine in urine samples. This method would be very useful in a workplace testing or forensic toxicology setting. Preliminary data collected from this study provide information on the concentrations of ketamine and norketamine typically found in urine samples collected from individuals frequenting pubs and clubs in Malaysia. Materials and Methods Reagents, standards, and specimens Ketamine ELISA kits (product number: ) were purchased from Neogen (Lexington, KY). The kits contained 96 antibody coated microplate wells, wash buffer (phosphate buffer saline solution containing a surfactant, ph 7), ketamine enzyme conjugate labeled with horseradish peroxidase (HRP), 3,3',5,5'-tetramethylbenzidine (TMB) substrate solution with hydrogen peroxide (H 2 O 2 ), a red stop solution containing 1 N H 2 SO 4, and negative (0 ng/ml) and positive controls (2000 ng/ml) prepared in synthetic urine. Phosphate buffer saline (PBS) ph 7 was purchased from Immunalysis (Pomona, CA) and contained bovine serum albumin and non-azide preservatives. Ketamine, norketamine, ketamine-d 4, norketamine-d 4, and the drug standards used to test the cross-reactivity of the ELISA kit, amphetamine, methamphetamine, 3,4- methylenedioxyamphetamine (MDA), 3,4-methylenedioxymethylamphetamine (MDMA), 3,4-methylenedioxyethylamphetamine (MDEA), cocaine, benzoylecgonine, diazepam, nordiazepam, morphine, methadone, 6-monoacetylmorphine (6-MAM), and phencyclidine (PCP), were obtained from Promochem (Teddington, U.K.). Tiletamine HCl powder was purchased from Sigma- Aldrich (Dorset, U.K.). Ammonium acetate and ammonium formate were purchased from Fluka (Buchs, Switzerland). HPLCgrade formic acid, methanol, and acetonitrile were from BDH (Poole, U.K.). β-glucuronidase Type HP-2 from Helix pomatia (100 U/μL) was obtained from Sigma-Aldrich. Synergi Hydro RP column (150-mm length, 2.0-mm i.d., 4-µm particle size) was purchased from Phenomenex (Torrance, CA) along with a guard column (4.0 mm 2.0 mm, 5 µm) with the same packing as the main column. World Wide Monitoring Clean Screen columns (ZSDAU 020) used for drug extraction in the LC MS MS method dehydronorketamine 311

3 were purchased from United Chemical Technologies (Bristol, PA). Thirty-four urine samples were obtained from the Narcotic Department of the Royal Malaysian Police. All urine specimens (10 ml volume in each container) were freeze-dried at 54 C by the Drug Laboratory, Pathology Department, Kuala Lumpur General Hospital in Malaysia using a Cole Palmer 1 L Benchtop Freeze Dry System. The lyophilized samples were sent by courier to the forensic toxicology laboratory at the University of Glasgow. Each of the samples was reconstituted with 10 ml deionized water prior to analysis by ELISA and LC MS MS. These samples were previously screened using a qualitative GC MS method with a cutoff of 350 ng/ml at the Drug Laboratory, Pathology Department, Malaysia. These GC MS results were compared to the ELISA and LC MS MS results obtained in this study (Table I). Ten negative urine controls were obtained from volunteers among the Glasgow laboratory personnel. Microplate well ELISA Optimization of ELISA procedure Combinations of different sample volumes (20 and 40 µl), enzyme conjugate volumes (90 and 180 µl) and sample pre-incubation times before addition of enzyme conjugate (0, 30, and 60 min) were tested. A Miniprep 75 automatic pipettor purchased from Tecan (San Jose, CA) was used to dilute samples and pipette all calibrators, quality control samples and case samples into the microplate wells. Twenty microliters of diluted calibrator, quality control, or case sample (diluted 1:10 online with PBS) was loaded into the microplate wells in duplicate. After pipetting the samples, 180 µl of diluted enzyme conjugate (diluted 1:180, v/v, according to kit manufacturer instructions) was added into each microplate well. The microplate was then left in the dark at room temperature for 45 min. Then the wells were washed manually five times with 300 µl wash buffer to remove any unbound sample or residual conjugate reagent remaining in the wells. K-Blue substrate (150 µl) was then added to the microplate wells and left to incubate in the dark at room temperature for a further 30 min. The reaction was stopped by adding 50 µl stop solution (1 N H 2 SO 4 ) to each well, turning the contents yellow. The plate was read at a wavelength 450 nm using a Sunrise Remote EIA reader by Tecan (Grödlg, Austria). Method validation for ELISA Dose-response curve. A dose-response curve was generated for urine spiked at concentrations of 0, 10, 25, 50, 100, 500, 1000, 2000, 4000, and 8000 ng/ml ketamine. The 312 test was performed in duplicate and the data were expressed as mean of the B/B 0 (%) readings where B is the absorbance reading of the bound calibrator and B 0 is the absorbance value of the blank calibrator. The ketamine concentration range used in the dose-response curve spanned the range previously reported in the literature for ketamine concentrations detected in urine (2,8,10,12,14,15,17). Limit of detection (LOD). Ketamine was spiked at concentrations of 0, 0.5, 1, 2, 4, 5, 10, 25, and 50 ng/ml to establish a calibration curve. Ten negative urine samples were used in the calculation. The LOD absorbance value was calculated as the concentration having a signal-to-noise ratio of 3:1. This absorbance value was matched to the absorbance value produced by one of the ketamine-spiked calibrators on the graph. Intra- and interday precision. The intraday precision of B/B 0 (%) values was determined by spiking 10 1 ml Table I. GC MS, ELISA, and LC MS MS Urine Sample Results GC MS Qualitative ELISA LC MS MS LC MS MS Ratio of Sample Screening at Concentration Ketamine* Norketamine* Norketamine/ Number Cutoff 350 ng/ml (ng/ml) (ng/ml) (ng/ml) Ketamine 1 Positive > Positive > Positive > Positive > Positive > Positive > Positive > Positive > Positive > Positive > , Positive > Positive Positive > , Positive Positive Positive Positive > Positive > Positive > , Positive > Positive > , Positive > Positive > Positive > Positive > Positive > Positive > , Positive > Positive > Positive > Positive > Positive > Positive > , Positive > * High concentrations are those equal to or higher than the high control, 1000 ng/ml.

4 drug-free urine with 25 ng ketamine (cutoff value). Each spiked sample was analyzed in duplicate. The test was carried out on the same plate and on the same day. For interday precision, the intraday test was carried out on five different plates on five different days. The mean B/B 0 (%) was calculated for n = 50. Cross-reactivity study. Ketamine was spiked into blank urine to produce concentrations of 2, 4, 5, 10, 20, and 50 ng/ml and norketamine was spiked into other blank urine samples to produce concentrations of 25, 50, 75, 150, and 200 ng/ml. A blank urine sample was also prepared (B 0 ). The crossreactivity of norketamine to ketamine was calculated at 200 ng/ml by comparing the ketamine and norketamine calibration curves plotted in Microsoft Excel. Thirteen commonly abused drugs and tiletamine (an anesthesic with a similar chemical structure to ketamine) were individually tested in separate microplate wells in duplicate at a concentration of 10,000 ng/ml to test the kit cross-reactivity to these compounds. Sensitivity and specificity. The ELISA test sensitivity and specificity were calculated using Eq. 1 and 2. Sensitivity = (TP 100)/(TP + FN) Eq. 1 Specificity = (TN 100)/(TN + FP) Eq. 2 To define the equation parameters, a true-positive result (TP) produced both positive screening and confirmation results, a false-positive result (FP) produced a positive screening and negative confirmation result, a true-negative result (TN) produced a negative result for screening and confirmation, and a false-negative result (FN) was negative for screening but positive in the confirmation test. Procedure for sample analysis using ELISA Urine (100 µl) was pipetted into disposable borosilicate glass culture tubes (75 12 mm). The samples were vortex mixed then diluted 1:10 online with PBS ph 7 (900 µl). This procedure was performed to minimize potential urine matrix effects. Each ELISA run consisted of a set of calibrators including blank urine and three blank urine samples spiked to produce 25, 50, and 125 ng/ml ketamine. The ELISA administration cutoff was set at 25 ng/ml based on the LOD calculation. In addition, this concentration formed part of the linear portion of the S-shaped dose-response curve. Negative and positive controls (0 and 2000 ng/ml) were provided by the manufacturer to verify the performance of the ELISA test. These were diluted in the same manner as the calibrators and samples, vortex mixed, and distributed at the beginning and end of the plate to monitor assay performance. LC MS MS analysis Sample hydrolysis Standards were prepared by spiking 1 ml of blank urine with 50, 100, 200, 400, 800, and 1200 ng ketamine and norketamine standards. One blank urine with no standard or internal standard, and one blank urine containing 100 ng internal standards were also prepared. One milliliter of 1 M sodium acetate buffer (ph 5.0) was added to each tube along with 40 µl of β-glucuronidase crude enzyme solution (Helix pomatia) was added to each tube. The tubes were capped and placed in an oven at 60 C. After 3 h, the tubes were removed from the oven and left at room temperature to cool. After cooling, 3 ml phosphate buffer (0.1 M, ph 5.0) was added to each tube, and the mixture was vortex mixed. The ph was adjusted to ph 5.0 using 1 M potassium hydroxide. All tubes were centrifuged at 2500 rpm for 10 min prior to loading on solid-phase extraction (SPE) columns. Table II. Gradient Program of the Mobile Phase 3 mm Ammonium Time Formate % Acetonitrile Flow Rate (min) Formic Acid (%A) (%B) (µl/min) Table III. The LC MS MS Optimized Parameters for Ketamine and Norketamine Sheath Auxiliary Capillary Collision Precursor Product Compound Gas (AU) Gas (AU) Temperature ( C) Energy (%) Ions Ions Ketamine *, 207 Norketamine *, 206 * Quantitation ions. SPE World Wide Monitoring (ZSDAU020) SPE cartridges were first conditioned with 3.0 ml methanol, followed by 3.0 ml distilled water, and 1.0 ml phosphate buffer (0.1 M, ph 5.0) without the application of a vacuum. Then the urine samples were loaded onto the SPE cartridges. The cartridges were washed sequentially with 3 ml phosphate buffer (0.1 M, ph 5.0) and 1.0 ml acetic acid (1.0 M, ph 5.0) in an attempt to remove potential interferences present in the in urine matrix. The columns were dried thoroughly under full vacuum for 5 min. Two milliliters of methanol/aqueous ammonium hydroxide (98:2, v/v) was used to elute the analytes and deuterated internal standards. The SPE eluant was evaporated to dryness under nitrogen gas, using a heating block set at 40 C. Finally, the residues were reconstituted with 150 μl mobile phase and vortex mixed. LC MS MS set up A Surveyor HPLC system (Thermo Finnigan, San Jose, CA) linked to an LCQ Deca XP Plus ion trap MS (Thermo Finnigan) was used for detection. The equipment also came with a 313

5 Surveyor Autosampler (AS 3000) and a data processing system, Xcalibur 1.3. LC analysis was carried out using a mobile phase gradient program combining 3 mm ammonium formate buffer % formic acid (ph 3) and acetonitrile at a flow rate of 0.25 ml/min as shown in Table II. A 20-µL sample was injected onto the Synergi Hydro RP LC column using partial loop mode. The mass spectral data acquired for both analytes and their deuterated internal standards are based on electrospray (ESI) positive ion mode [M + H] +. The probe voltage used was 4.5 kv. The capillary temperature, sheath and auxiliary gas flow rates, and collision energies were optimized during tuning for each analyte. The optimized parameters are given in Table III. Internal standard data was collected by selected ion monitoring (SIM) for identification of the parent ions of ketamine-d 4 (m/z 242) and norketamine-d 4 (m/z 228) and analyte data was collected by selected reaction monitoring (SRM) over the mass range m/z SRM was used where one parent ion and two product ions were identified for a compound and this fulfilled the European Union requirement for identification and confirmation of illicit drugs (24). The precursor ions for ketamine and norketamine were at m/z 238 and m/z 224, respectively. The precursor and product ions are shown in Table III. The quantitation ion was the major product ion produced on precursor fragmentation. The ratios of quantitation ion to internal standard and qualitative ions to quantitation ion were calculated. For positive sample identification, the ratio of quantitation ion to internal standard was either greater than or within ± 20% of the ratio for the lowest calibration standard. The qualitative ion to quantitation ion ratio was also calculated, and the qualitative ion ratios were acceptable if they were within ± 20% of the corresponding calibrator or control. Method validation For the LC MS MS method, the parameters investigated as part of the validation procedure were limit of detection (LOD), lower limit of quantification (LLOQ), linearity, recovery, intra- and interday precision, accuracy, and urine matrix effects. Linearity and determination of LODs and LLOQs. Urine calibration standards were prepared by spiking the appropriate amount of the ketamine and norketamine stock solutions (1 and 10 ng/µl) into test tubes containing drug-free urine to provide final concentrations of ketamine and norketamine of 0, 25, 50, 100, 200, 400, 800, and 1200 ng/ml. 100 ng of deuterated IS was added to each tube with exception of the blank. Two replicate analyses were performed for each concentration to evaluate linearity for statistical purposes. The slope and standard error of the calibration curves were calculated and the peak-area ratios of analyte to deuterated internal standard were used for sample quantification through these calibration curves. 314 LODs were calculated statistically using Eq. 3 and 4, where y B is the intercept, s B is the standard error of the regression line, and m is the slope (28,29). y LOD = y B + 3s B Eq. 3 LOD = (y LOD y B )/m Eq. 4 LLOQs were calculated using a similar method, but 10 times the standard error of the regression line was used (Eq. 5 and 6). y LOQ = y B + 10s B Eq. 5 LOQ = (y LOQ y B )/m Eq. 6 Precision and accuracy. The intraday precision was determined by spiking five replicates of 50, 500, and 1000 ng/ml ketamine and norketamine standards in blank urine and analysing them on the same day. The interday precision was the same experiment as intraday only carried out on five different days to produce (n = 25) results. The accuracies were determined by comparing the mean calculated concentration of the five spiked urine samples with the target concentration. Matrix effect assessment. This study was conducted to assess the potential interference caused by the urine matrix during analysis. Three replicates of 1 ml blank urine from six individuals were run through the SPE procedure together with three replicates of 2 ml SPE loading buffer (0.1 M phosphate buffer, ph 5.0). Each of the SPE eluants was spiked with 50, 500, and 1000 ng/ml of ketamine and norketamine and 100 ng of internal standard after the extraction. The % matrix effect was calculated according to Eq. 7, where a and b are defined as the peak-area ratio of analyte to internal standard in Table IV. Experimental Conditions for Method Optimization Experimental Condition: Urine Diluted 1:10 (v/v) with PBS Volume of Volume of Sample sample enzyme pre-incubation Average (µl) conjugate (µl) time (min) B/B 0 (%) S.D. % R.S.D. 20* * 0* * Parameters used for method validation.

6 SPE loading solution without urine matrix and human urine matrix, respectively. Percentage of matrix effect = b/a 100% Eq. 7 A value of less than 100% indicates analyte ion suppression whereas a value of more than 100% indicates analyte ion enhancement. Both effects are attributed to the urine matrix effect. Recovery study. Ketamine and norketamine at 50, 500, and 1000 ng/ml were spiked into 1 ml of blank urine (n = 3) and extracted using the SPE procedure. Two unextracted standards were also prepared at each concentration without internal standards and were kept in the refrigerator throughout the extraction. Ketamine-d 4 and norketamine-d 4 internal standards (100 ng) were added to each tube before the samples were evaporated under nitrogen at 40 C. The % recovery was determined by comparing peak-area ratios obtained from extracted ketamine and norketamine versus the corresponding peak-area ratios of the same concentration of unextracted standards. Results ELISA results Method development Sample pre-incubation time and sample volume were not Figure 2. Ketamine ELISA dose-response curve in urine ( ng/ml). significant factors and different times and volumes demonstrated similar % B/B 0 results (± 13%). The most important variable was clearly the volume of enzyme conjugate used, which showed that a larger volume (180 µl as recommended by the kit manufacturers) was better than a smaller volume (90 µl), producing a decrease in B/B 0 (%) response of 10% (Table IV). An assay cutoff of 25 ng/ml was chosen because it produced an average B/B 0 (%) of 75%. If a calibrator with a higher B/B 0 (%) value was used, the number of false positives would be increased. In this study, no sample pre-incubation time, a smaller sample volume (20 µl) and a larger volume of enzyme conjugate (180 µl) were optimum and hence chosen for method validation This finding agrees with a study for optimization of an ELISA method for cocaine in hair done by Lachhenmier et al. (30). Method validation The ELISA validation procedure in this study was based on ELISA validation procedures published previously in the literature for the analysis of drugs of abuse within the forensic toxicology field (31 33). The ELISA kit manufacturers determined a plate sensitivity of 8 ng/ml ketamine in buffer compared to 10 ng/ml ketamine (1.25-fold lower sensitivity) in neat urine. Therefore, the blank, calibrators, and quality control samples were prepared in blank urine for this study in order to take the urine matrix effect into account. Dose-response curve. The B/B 0 (%) values were calculated where B is the absorbance value of the bound calibrator and B 0 is the absorbance value of the blank calibrator. Both the x and y-axes were converted into log-scales and the results, as expected, demonstrated the inverse relationship between concentration and absorbance. The higher the ketamine concentration, the lower the B/B 0 % due to the lower quantity of enzyme conjugate that binds to the antibody sites on the microplate wells compared to the analyte. The graph also indicated that the ketamine assay S-shaped binding curve had a linear portion between 25 and 500 ng/ml ketamine, which leveled off after 2000 ng/ml ketamine. (Figure 2). Based on the dose-response curve in which 25 ng/ml ketamine was the lowest concentration on the linear portion of the binding curve, this concentration was selected for the assay cutoff, which is in agreement with manufacturer s recommendations. Also, by selecting a range of calibrators related to the linear portion of Table V. LC MS MS Parameters for Ketamine, Norketamine, Ketamine-d 4, and Norketamine-d 4 in the ESI Positive Mode Precursor Ion Product Ions Collision Energy Compound (m/z) (m/z) (%) Ketamine *, Norketamine *, Ketamine-d *, Norketamine-d *, Figure 3. The cross-reactivity of norketamine to ketamine in the Neogen ELISA kit. * Quantitation ions (100% relative abundance compared to 10% relative abundance of precursor ion). 315

7 this graph, a semi-quantitative estimation of sample concentration was obtained. The absorbance response for the main metabolite norketamine stated in the manufacturer s kit insert and also in the cross-reactivity conducted for this study, were very low (4.6% and 2.1%, respectively) and was therefore not a significant factor in the preparation of the dose-response curve. Assay precision and LOD. The test demonstrated excellent intra- and interday precision results. The % R.S.D. for the intraday precision for 10 blank urine samples spiked with ketamine at 25 ng/ml was 2.5%, and the % R.S.D. for the interday precision for 10 samples spiked at the same concentration and analyzed on five different days (n = 50) was 4.8%. This study determined the LOD to be approximately 5.0 ng/ml ketamine which was lower than the LOD reported by the manufacturer in the kit insert. This LOD value was not used as the cutoff value of this test to minimize the number of false-positive (FP) results. An administrative cutoff of 25 ng/ml ketamine which produced a B/B 0 (%) value of approximately 75% was selected and used throughout the study in an attempt to make a better distinction between positive and negative results. Cross-reactivity studies. Referring to the manufacturer s insert, norketamine is the only related drug that cross-reacts with the ketamine assay. The manufacturer found that a concentration of ng/ml ketamine corresponded to a norketamine cross-reactivity of 4.6% towards the ketamine assay. The microplate is directed towards ketamine and cross-reacts 100% with it. As shown in Figure 3, 200 ng/ml norketamine cross-reacted to ketamine at 4.2 ng/ml. The cross-reactivity was calculated relative to ketamine calibration curve using Microsoft Excel, and the value was found to be 2.1%. The common drugs of abuse, which were tested at a concentration of 10,000 ng/ml, including amphetamine, methamphetamine, MDA, MDMA, MDEA, cocaine, benzoylecgonine, diazepam, nordiazepam, morphine, methadone, 6-MAM, PCP, and structurally similar anesthetic, tiletamine, did not produce absorbance values in the assay equal to or less than the assay cutoff level 25 ng/ml and therefore were noted as being non cross-reactive with this ELISA assay at this concentration. Case samples results: sensitivity and specificity of the ELISA assay. Forty-four samples (34 positive and 10 negative) were analyzed using the ketamine ELISA kit. The assay cutoff 25 ng/ml was chosen based on the doseresponse curve. If samples produced B/B 0 (%) 316 Blank urine extract absorbance greater than the cutoff concentration, the results were reported as negative. The assay is semi-quantitative and samples with B/B 0 (%) between the cutoff concentration (25 ng/ml) and the highest calibrator (125 ng/ml) were reported as values between these concentrations (i.e., reported as ng/ml). Some samples produced B/B 0 % values which were less than the highest calibrator and these were reported as > 125 ng/ml. Ten samples screened as negative, and Extracted 5 ng/ml ketamine and norketamine standards Positive case sample containing ketamine at 17,260 ng/ml and norketamine at 1040 ng/ml Figure 4. Chromatograms of blank urine (A), extracted 5 ng/ml ketamine and norketamine standard (B), and positive ketamine and norketamine case sample (C). A B C

8 34 samples screened as positive. Using 25 ng/ml ketamine as the cutoff and based on the comparison of the ELISA and the LC MS MS results (Table I), the sensitivity for the test was 100% and the specificity was 100% according to Eq. 1 and 2. ELISA is a semi-quantitative assay and the sensitivity and specificity were calculated based solely on a positive or negative results and did not take ELISA concentrations into account. For example, norketamine was present in sample 15 at 1080 ng/ml with a cross-reactivity of 2.1%, which translates into 23 ng/ml plus the 22 ng/ml ketamine in the sample, it is not surprising that the sample screened at that level. A similar result was obtained for sample 23 because of the cross-reactivity of norketamine. LC MS MS results The optimum tuning parameters, precursor and product quantitation and qualification ions are shown in Table V. These ions were used for method validation. The product ion chromatograms and spectra for blank urine Table VI. Analytical Characteristics of LC ESI-MS MS Method for Ketamine and Norketamine in Urine Average Linear Number of Coefficient (R 2 ) Range Calibration Linear Correlation Compound (ng/ml) Points Equation (n = 3) Ketamine y = 0082x Norketamine y = x without analyte or deuterated internal standards, 5 ng/ml ketamine and norketamine standard plus deuterated internal standard and a positive case sample for ketamine and norketamine plus deuterated internal standard are shown in Figure 4. The chromatograms show, from top to bottom, precursor ions for ketamine and norketamine, two product ions and a deuterated internal standard precursor ion. Linearity, LOD, and LLOQ The correlation coefficients (R 2 ) of the calibration curves were greater than 0.99 for both ketamine and norketamine (n = 3). The linearity of the LC ESI-MS MS method was evaluated within the range ng/ml, using eight points across the curve. The average slopes, intercepts, and R 2 values are summarized in Table VI and Figure 5. The LODs for ketamine and norketamine were both approximately 0.6 ng/ml as calculated statistically according to Eq. 3 and 4. The LLOQ values were determined as 1.9 ng/ml for ketamine and 2.1 ng/ml for norketamine. Intra- and interday precision and accuracy The intraday precision and accuracy were evaluated by analyzing five replicates of three spiked samples in the same day. The precision between extracts was calculated as the % relative standard deviation (% R.S.D.). Interday precision and accuracy were determined by analyzing five spiked samples on five different days (n = 25). The % R.S.D.s for the intra- and interday precision are shown in Table VII. The % accuracy was determined by comparing the mean calculated concentration of the spiked urine samples with the target concentration. The intraand interday accuracies for all the samples ranged from 96.6% to 105.2%. Therefore, the precision and accuracy values were excellent and were well within the recommendations issued by the Society of Forensic Toxicologists of ± 20% (34). Figure 5. Calibration curves for ketamine and norketamine. Table VII. Intraday and Interday Imprecision and Accuracy of the Spiked Samples of Ketamine and Norketamine in Urine Spiked Intraday Imprecision (n = 5) Interday Imprecision (n = 25) Concentration Accuracy R.S.D. Accuracy R.S.D. Analyte (ng/ml) Mean ± S.D. (%) (%) Mean ± S.D. (%) (%) Ketamine ± ± ± ± ± ± Norketamine ± ± ± ± ± ± Matrix effect analysis Most researchers include the assessment of matrix effects in their LC MS MS method development to ensure that the chromatographic separation developed using different matrices is not affected or that the separation shows minimal or acceptable effects, at least when doing quantitative analysis (35). The matrix effects caused by the interferences in the urine were acceptable for both ketamine and the metabolite nor-ketamine. Matrix effects analysis for ketamine showed ion suppression of 8.6% for 50 ng/ml, 4.7% for 500 ng/ml, and 4.0% for 1000 ng/ml. Norketamine showed ion enhancement of 19.7% for 50 ng/ml, 6.3% for 500 ng/ml, and 12.9% for 1000 ng/ml. A summary of the observed matrix effects is given in Table VIII. Recovery studies Ketamine and norketamine recoveries in human urine samples for low (50 ng/ml), medium (500 ng/ml), and high (1000 ng/ml) concentrations are presented in Table IX. The recoveries for both analytes at all three con- 317

9 centrations were > 97%, with excellent % RSD values < 8%. Application of LC MS MS method to urine samples Thirty-four urine samples obtained from the Royal Malaysian Police were tested using a calibration curve with ketamine and norketamine concentrations ranging from 0 to 1200 ng/ml. Case samples which contained ketamine and norketamine concentrations greater than 1200 ng/ml were diluted 10 times or in some cases 100 times with phosphate buffer (0.1 M, ph 5.0) and reanalyzed so that the diluted concentration fell within the set calibration concentration range. A urine specimen was considered positive by LC MS MS if ketamine or norketamine was detected at a concentration higher than the method LLOQ of 2 ng/ml. For the LC MS MS analysis, a concentration equal to or more than the high control (1000 ng/ml) is defined as a high concentration. All 34 samples were confirmed positive by LC MS MS with most of them containing high concentrations of ketamine and norketamine. The data showed no consistent ratio of norketamine to ketamine. The average ratio was 14.2 and the ratio median was 4.3. The ratio of norketamine to ketamine ranged from to 82.2 with norketamine being higher than ketamine in 27 out of 34 samples. The results are listed in Table I. Discussion Immunoassay tests for drugs of abuse screening require to be rapid and efficient because of the large number of samples typically encountered in forensic toxicology laboratories. It is also extremely important that these screening tests are highly sensitive and specific. In this study the ELISA test was quick and simple to perform (3 h for analysis). The optimized test conditions required a small sample volume of diluted urine (20 µl diluted urine) with no sample pre-incubation time included. The manufacturer s kit insert claimed that the ELISA test sensitivity was better in buffer compared to urine with a B/B 0 % approximately 20% lower in buffer than in urine. In this study the comparison was tested using spiked ketamine at 25 ng/ml in buffer (n = 3) and urine (n = 3); with a blank buffer and a blank urine included and the results were found to be similar to the manufacturer s findings. This was in agreement with Huang et al. (20) who found that, when using the Neogen kit for urine samples, the lower concentration ketamine standards generated higher B/B 0 (%) values than the same standard concentration prepared in buffer. Conversely, the authors also observed that for higher concentration ketamine standards, the B/B 0 (%) was lower than for same standard concentration prepared in buffer. The test was also very specific to ketamine and showed a low cross-reactivity to norketamine and no cross-reactivity to 13 commonly tested drugs of abuse at a concentration of 10,000 ng/ml. In the case of sample 23, in which the ketamine concentration determined by ELISA was 33 ng/ml and the norketamine concentration was 740 ng/ml, the large norketamine concentration resulted in a relatively large cross-reactivity to produce an ELISA response greater than 125 ng/ml. Huang et al. (20) found that the Neogen Ketamine ELISA kit showed some cross-reactivity (0.3%) to the other ketamine metabolite, dehydronorketamine (DHNK), and that this decreased with increasing concentrations of DHNK (20). The ketamine ELISA assay was shown to be a sensitive test (LOD 5.0 ng/ml), in agreement with the manufacturer s findings. The LOD was not used as the cutoff value, however to minimize the number of false-positive results. Ketamine is a weakly basic drug (pk a 7.5) and is metabolized into glucuronide conjugates during phase II metabolism before excretion in urine. Because of low cross-reactivity of norketamine and dehydronorketamine, the use of a 25 ng/ml ketamine cut off could result in some case samples with higher concentrations of norketamine or dehydronorketamine or their glucuronide producing a negative screening result if the ketamine concentration in the sample is low. The cutoff value as determined for this assay was shown to be fit for purpose and had no norketamine cross-reactivity issues at the levels that were detected for these particular samples. However, this may pose a problem in a clinical setting involving low therapeutic concentrations of ketamine. One possible solution in this scenario would be to include a hydrolysis step in the ELISA protocol in order to release conjugated ketamine and norketamine. Wieber et al. (36) found that a very small percentage of unchanged ketamine (2.3%), norketamine (1.6%), and Table VIII. Matrix effects for Ketamine and Norketamine in Urine (n = 6 individuals) Spiked Concentration Matrix Effect (%) Analyte (ng/ml) [R.S.D.] Ketamine [2.3] [10.4] [5.1] Norketamine [13.2] [11.9] [10.7] Table IX. Recoveries for Ketamine and Norketamine in Human Urine Samples Spiked Concentration % Recovery Analyte (ng/ml) (n = 5) S.D. % R.S.D. Ketamine Norketamine

10 dehydronorketamine (16.2%) are eliminated in urine, whereas 80% is present as the glucuronide conjugates of hydroxylated metabolites of ketamine. The aim of this study was to detect total and free ketamine and norketamine in urine samples. Although LC MS MS analysis could en-compass the detection of glucuronide conjugates (37), there were no ketamine or norketamine glucuronides standards available in the commercial market at the time of conducting this study. The LC MS MS method detected free ketamine and norketamine in the urine samples and, because the sample concentrations were unknown at the time of testing, it was thought best to enhance ketamine and norketamine sensitivity by cleaving the glucuronides. A previously published study by Kronstrand et al. (37) determined that direct LC MS MS analysis of buprenorphine and norbuprenorphine glucuronides without hydrolysis was only suitable for a screening method and this could be compared to the ELISA method in this current study. Their study found that inclusion of a hydrolysis step and SPE clean up increased the sensitivity of the LC MS MS method 20-fold (from 20 to 1 µg/l) for both analytes (37). With this procedure, the LODs for ketamine and norketamine obtained were low, (both approximately 0.6 ng/ml), and the LLOQ values were 1.9 and 2.1 ng/ml, respectively. SPE was used to extract ketamine and norketamine. World Wide Monitoring Clean Screen columns (ZSDAU 020) columns have been used in many studies for detection of drugs of abuse with good recoveries (38,39). The SPE method selected operated by a mixed-mode cationic exchange mechanism based on the sorbent composition of C8 chains and benzene sulfonic acid (BSA) residues. Therefore, ketamine and its metabolite norketamine are most likely retained on the column via both hydrophobic and ionic interactions resulting in high recoveries for both analytes. The SPE step also helped in obtaining low LODs and reducing the matrix effects as shown for the chromatogram of the blank urine sample in Figure 3 (35). Both ketamine and norketamine were detected by LC MS MS in all of the 34 urine specimens tested from the Royal Malaysian Police. However, the concentrations of ketamine and norketamine in each sample vary widely with no consistency of ratios between ketamine and norketamine. This could be caused of a number of factors such as dose and route of administration of ketamine, the time interval between administration and collection of the urine sample, the subject s rate of metabolism, weight and health; thus the duration of the drug in the body and severity of effects varies from one person to another (40). The concentration range of ketamine in the urine samples was between 22 and 31,670 ng/ml, and the concentration of norketamine was between 25 and 10,990 ng/ml. Wieber et al. (36) reported that the half-lives of ketamine and norketamine in urine were 3.37 ± 0.14 h and 4.21 ± 0.35 h, respectively, and that both analytes were undetectable after 22 h (approximately 1 day). Therefore those samples which have ketamine concentrations greater than norketamine in this study were probably collected soon after use. The high concentrations of ketamine and norketamine in most samples may suggest regular intake of ketamine by those users (20). In this study, a comparison of ELISA and LC MS MS results has been made for ketamine and norketamine detection in urine. The linear range of Neogen ELISA kit for ketamine was evaluated as ng/ml Another paper which used an ELISA cutoff of 100 ng/ml ketamine as well as a GC MS method with a cutoff of 2.6 ng/ml demonstrated poor correlation between methods (less than 30%) in 43 samples (21). In another study (15), the data demonstrated 90.9% sensitivity and 98.9% specificity with 1% false-positive results when an ELISA cutoff concentration of 10 ng/ml ketamine was used with a GC MS cutoff value of 15 ng/ml ketamine. In this study, the use of an ELISA cutoff concentration of 25 ng/ml ketamine and an LC MS MS cutoff concentration of 2 ng/ml yielded 100% sensitivity and specificity on comparison of the two methods. The combination of an ELISA screening test and an LC MS MS confirmation analysis produced an acceptable, highly sensitive, specific and efficient system for the determination of ketamine and norketamine in urine samples. Conclusions A simple, rapid, and efficient ELISA test for ketamine has been optimized and validated for the analysis of ketamine and norketamine in urine samples. The Neogen ELISA kit is extremely sensitive and specific for ketamine screening at a cutoff concentration of 25 ng/ml coupled with an LC MS MS cutoff of 2 ng/ml. The ELISA test cross-reacts with ketamine by 100% and demonstrated some cross-reactivity to its main metabolite norketamine. The kit demonstrated excellent precision for ketamine. An LC MS MS confirmation method was validated for the quantitation of ketamine and norketamine in human urine. The method demonstrated excellent linearity, LOD, LLOQ, accuracy, and precision with acceptable matrix interference effects. The screening efficiency of ELISA has been validated and evaluated together with this LC MS MS confirmation method using 34 urine specimens collected from suspected drug users were tested and found to be positive. The combination of tests demonstrated excellent efficiency; sensitivity and specificity, with no false positive and false negative results for this particular set of samples. Therefore, a combination of ELISA and LC MS MS can be used reliably as components of a two-approach routine test strategy (screening and confirmation) for the determination of ketamine in urine specimens. This study also found that ketamine and norketamine are present in all specimens with 47% ketamine and 79% norketamine present at high concentrations (> 1000 ng/ml) and highlighted that ketamine is being abused in Malaysia. Acknowledgments This study was supported by a scholarship from the Depart- 319

11 ment of Public Service and the Ministry of Health, Malaysia for the corresponding author s postgraduate study in forensic toxicology at the University of Glasgow, Scotland. The authors also gratefully thank the Narcotic Department within the Royal Malaysian Police for the provision of urine samples. References 1. H.S. Leong, N.L. Tan, C.P. Lui, and T.K. Lee. Evaluation of ketamine abuse using hair analysis: Concentration trends in Singapore population. J. Anal. Toxicol. 29: (2005). 2. A.C. Lua, H.R. Lin, Y.T. Tseng, A.R. Hu, and P.C. Yeh. Profiles of urine samples from participants at rave party in Taiwan: prevalence of ketamine and MDMA abuse. Forensic Sci. Int. 36: (2003). 3. K.A. Moore, E.M. Kilbane, and R. Jones. Tissue distribution of ketamine in a mixed drug fatality. J. Forensic Sci. 42(6): (2007). 4. H.V. Curran and C. Morgan. Cognitive, dissociative and psychogenic effects of ketamine in recreational users on the night of drug use and 3 days later. Addiction 95(4): (2000). 5. European Monitoring Centre for Drugs and Drug Addition (EMCDDA). Report on the Risk Assessment of Ketamine in the Framework of the Joint Action on New Synthetic Drugs. ne_final_riskassessment_report.pdf 6. A.R.W. Jackson and J.M. Jackson. Forensic Science, 2nd ed. Pearson Education, Harlow, U.K., J.A. Morgan, M. Ricelli, C.H. Maitland, and H.V Curran. Long term effects of ketamine: evidence for a persisting impairment of source memory in recreational users. Drug Alcohol Depend. 75: (2004). 8. R.L. Lin and A.C. Lua. Detection of acid-labile conjugates of ketamine and its metabolites in urine samples collected from pub participants. J. Anal. Toxicol. 28: (2004). 9. A. Negrusz, P. Adamowicz, and B.K. Saini. Detection of ketamine and norketamine in urine of nonhuman primates after a single dose of ketamine using microplate enzyme-linked immunosorbent assay (ELISA) and NCI-GC MS. J. Anal. Toxicol. 29: (2005). 10. M.E.C. Tan, H.Y. Moy, L.C.P. Moy, and T.K. Lee. Evaluation of an ELISA test kit in the screening of ketamine in urine. Forensic Sci. Int. 136(suppl 1): 305 (2003). 11. S.L.Chou, M.H. Yang, Y.C. Ling, and Y.S. Giang. Gas chromatography-isotope dilution mass spectrometry preceded by liquid liquid extraction and chemical derivatization for the determination of ketamine and norketamine in urine. J. Chromatogr. B 79: (2004) 12. M.K. Huang, C. Liu, J.H. Li, and S.D. Huang. Quantitative detection of ketamine, norketamine and dehydronorketamine in urine using chemical derivatization followed by gas chromatography mass spectrometry. J. Chromatogr. B 820: (2005). 13. E.M. Kim, J.S. Lee, S.K. Choi, M.A. Lim, and H.S. Chung. Analysis of ketamine and norketamine in urine by automatic solidphase extraction (SPE) and positive ion chemical ionization gas chromatography mass spectrometry. (PCI GC MS). Forensic Sci. Int. 174: (2008). 14. C.H. Wu, M.H. Huang, S.M. Wang, C.C. Lin, and R.H. Liu. Gas chromatography mass spectrometry analysis of ketamine and its metabolites a comparative study on the utilization of different derivatization groups. J. Chromatogr. A 1157: (2007). 15. P.S. Cheng, C.Y. Fu, C.H. Lee, C. Liu, and C.S. Chien. GC MS quantification of ketamine, norketamine, and dehydronorketamine in urine specimens and comparative study using ELISA as the preliminary test methodology. J. Chromatogr. B 852: (2007). 16. P. Xiang, M. Shen, and X. Zhuo. Hair analysis for ketamine and its metabolites. Forensic Sci. Int. 162: (2006). 17. K.A. Moore, J. Sklerov, B. Levine, and A.J. Jacobs. Urine concentrations of ketamine and norketamine following illegal consumption. J. Anal. Toxicol. 25: (2001). 18. J.Y.K Cheng and V.K.K. Mok. Rapid determination of ketamine in urine by liquid chromatography tandem mass spectrometry for a high throughput laboratory. Forensic Sci. Int. 142: 9 15 (2004). 19. A.C. Lua and H.R. Lin. A rapid and sensitive ESI-MS screening procedure for ketamine and norketamine in urine samples. J. Anal. Toxicol. 28: (2004). 20. M.H. Huang, M.Y. Wu, C.H. Wu, J.L. Tsai, H.H. Lee, and R.H. Liu. Performance characteristics of ELISA for monitoring ketamine exposure. Clin. Chim. Acta 379: (2007). 21. P. Adamowicz and M. Kala. Urinary excretion rates of ketamine and norketamine following therapeutic ketamine administration: method and detection window considerations. J. Anal. Toxicol. 28: (2004). 22. Y. Hijazi and R. Bolieu. Contribution of CYP3A4, CYP2B6 and CYP2C9 isoforms to N-methylation of ketamine in human liver microsomes. Drug Metab. Dispos. 30: (2002). 23. R.C. Baselt. Disposition of Toxic Drugs and Chemicals in Man, 6th ed. Biomedical Publications, Foster City, CA, T.M. Pizzolato, M.J. Lopez de Alda, and D. Barceló. LC-based analysis of drugs of abuse and their metabolites in urine. Trends Anal. Chem. 26: (2007). 25. M.C. Parkin, S.C. Turfus, N.W. Smith, J.M. Halket, R.A. Braithwaite, S.P. Elliott, M.D. Osselton, and D.A. Cowan. Detection of ketamine and its metabolites in urine by ultra high pressure liquid chromatography tandem mass spectrometry. J. Chromatogr. B 876: (2008). 26. A.E. Elian. ELISA detection of clonazepam and 7-aminoclonazepam in whole blood and urine. Forensic Sci. Int. 134: (2003) &title=narkotik%20-%20rampasan 28. A.I. Al-Asmari and R.A. Anderson. Quantification of opioids and their metabolites in autopsy blood by liquid chromatography tandem mass spectrometry. J. Anal. Toxicol. 31: (2007). 29. M.M. Ariffin and R.A. Anderson. LC MS/MS analysis of quaternary ammonium drugs and herbicides in whole blood. J. Chromatogr. B 842: (2006). 30. K. Lachenmeier, F. Musshoff, and B. Madea. Determination of opiates and cocaine in hair using automated enzyme immunoassay screening methodologies followed by gas chromatographic mass spectrometric (GC MS) confirmation. Forensic Sci. Int. 159: (2006). 31. E.I. Miller, H.J. Torrance, and J.S. Oliver. Validation of the Immunalysis microplate ELISA for the detection of buprenorphine and its metabolite norbuprenorphine in urine. J. Anal. Toxicol. 30: (2006). 32. M. Laloup, G. Tilman, V. Maes, G. De Boeck, P. Wallemacq, J. Ramaekers, and N. Samyn. Validation of an ELISA-based screening assay for the detection of amphetamine, MDMA and MDA in blood and oral fluid. Forensic Sci. Int. 153: (2005). 33. E.I. Miller, F.M. Wylie, and J.S. Oliver. Detection of benzodiazepines in hair using ELISA and LC ESI-MS MS. J. Anal. Toxicol. 30: (2006) E. Chambers, D.M. Wagrowski, Z. Lu, and J.R. Mazzeo. Systematic and comprehensive strategy for reducing matrix effects in LC/MS/MS analyses. J. Chromatogr. B 852: (2007). 36. J. Wieber, R. Gugler, J.H. Hengstmann, and H.J. Dengler. Pharmacokinetics of ketamine in man. Anaesthesia 24: (1975). 37. R. Kronstrand, T.G. Selden, and M. Josefsson. Analysis of buprenorphine, norbuprenorphine and their glucoronides in urine by LC MS. J. Anal. Toxicol. 27: (2003). 320