Direct Analysis of Opiates in Urine by Liquid Chromatography-Tandem Mass Spectrometry

[ TechnicalNote Direct Analysis of Opiates in Urine by Liquid Chromatography-Tandem Mass Spectrometry Leslie E. Edinboro 1,2,*, Ronald C. Backer 3, and Alphonse Poklis 2 7Department of Pharmaceutics, School of Pharmacy, 2Department of Pathology, School of Medicine, Virginia Commonwealth University, Richmond, Virginia and 3Ameritox, Ltd., Midland, Texas Abstract I A method for the direct analysis of 10 opiate compounds in urine was developed using liquid chromatography-mass spectrometry-mass spectrometry (LC-MS--MS) with electrospray ionization interface (ESI). Opiates included were morphine-3-~glucuronide, morphine-6-~glucuronide, morphine, oxymorphone, hydromorphone, norcodeine, codeine, oxycodone, 6-monoacetylmorphine (6MAM), and hydrocodone. Urine samples were prepared by centrifugation to remove large particles and direct injection into the LC-MS-MS. Separation and detection of all compounds was accomplished within 6 rain. Linearity was established for all opiates except 6MAM from 50 ng/ml to 10,000 ng/mt; 6MAM from 0.25 ng/mt to 50 ng/ml with all correlation coefficients (r) > 0.99. Interrun precision (%CV) ranged from 1.1 % to 16.7%, and intrarun precision ranged from 1.3% to 16.3%. Accuracy (% bias) ranged from -7.3% to 13.6% and -8.5 % to l!.8 for inter- and intrarun, respectively. Eighty-nine urine samples previously analyzed by gas chromatography-ms were re-analyzed by the LC-MS-MS method. The qualitative results found an 88% agreement for negative samples between the two methods and 94% for positive samples. The LC-MS-MS method identified 19 samples with additional opiates in the positive samples. Overall, the direct injection LC-MS-MS method performed well and permitted the rapid analysis of urine samples for several opiates simultaneously without extensive sample preparation. Introduction The use of liquid chromatography-mass spectrometry (LC-MS) for the analysis of opiates has increased since first being reported in 1995 (1). One application has been in the area of pharmacokinetic studies (1-5). These reports have primarily focused on morphine and its glucuronide metabolites. Oxycodone (6), hydrocodone, and hydromorphone (7) methods have been reported as well. As opiates play a significant role in 9 Author to whom correspondence should be addressed. Current address: Commonwealth of Virginia, Department of Forensic Sciences, Central Laboratory, 700 North 5lh Street, Richmond, Virginia 23219. E-mail: les.edinboro@dfs.virginia.gov. the field of forensic toxicology, methods for their determination in various biological matrices of forensic importance have been published (8-12). Most of these methods employ some kind of sample preparation technique such as liquid-liquid or solid-phase extraction. Fitzgerald et al. (10) utilized a sophisticated column switching technique to purify directly injected urine prior to introduction into the LC-MS. This application was for broad urine drug screening and quantitation of the screened opiates was not presented. More recently, Dams et al. (11) reported a simple protein precipitation using acetonitrile for the analysis of 27 compounds in oral fluids. Heroin, morphine, codeine, and their metabolites were included in this application. Protein precipitation techniques or "dilute-andshoot" provide a rapid and simple sample pretreatment, but suffer from the potentially adverse effects of ion suppression discussed by Darns et al. (11). Ion suppression effects or matrix effects result in loss of signal due to competition for available charge within the ionization source. Electrospray ionization (ESI) is more susceptible to matrix effects; therefore, Dams et al. (12) applied atmospheric pressure chemical ionization (APCI) LC-MS-MS in developing a urine drug screening method utilizing the direct injection of urine. The opiates included in this method were morphine, normorphine, codeine, norcodeine, acetylcodeine, heroin, and 6-monoacetylmorphine (6MAM). This report presents a method for the analysis of 10 opiates, including the glucuronide metabolites of morphine, by the direct injection of urine and ESI-LC-MS-MS. Experimental Materials LC-MS-MS. Hydrocodone bitartrate salt, oxycodone HC1, morphine-3-glucuronide (M3G), and morphine-6-glucuronide (M6G) were purchased from Sigma (St. Louis, MO). Morphine, oxymorphone, codeine, norcodeine, and 6MAM were purchased from Cerilliant Corporation (Round Rock, TX) as 1.0 mg/ml standards in methanol except 6MAM, which is in acetonitrile. 704 Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.

Table I. LC-MS-MS Acquisition Parameters Dwell Collision Cone Segment Precursor Product Time Energy Voltage Time Channel (m/z) (m/z) (s) (ev) (V) (min) M3G 462.00 286.00 0.50 30 40 0.00-3.10 M6G 462.00 286.00 0.50 30 40 3.10-4.25 Morphine-d3 289.00 16 0.10 30 40 3.10-4.25 Morphine 285.90 200.90 0.10 30 40 3.10-4.25 Oxymorphone 302.00 226.90 0.10 30 40 3.10-4.25 Hydromorphone 286.10 184.50 0.10 30 40 4.00- Norcodeine 286.10 22 0.10 30 40 4.00- Codeine 300.00 198.80 0.10 30 40 4.00- Oxycodone 316.00 240.90 0.10 30 40 4.90-6.00 6MAM-d6 334.00 164.90 0.25 40 40 4.90-6.00 6MAM 328.00 164.90 0.25 40 40 4.90-6.00 Hydrocodone 300.00 198.80 0.10 30 40 4.90-6.00 R S 1 5 0 1 0. 2 5 9 6-MONOACETYLMORPHINE- d~ 01... i... ~... ~... t... ~...,....... i...,, ~ 6-MONOACETYLMORPHINE 100~ OXYCODONE lt~'/on0 ~ HYDROCODONE 5.19 334 9 164.9, 6.72e3 4: MRM 0{' 4 Channels ES+ 5.21 328 > 164.9 170 5.11 316 9 240.9 5.15e3 5.45 300 9 198.8 \ 2.81e4 ln.,... 4.77 300 > 198.8 CODEINE // 1.42e4... i... i... p... i... i... i... q... r... r... i... ~... i... J... i... J lnn_. 4.24 286.1 9 184.5 o;o. ~o~),~ HYDROMORPHONE A 2.86e4... t... i... i... i... ~...,... i... i... f... i... ~... i... ~... J..., 101._. 4.50 286.1 9 225 ;~ NORCODEINE 4.25A 1.81e3 ol... ~.J,t............................. MIP110903002 1,,,n~ 3.99 302 9 228.9 OXYMCRPHONE.~ 3.t4e3 10 n- 3.83 :/:~ MORPHINE-d3 A 0... i... i... i... i... i... ~... i..., 3.85 10/o0 ~ MORPHINE... ~,, 289 9 165 1.6665... i... i... i... i... i... r... i 285.9 9 200.9 5.20e3 0 `1... i... i... i... ~... J... J..,... ~... i... ~... J... i... J... J 100-~ 3.51 462 > 286 'Y~ MORPHIN E-6-GLUCURON IDE A 600 01...,,-,~, ~, i,.i... i... i,, J, '..i... ~... i... i... i...,... i... i 1: MRM of 1 Channel ES+ 10 2.66 ~AO~ A MORPHINE',.~GLUCURONIDE 462 > 380288 ~! \ 3.oi 1.00 2.00 3.00 4.00 8.00 7.00 Time (min) Figure 1, Non-extracted standard containing 0.25 ng/ml 6MAM and 50 ng/ml all other opiates. Deuterated internal standards (IS), morphine-d3 (1.0 mg/ml in methanol), and 6MAM-d6 (100 IJg/mL in acetonitrile) were purchased from Cerilliant. Blank pooled urine was provided by the Bioanalytical Laboratory, Virginia Commonwealth University and Quality Assurance Services (QAS, Augusta, GA). Formic acid (99%) and ammonium formate were purchased from Aldrich Chemicals (Milwaukee, WI). Acetonitrile and methanol were B&J brand high-purity solvents from Burdick and Jackson (Muskegon, MI). Deionized water (DI H20) was produced by a NANOpure Diamond TM water purification system from Barnstead International (Dubuque, IA). Gas chromatography (GC)-MS. All primary reference materials of oxycodone, oxymorphone, and other drugs were obtained as 1.0 mg/ml methanolic solutions from Cerilliant Corporation. Drug-free urine and GC-MS urine controls were obtained from QAS. Apparatus LC-MS-MS. Two Shimadzu (Columbia, MD) LC-10AVP pumps were connected through a high-pressure mixing tee with a 0.2-1Jm inline filter prior to the autosampler. Mobile phases were passed through a Shimadzu DGU-14 solvent degasser prior to entering the pumps. The autosampler was from LEAP Technologies (Carrboro, NC) with CTC Analytics HTS PAL with Peltier cooling stacks and a TRIOS switching valve. A 10-1JL syringe was fitted in the autosampler. A Shimadzu SCL-10 controller was integrated to control communication with the pumps, autosampler, and the MS. The MS was a Waters (Milford, MA) Micromass Quattro LC triple-quadrupole with an electrospray ionization (ESI) interface. Liquid N2 was used to produce the desolvation/nebulizer gases needed. Argon was used as the collision gas. MassLynx version 3.5 was installed as the operating software performing data acquisition and processing functions. The analytical column was an Agilent Zorbax SB Phenyl (2.1 mm 150 mm, 5-1Jm particle size). A 0.2-1Jm inline filter was connected directly to the frontend of the analytical column using a zero dead volume connector. The HPLC column was placed in a Jones Chromatograhy (Cardiff, Wales) column heater and maintained at 30~ The mobile 705

phase was established as A: 95% (0.05% formic acid with 10mM ammonium formate in DI-H20) and 5% (0.05% formic acid in acetonitrile) and B: 95% (0.05% formic acid in acetonitrile) and 5% (0.05% formic acid with 10mM ammonium formate in DI- H20). Flow was 0.25 ml/min. The following gradient was used: 0.00-0.50 min, 0% B, linear gradient to 18% B at 3.00 min, hold 18% B for 2.5 rain, then return to 0% B. Injection volume was 5 tjl, and autosampler temperature was set at 6~ Autosampler wash solutions were Wash 1: methanol and Wash 2: Mobile Phase A. An Eppendorf refrigerated centrifuge was used to centrifuge all samples prior to injection. GC-MS. The samples were analyzed on an Agilent (Pale Alto, CA) 6890 GC with a split/splitless injection port, a 7673 auto-sampler and a 5973A mass selective detector (MSD). The column was an HP-5 capillary column (5.0 m x 0.1-mm i.d., 0.40-1Jm film thickness). Column flow rate was 1.0 ml/min; inlet pressure at 60.53 psi; and injection was in the 4:1 split mode with a split flow of 4.0 ml/min. The oven temperature program was as follows: initial 170~ for 0 min, then ramped to 280~ at 30~ and held for 0.33 min. Under these conditions, the retention times of opiate TMS derivatives were 2.85 min for codeine; 3.02 min for morphine; 3.10 min for hydrocodone-oxime; 3.17 min for hydromorphone-oxime; and 3.32 min for oxycodone-oxime. The MSD was operated in the SIM mode using the following ions (quantitative ions are in parentheses): 234, 343, (371) for codeine; 346 (374) for codeine-d3; 234, 401, (429) for morphine; 417, (432) for morphine-d3; 297, 371, (386) for hydrocodone-oxime-tms; 368, 444 (459) for oxycodone-oxime-tms; and 429, 444, (355) for hydromorphoneoxime-tms. Methods Stock solutions. Standard 1.0 mg/ml solutions in methanol were prepared for all solid compounds with the exception of M3G and M6G, which were prepared in DI H20. A 0.01 mg/ml 6MAM standard was prepared by serial dilution with methanol of the 1.0 mg/ml stock. A 0.01 mg/ml 6MAM-d~ solution was prepared by dilution of the stock standard. Separate stock solutions were prepared for preparation of calibrators and controls. Final mixed internal standard solution was prepared by transferring 100 IJL of each of the 1.0 mg/ml morphine-d3 and the 0.01 mg/ml 6MAM-d~ into a ]2-ram x 75-ram glass tube. The methanol was evaporated at 40~ under N2. A 1.0 ml solution of 95:5 DI H20/ACN was added to the tube, which was then vortex mixed and capped. Final IS concentrations are 100 tjg/ml 706 morphine-d3 and 1 IJg/mL 6MAM-d6. Urine calibrators~controls. A 10,000/50 (opiates/6mam) ng/ml mixed calibration urine standard was prepared in a single 20.0-mL volumetric flask using the appropriate dilutions of the stock standards with drug-free pooled urine as the diluent. Six additional calibration curve standards were prepared by serial dilution of 10,000/50 ng/ml mixed calibration standard. An additional 10,000/50 ng/ml mixed urine control solution was prepared separately from the primary reference stock solutions as described for the calibrators. Additional urine controls at 1000/5.0 ng/ml and 100/0.5 ng/ml were prepared. A 10,000 ng/ml mixed standard was prepared as described except that 95:5 DI H~O/ACN was used as the diluent. Final urine calibrators were prepared by transferring 1.0 ml of the appropriate calibration standard to a 1.5-mL autosampler vial. Ten microliters of IS solution was then added, and the vial was capped and vortex mixed. Samples were then placed in an Eppendorf refrigerated centrifuge (15~ and spun for 10 rain U1 BLANK 1 4: M RM of 4 Channels ES+ 334 9 5 ~ 1(01 5. 40. 0 5.545'84~ 7 ~ ^, "u' 164.9 1.72e3 0 ~.... i.... i.... i.... i.... L.... i.... L.... i.... i................ i.... [.... i % 5.0 54-5.44 5"85"592 328> 164.9.... i.... i....,.... i.... i.... J.... ~.... F.... i.... i.... i.... i.... i.... J.... 5.3 ~ 5.67~ 5:895.99 316 9 240.9 10 % 5.01~,~'-~f?" 4.98e4.... i.... i.... i.. ~ w ~ - ~.. i.... i.... b.... i.... =.... i.... i.... i.... i.... i.... i 4; MRM of 4 Channels ES+ 1003 5nl 54&5.85-,535 300 9 198.8 o,fl v~vr 2.71e5 o,............................ 100~ 4 174A~ 4.654,98 300>198.8 1 0 0 ~ 4 1 6 4 4 8 4 6 7 4 9 9 286.1 9 1 8 4 5 off o....... 471 4.96 286.1 9 225.... r.... i.... i....,.... j....,.... i.... i.... J.... i.... i.... i.... i 2: MRM of 3 Channels ES+ 1007, 3.77,,&86 _4.03 302 9 226.9 %i ~1~ 2.99e4 0.... i.... i.... i, i.... i.... i.... i.... i.... ~.... i.... ~,,, J.... i.... i.... i 2: MRM of 3 Channels ES+ 100~ 3.71 3.94 289 9 271.2 :1 oo9 ~,. i.... J.... i.... i.... i.... i.... J.... J.... J.... i.... i.... i.... i.... i.... i 2: MFIM of 3 Channels ES+ 1001~ 366~, ~P/~1"3"86~3'96 285.99 ~ 1.... i.... i....,.... t.... i.... J....,.... ~.... ~.... i.... ~.... r.... =.... i.... 1: MRM of 1 Channel ES+ 10 1.28 462 9 286 %0~,04 0,s ~.~ 6.73e4 0~... i 2 ~ 1.00 2.00 3.00 4,00 6.00 7.00 Time (min) Figure 2. Evaluation of ion suppression, Ion channels as described in Figure 1.

"L Journal of Analytical Toxicology, Vol. 29, October 2005 U BLANK MIP110903D91 4: MRM of 4 Channels E$4-510530 334>1649 1~176 "-. "L5.6% ~ip5 1~8 MtPl 10g03D0"l 1001 4.94 6,125.27 5.89 328 > 164.9.... i.... i.... i.... i.... i... ~.... i.... i.... i.... i.... i.... ~.... J.... i.... i 100--~ 4.96 5.25.33 316 9 240.9 ~176..................... i i i.. b, i i... i... ~ i i. i.... i.... i.... i............ ~.... 4: MRM of 4 Channels lee+ 10 ~ 5.63 2,03e3 0....,....,....,....,....,....,....,....,.... ~... : 7 ~ ' ; ~ '. ~,.. " ~ : ' ~....,....,...., 3: MRM of 3 Channels ES 10~ ~ 300 9 198.8 4.2,87 1.27e3... i... ~... i... f... i... i... i... J... i... i... i... i...,... i... i MIP11 0903D01 ~0~ 444 4.53 286.1 > 184,5 1 4 ;6 ~4.55 1.68~ ~ ' "~.~l ~.~ ~4.85 0....~,",, ~... 1~ 4-534.70 286.1 > 225 MIP1109031301 4. 1 4. 4 5 ~. 8 6 344 2: MRM of 4 Chennels ES+ 4.06 462 > 286 10%0~ 3. 453'77 ' 282 3.3 LLJ MIP 110903D01 2: M RM Of 4 Channels ES+ 10 P~ 3.28 4.11 289 > 168 ~:,~ 8231 3.3"t 14.~4,340 MIP11GgO3D01 3.74 288,9 9 200.9 0... ~... ~... ~...,... ~... ~,,"]K"ET~ v. ",~ 7. 9.,... ~...,...,...,...,..., 1003,~0.06 0 96. o " " " '.15 1 : MRM of 1 Channel ES+ 2 42 2 81 462 9 286 119 1.O0 2.00 3.00 4.00 6.00 7.00 Time (min) Figure 3. Direct injection of drug-free urine. Ion channels as described in Figure 1. Table II. Method Validation Parameters* Precision (%CV) Accuracy (%Bias) Opiate r Inter lntra Inter Intra M3G 0.9995 3.1-6.7 3.4-7.6-2.4-8.0-0.3-3.6 M6G 0.9991 3.5-7.8 1.3-5.5-6.6-10.1-4.3-6.0 Morphine 0.9974 2.3-6.9 4.6-5.8-2.1-2.2-3.6-2.2 Oxyrnorphone 0.9991 2.6-11.9 2.1-5.3-7.2-6.8-3.3-11.8 Hydromorphone 0.9989 2.9-6.1 4.4-8.7-4.1-6.2-5.2-2,4 Norcodeine 0.9993 4.1-16.7 2.5-10.9-4.6-11.t -4.4-2.3 Codeine 0.9939 1.1-8.4 1.6-3,4-5.7-5.9-8.5-3.5 Oxycodone 0.9992 3.4-8.0 2.9-5.3-4.9-6.3 0.2-6.0 6MAM 0.9986 3.0-13.1 5.6-16.3-7.2-13.6 2.0-8.3 Hydrocodone 0.9990 2.7-8.0 1.5-3.6-7.3-7.9-9.6-0.9 * n = 6 for all determinations. at 4500 x g. Urine controls were processed in the same manner. Tuning solutions. Tuning solutions for the LC-MS-MS were made for each compound by preparing serial dilutions of the 1.0 mg/ml stocks to 1 pg/ml in 50:50 acetonitrile and water. A 500-1JL Hamilton gas tight syringe filled with tuning solution was installed in a Harvard Infusion Pump connected into the mobile phase stream via a tee connector. The infusion rate was set at 10 IJL/min into the mobile phase flow at 0.25 ml/min. Final tuning parameters for MRM acquisition are listed in Table I. Ion suppression~matrix effect. An ion suppression experiment was conducted to determine the presence of a matrix effect that would adversely affect the results. In this experiment the infusion pump is connected post-column via a tee connector. The 10,000 ng/ml mixed standard in solvent was diluted 1:100 with 95:5 DI H20/ACN to yield a 100 ng/ml solution. This solution was loaded in the gas-tight syringe, and the infusion rate was set at 10 IJL/min. The baseline was monitored until it became stable. Injections of drug-free urine were then made as normal acquisitions. LC--MS-MS sample analysis. An aliquot of urine was centrifuged in the Eppendorf refrigerated centrifuge at 4500 x g for 10 min at 15~ to pellet large particles; 1.0 ml of sample was then transferred with a micropipette to a 1.5-mL glass autosampler vial containing 10 ljl of IS solution. The vial was capped and vortex mixed. All specimens that exceeded the upper limit of linearity were diluted with drug-free human urine and reanalyzed. GC-MS sample analysis. To 2.0 ml urine calibrators, controls, and diluted patient sample (500 I~L of urine was added to 1.5 ml of distilled water), 20 l~l (33 ng/ljl) of tri-deuterated internal standards of codeine and morphine was added, followed by 75 IJL of [3-glucuronidase and 300 IJL of 2M acetate buffer/10% hydroxylamine mixture. The mixture was vortex mixed and placed in a water bath at 55~ for 2 h to hydrolyze opiate glucuronide conjugates and convert the ketone group on the phenanthrene ring to an oxime to prevent tautomeric shifts. The mixture was removed from the water bath and diluted with 5.0 ml of ph 9.1 buffer (Biochemical Diagnostics Corp., Edgewood, 707

NY). Opiates were isolated from urine by solid-phase extraction in a Detectabuse TM Gravity GV-65 column as described by the manufacturer (Biochemical Diagnostics Corp.). The final column wash (acetonitrile/n-butylchloride, 30:70) containing extracted opiates was evaporated to dryness under nitrogen with low heat. The residue was dissolved in 50 pl of ethyl acetate, and the opiates were derivatized by the addition of 50 pl ofn, O-bis(trimethylsilyl)trifluoroacetamide (BSTFA). The mixture was transferred to autosampler vials that were heated in a bead bath at 75~ (+ 20%) for 15 min. One to 3 pl of the mixture was injected into the GC-MS. The GC-MS method was previously validated at Ameritox. The linear range was 80 ng/ml to 20,000 ng/ml with the lower limit of quantitation for each opiate 80 ng/ml. Oxymorphone and 6MAM were not analyzed by this method. Results LC-MS-MS method development and validation Initial development of the method used isocratic conditions in order to decrease retention time and take advantage of the capabilities of the LC-MS-MS to spectrally separate the opiates. Product ions were chosen that provided characteristic fragments of the precursor ion and provided maximal response. Morphine and hydromorphone, however, share common precursors and products ions, and isocratic conditions did not provide sufficient chromatographic separation of morphine and hydromorphone. Therefore, a gradient was instituted that started with 5% organic solvent, held briefly, and then increased to approximately 20% over a short period of time. The final separation is shown in Figure 1. In this system, morphine and oxymorphone are not totally separated in time, but spectrally they were distinguishable. Oxymorphone's parent mass-tocharge ratio is 302.00 versus 285.90 for morphine, and they do not share common product ions. Hydromorphone was well separated from morphine and other components. Single product ion transitions were considered acceptable for forensic identification purposes as the urine specimens used in the cross-validation were obtained from a pain management clinic and had been submitted specifically for opiate analysis. In the event of the analysis of an unknown specimen, an additional transition could be added to provide additional confirmation. Selectivity of the method was determined in two ways: injection of multiple drug-free urines specimens obtained from different donors and determination of matrix effects on ion suppression. Ten different urine specimen pools were tested. In the infusion experiment, some ion suppression was noted although not all the compounds may potentially be affected. Figure 2 demonstrates a typical ion suppression pattern for a blank urine sample. A drop in the baseline indicates potential ion suppression. Based upon the retention times of all compounds, hydromorphone, norcodeine, codeine, and oxycodone appear to have the greatest potential for ion suppression issues. Figure 3 depicts the same drug-free urine presented in Figure 2 injected using the MRM conditions. No peaks were observed that would interfere with the analysis of any of the compounds. Name: MIP110903D02 Textr U $1 50/0.25 i: Mor 3-B-glucur..2: Morphine - d3 IS3: Mor 6-B-glucur.4: Morphine 5: 0xymorphone 6: Hydromorphone MIPl10903D02 F11 MIP110903D02 F2 MIP110903D02 F2 MIP110903D02 F; MIP110903D02 F2 I MIP110903D02 F3 100~ 462 > 286 100 462 > 286 1007 28[ ) > 200!. 100-1 3C > 226.91 10@ 286.1 > 184.5 q 1.53e:t 1i ',891.48e[ > 161 1.7003 5.06e,~ ~[ 6.68e3 2.79e4 Are~ / Are~ Are~ %H 0,~,,,~;,;,~ Time 2.00 0 ],,,,,,,,.,,,,,,~,, Time 3.50 4.00 %-t 0...,'.-.,... "rime 3.50 4.00,4 3.50 4,00 %-1 T,me ~ ~ Time 3.50 4.00 I %. 0,.,.,.~..,,,,,,, Time 4.50 7: Norcodeine 8: Codeine 9: Hydrocodone i0: 6-MAM ii: 6-MAM-d6 12: Oxycodone MIP110903D02 F3 MIP110903D02 F3 MIPl10903D02 F4I MIP110903D02 F4 MIP110903D02 F4 I MIP110903D02 F4.86.1 > 22. 100 100 > 198.8 i 100 328 > 164.9 100~ 334> 164.9 I 100-316 > 240.s 100] 1.27e: 1o, l 3c,1981 1.64e 3.88e41 232.~ 5.81e3 I 7.21 e3 Are; Are~ l %4 %-1 %- 0 'nme 4.50 04~d ~- 9,- Time 4.50 0,,~,,,,,,Lr~-., Time 0% Time 0 ',... ~... Time 0- ~ Time Figure 4. Direct injection of LOQ urine calibrator containing 0.25 ng/ml 6MAM and 50 ng/ml all other opiates. 708

None of the 10 drug-free urine pools tested demonstrated peaks that interfered with analysis of the opiate analytes. All the calibrations of the opiate analytes demonstrated a correlation coefficient (r) greater than 0.99 (Table II). A typical chromatogram of the lowest opiate calibrator is shown in Figure 4. Interassay accuracy and precision were determined from the back-calculated calibrators (Table II). The method demonstrates acceptable accuracy and precision. The greatest bias, 13.60%, was found with the 0.25 ng/ml 6MAM calibrator, and the greatest CV was the norcodeine 50 ng/ml calibrator. All others demonstrated good accuracy and precision. Intrarun accuracy and precision was demonstrated by six replicates at three concentrations, 100 ng/ml, 1000 ng/ml, and 10,000 ng/ml (Table II). The greatest bias was -9.59% with the hydrocodone 1000 ng/ml control. 6MAM at 0.50 ng/ml demonstrated the greatest imprecision at 16.93% CV. The lower limits of quantitation (LOQ) were arbitrarily established at the lowest calibrator for each opiate analyte. At these concentrations, the accepted requirements for LOQ were met: signal-to-noise ratio greater than 10 and precision and accuracy less than 20%. Cross-validation with GC-MS Eighty-nine urine specimens previously analyzed by GC-MS were re-analyzed by LC-MS-MS. The LC-MS-MS results yielded an 88% agreement with GC-MS results for opiate negative samples and 94% agreement for opiate positive samples. The LC-MS-MS method identified additional opiates in 19 specimens opiate positive by GC-MS. Oxymorphone was identified by LC-MS-MS in 10 specimens that the GC-MS method was not set up to detect. However, 9 of these 10 specimens were positive for oxycodone by GC-MS. Because oxymorphone is a metabolite of oxycodone, the presence of the parent opiate was not missed. Similarly hydromorphone was detected by LC-MS-MS in three specimens that were positive for hydrocodone by GC-MS, which can also be explained by hydrocodone metabolism. The remaining LC-MS-MS opiate-positive specimens not detected by GC-MS contained oxymorphone, oxycodone, hydrocodone, and hydromorphone. Quantitatively, it was difficult to compare the two methods because the GC-MS method utilized a glucuronidase step that was not included in the LC-MS-MS method. Therefore, only the morphine results could be compared if the LC-MS-MS results for the glucuronide metabolites were summed on a molar basis with the morphine. A scatter-plot of GC-MS results versus LC-MS-MS results yielded a coefficient of determination (r ~) equal to 0.2975. Assuming 100% hydrolysis, the M3G and M6G (on a molar basis) were added to the morphine value and the scatter-plot was redrawn. The r 2 increased to 0.8127. Discussion With the exception of 6MAM, the method was linear for all opiate analytes from 50 ng/ml to 10,000 ng/ml. 6MAM was linear from 0.25 ng/ml to 50 ng/ml. The r was > 0.99 for all analytes. Accuracy and precision at all concentrations were within 20% for bias and a CV of 20% is a normal accep- tance criterion for bioanalytical methods. LOQ for all opiates but 6MAM was 50 ng/ml; LOQ for 6MAM was 0.25 ng/ml. Dams et al. (12) reported a direct injection LC-MS--MS method using urine as the matrix and several drugs, including opiates were analyzed. Sample preparation consisted of adding internal standard, mixing, centrifuging, and transferring to the LC-MS for analysis. APCI was the interface chosen because of the decreased ion suppression caused by the urine matrix. ESI typically has greater ion suppression in a urine matrix because several endogenous compounds found in urine are polar by nature of the excretion process. The competition of these polar compounds for charge created in the ESI probe is the suspected mechanism of the ion suppression. ESI, however, provides a greater response to ionized compounds then does APCI. Therefore, ESI was chosen for this method to try to provide sufficient sensitivity with direct injection for 6MAM below the 10 ng/ml reported by Dams and colleagues (12). To try to circumvent the ion suppression problem, a high aqueous initial mobile phase coupled with a gradient was used. The high aqueous environment would permit the opiates to be retained on the HPLC column, whereas some of the more polar compounds would be passed directly through the column. After an initial 's~vashing" period, the gradient is started to elute the opiates from the column. Of course this methodology is not selective, but it was hoped that it would simply decrease the ion suppression sufficiently to allow the analysis to occur. Considering that LC-MS-MS methods are routinely reported in the picograms-per-milliliter range, 10 ng/ml is a significant concentration to use. The gradient would further permit the separation of the critical analyte pair of morphine and hydromorphone. The results of the validation demonstrate that the method is usable for the determination of opiates by direct injection of urine into the LC-MS-MS. Stability of the method was not evaluated, however, for the following reasons. It is well documented in the literature that 6MAM is not stable for long periods of time when stored in urine. Rop et al. (13) found that 6MAM stored at 20~ lost 10% within 4 h, and within 24 h, 65% of the original amount was lost. Increasing the pi-i of the sample accelerated the loss of 6MAM. If the samples were stored at -20~ loss was limited to 9%. Additionally, when the samples were treated with ~-glucuronidase to hydrolyze the glucuronide conjugates of morphine, 100% of the 6MAM was lost within 8 h. The loss of morphine under the same conditions (-20~ and hydrolysis) was 0%. Stored methanolic solutions of both morphine and 6MAM were stable for at least 14 days in that study. Stability of methanolic solutions must be longer because Cerilliant, the manufacturer of the standards used in this study, includes one-year expiration dates on their products. Given the ease with which the calibration curves can be prepared, it seems more practical to make the calibration curves for each analysis on an as needed basis. Although this may potentially cause a decrease in the interrun performance of the method, that was not a factor in this validation. To further validate this method, a cross-validation was performed, by analyzing samples previously reported by a GC-MS method. Initially, a poor correlation was found between the two methods. Removing the compounds for which no glucuronide 709

standards were available improved the r 2. Despite the correction used by summing morphine and its glucuronide metabolites, some samples still were vastly different between GC-MS and LC-MS-MS. For example, one sample by GC-MS was morphine positive with 18,350 ng/ml and LC-MS-MS positive with 37,685 ng/ml. This large difference cannot be explained, but it appears to be reproducible. Only one other direct injection method for urine using LC-MS-MS has been reported (12). APCI interface was used, however, although similar LOQs and upper limits of linearity were reported. The LOQ for 6MAM at 0.25 ng/ml for our method is far superior to the 10 ng/ml reported by Dams et al. (12). In that report, the glucuronide metabolites of morphine, hydromorphone, hydrocodone, oxycodone, norcodeine, and oxymorphone were not included in the method. Normorphine, heroin, and acetycodeine were included, and our method does not analyze for those opioids. Another advantage of this new method is shorter run times. In this method, all analyses are complete within 6.00 rain. In the Dams et al. (12) report, the last opioid is eluted at 19.3 rain. Conclusions A method for the analysis to 10 opiates by direct injection of urine and detection by LC-MS-MS has been developed. The method is simple and provides a rapid analysis of the 10 opiate analytes included. The LOQ, linear range, precision, and accuracy validated for this method would permit its use for a variety of bioanalytical applications. The LOQ for 6MAM, 0.25 ng/ml, presents an improvement in previously reported limits for this analyte. References 1. R. Pacifici, S. Pichini, I. Altieri, A. Caronna, A.R. Passa, and P. Zuccaro. High-performance liquid chromatographic-electrospray mass spectrometric determination of morphine and its 3- and 6-glucuronides:application to pharmacokinetic studies. J. Chromatogr. B 664:329-334 (1995). 2. N. Tyrefors, B. Hyllbrant, L. Ekman, M. Johansson, and B. Langstrom. Determination of morphine, morphine-3-glucuronide and morphine-6-glucuronide in human serum by solidphase extraction and liquid chromatography-mass spectrometry wih electrospray ionisation. J. Chromatogr. 729:279-285 (1996). 3. W. Naidong, J.W. Lee, X. Jiang, M. Wehling, J.D. Hulse, and P.P. Lin. Simultaneous assay of morphine, morphine-3-glucuronide and morphine-6-glucuronide in human plasma using normal-phase liquid chromatography-tandem mass spectrometry with a silica column and an aqueous organic mobile phase. J. Chromatogr. B 735:255-269 (1999). 4. G. Schanzle, S. Li, G. Mikus, and U. Hofmann. Rapid, highly sensitive method for the determination of morphine and its metabolites in body fluids by liquid chromatography-mass spectrometry. J. Chromatogr. B 721 : 55-65 (1999). 5. W.Z. Shou, M. Pelzer, T. Addison, X. Jiang, and W. Naidong. An automated 96-well solid phase extraction and liquid chromatography-tandem mass spectrometry method for the analysis of morphine, morphine-3-glucuronide and morphine-6-glucuronide in human plasma. J. Pharm. Biomed. Anal. 27:143-152 (2002). 6. M. Dawson, F. Bronwyn, T. Kelly, J. Keegan, and L.E. Mather. A rapid and sensitive high-performance liquid chromatographyelectrospray ionization-triple quadrupole mass spectrometry method for the quantitation of oxycodone in human plasma. ]. Chromatogr. 5ci. 40:40-44 (2002). 7. Y.L. Chen, G.D. Hanson, X. Jiang, and W. Naidong. Simultaneous determination of hydrocodone and hydromorphone in human plasma by liquid chromatography with tandem mass spectrometric detection. J. Chromatogr. B 769(1): 55-64 (2002). 8. M.J. Bogusz, R.D. Maier, and S. Driessen. Morphine, morphine- 3-glucuronide, morphine-6-glucuronide, and 6-monoacetylmorphine determined by means of atmospheric pressure chemical ionization-mass spectrometry-liquid chromatography in body fluids of heroin victims. J. Anal. Toxicol. 21:346-355 (1997). 9. M.J. Bogusz, R.D. Maier, M. Erkens, and S. Driessen. Determination of morphine and its 3- and 6-glucuronides, codeine, codeine-glucuronide and 6-monoacetylmorphine in body fluids by liquid chromatography atmospheric pressure chemical ionization mass spectrometry. J. Chromatogr. B 703:115-127 (1997). 10. R.L. Fitzgerald, J.D. Rivera, and D.A. Herold. Broad spectrum drug identification directly from urine, using liquid chromatography-tandem mass spectrometry. Clin. Chem. 45" 1224-1234 (1999). 11. R. Dams, C.M. Murphy, R.E. Choo, W.E. Lambert, A.P. De Leenheer, and M.A. Huestis. LC-atmospheric pressure chemical ionization-ms/ms analysis of multiple illicit drugs, methadone, and their metabolites in oral fluids following protein precipitation. Anal. Chem. 75:798-804 (2003). 12. R. Dams, C.M. Murphy, W.E. Lambert, and M.A. Huestis. Urine drug testing for opioids, cocaine, and metabolites by direct injection liquid chromatography/tandem mass spectrometry. Rapid Commun. Mass Spectrom. 17:1665-1670 (2003). 13. P.P. Rop, F. Grimaldi, J. Burle, M.N. De Saint Leger, and A. Viala. Determination of 6-monoacetylmorphine and morphine in plasma, whole blood and urine using high-performance liquid chromatography with electrochemical detection. J. Chromatogr. B 661 : 245-256 (1994). 710