Validation of Analysis of Amphetamines, Opiates, Phencyclidine, Cocaine, and Benzoylecgonine in Oral Fluids by Liquid Chromatography Tandem Mass Spectrometry Subbarao V. Kala*, Steve E. Harris, Tom D. Freijo, and Stan Gerlich One Source Toxicology Laboratory, 1213 Genoa Red Bluff, Pasadena, Texas 77504 Abstract The aim of the present study was to develop and validate a method for the detection and quantitation of drugs of abuse in oral fluids. Fortified oral fluid samples (made in-house) and samples from donors collected with Quantasil oral fluid collection kits from Immunalysis were screened on an Olympus 5400 using reagents purchased from Immunalysis. Amphetamines (AMPs), opiates, phencyclidine (PCP), and cocaine and its metabolite benzoylecgonine (BE) in oral fluids were quantitated by an Applied Biosystems 3200 QTRAP liquid chromatograph tandem mass spectrometer (LC MS MS). AMPs, opiates, PCP, cocaine, and BE were extracted from samples using liquid liquid or solid-phase extractions and the extracts were separated on a Shimadzu highperformance liquid chromatograph prior to the MS MS analysis. For each drug, two multiple reaction mode transitions were monitored using positive electrospray ionization coupled to an MS MS detector. Corresponding d 3, d 5, d 6, and d 11 internal standards were used to quantitate the results. The limit of detection/quantitation for AMPs, opiates, PCP, cocaine, and its metabolite BE were 10, 10, 2, 2, and 2 ng/ml of oral fluid, respectively, on a signal-to-noise ratio > 4. This corresponded to 25, 25, 5, 5, and 5 pg on column. The method was verified by participating in the North America Oral Fluid Proficiency Testing administered by Research Triangle Institute and by analyzing real samples from donors. In conclusion, LC MS MS provided a simple way to analyze and quantitate drugs of abuse in oral fluids. Introduction In recent years, oral fluid testing for drugs of abuse has become more prominent and is used in a wide variety of situations such as pre-employment, school athletics, roadside testing, prisons, and other institutions that are monitored by drug courts (1 4). Although urine is still the specimen of choice for testing of drug abuse in workplaces, oral fluid testing has become more important to detect the recent use of drugs because the drugs in oral fluids typically last between 12 to 24 h (5). A liquid chromatography tandem mass spectrometry (LC MS MS) method was developed for the confirmation and quantitation of the following drugs of abuse: amphetamine (AMP), methamphetamine (MAMP), 3,4-methylenedioxyamphetamine (MDA), 3,4-methylenedioxymethamphetamine (MDMA), codeine, morphine, hydrocodone, hydromorphone, oxycodone, oxymorphone, phencyclidine (PCP), and cocaine and its metabolite benzoylecognine (BE) in oral fluids. The method was verified by testing the samples obtained from real subjects as well as by participating in the North American Oral Fluid Proficiency Testing administered by Research Triangle Institute (RTI). Oral fluid drug testing in contrast to urine is advantageous, as collection is non-invasive, and because samples are collected under direct observation, issues regarding adulteration are of little concern. The advantages and disadvantages of oral fluid testing are described in great detail by other investigators (6 9). Several investigators have initiated the testing of oral fluid samples by LC MS MS as opposed to gas chromatography MS, as the cutoffs for positive identification of drugs in oral fluids are much lower (at least 10 times) than urine. Previous studies from other investigators have shown that LC MS MS is highly sensitive for the identification of drugs in urine, oral fluids, and hair samples (10 12). Additionally, both parent and metabolite compounds can be detected with greater sensitivity. This report describes the details regarding the validation of an LC MS MS method for the detection and quantitation of different classes of drugs of abuse in oral fluids. Experimental Materials and Methods * Author to whom correspondence should be addressed. Subbarao V. Kala, Ph.D., Scientific Director, One Source Toxicology Laboratory, 1213 Genoa Red Bluff, Pasadena, TX 77504. E-mail: skala@onesourcetox.com. Chemicals and reagents All drugs of abuse (AMP, MAMP, MDA, MDMA, AMP-d 11, Reproduction (photocopying) of editorial content of this journal is prohibited without publisher s permission. 605
MAMP-d 11, MDA-d 5, MDMA-d 5, codeine, morphine, hydrocodone, hydromorphone, oxycodone, oxymorphone, codeine-d 3, morphine-d 3, hydrocodone-d 3, hydromorphoned 3, oxycodone-d 6, cocaine, BE, cocaine-d 3, BE-d 3, PCP, and PCP-d 5 ) were purchased from Cerilliant (Round Rock, TX). High-performance liquid chromatography (HPLC)-grade ammonium acetate, acetonitrile, formic acid, methanol, and other reagents were purchased from Fisher Scientific (Houston, TX). Calibrators and quality controls All calibrators and controls were prepared in diluted drugfree saliva in an Immunalysis collection buffer (Pomona, CA). For all drugs, a calibrator, a low control (40% of cutoff), a high control (25% above cutoff), and a negative control were used for the analysis of oral fluid samples from donors. For the determination of limits of detection (LOD) and quantitation (LOQ), diluted negative (1:3 in Immunalysis buffer) oral fluid was fortified with different concentrations of drug, and the samples were analyzed in triplicate by LC MS MS after extraction (either liquid liquid or solid-phase). Deuterated internal standards were used to quantitate the results. Several criteria, including the shape of the chromatographic peak, the retention time (within ± 2%), and qualifying ion ratios (within ± 20%) were used to monitor the quality control of the analysis of the samples. 606 Extraction procedures The Immunalysis oral fluid collection device (Quantisal) allows the collection of 1 ml of oral fluid, and this is collected into 3 ml of buffer solution. During this process, the drug in the oral fluid is diluted 4 times. All drugs were extracted using 0.5 ml of diluted oral fluid (4 times) sample. For AMP and PCP extractions, deuterated internal standard (25 µl of 250 ng/ml of amphetamine mix or 25 µl of 500 ng/ml PCP), 2.0 ml of 100 mm phosphate buffer (ph 9.3), and 2.0 ml of hexane/ethyl acetate (50:50) were added to controls and samples. Samples were vortex mixed and placed on a mechanical rotator for 4 min to allow the extraction of the drugs. The extraction batch was centrifuged at 4000 rpm for 4 min and decanted into 10 75-mm test tubes. Mobile phase (125 µl of 5 mm ammonium acetate for AMPs or 0.1% formic acid for PCP) was added to each test tube and evaporated at 45 C under nitrogen stream until only the aqueous phase remained. The samples were then transferred to vials for the analysis by LC MS MS. For opiates, cocaine, and BE extractions, deuterated internal standard (25 µl of 200 ng/ml opiate mix or 25 µl of 400 ng/ml cocaine or BE) and 1.0 ml of 100 mm phosphate buffer (ph 6.0) were added to controls and samples. UCT Zsdau005 columns were prepped with 500 µl methanol, followed by 500 µl of DI water, and finally with 500 µl of 100 mm phosphate buffer (ph 6.0). The samples and controls were loaded onto the columns. The columns were then washed with 500 µl of DI water, followed by 500 µl of 100 mm acetic acid (for opiates) or 0.1 N HCl (for cocaine and BE), and finally with 500 µl of methanol. The columns were dried under vacuum and the drugs were eluted using 800 µl of MeCl/IPA/NH 4 OH (78:20:2). The samples were evaporated completely at 45 C under a nitrogen stream and 100 µl of mobile phase were added (0.1% formic acid and water) and transferred to the vials for the analysis by LC MS MS. The final concentrations of internal standards for AMPs, opiates, PCP, cocaine, and BE in diluted oral fluid were 12.5, 10, 25, 20, and 20 ng/ml, respectively, which corresponded to 50, 40, 100, 80, and 80 ng/ml of undiluted oral fluid. LOD/LOQ, precision, and accuracy studies For the precision and accuracy studies, 10 samples of low controls (40% of cutoff; AMPs at 20 ng/ml, opiates at 16 ng/ml, PCP at 4 ng/ml, cocaine at 8 ng/ml, and BE at 8 ng/ml) for each drug were analyzed by LC MS MS and the coefficient of variation and percentage of accuracy were calculated. LOD/LOQ were evaluated in triplicate. LOD is defined as the concentration producing a peak eluting within ± 0.05 min of the analyte s retention time for the lowest concentration of the drug with a signal-to-noise ratio of at least 3:1 and the qualifier ion ratios within ± 20% of the calibration standard. For LOQ, the quantitative results also should meet the criteria of being within 20% of target value in addition to satisfying all the previously mentioned criteria of LOD. Screening Samples from donors and the proficiency testing program were screened prior to confirmation by LC MS MS. Screening of oral fluids was performed using an Olympus AU5400 with reagents and controls from Immunalysis. The screening cutoffs for AMPs, opiates, PCP, cocaine, and BE were 50, 40, 10, 20, and 20 ng/ml, respectively. HPLC All HPLC work was carried out using Shimadzu HPLC system equipped with dual pumps, autosampler, and a column heater. The separation of AMPs was carried out using a Pinnacle II C18 column (50 4.6 mm, 5 µm, Restek, Austin, TX), whereas opiates, PCP, cocaine, and BE were separated on an Allure PFP propyl column (50 2.1 mm, 5 µm). (For the details of the mobile phases, please see the captions of Figures 1 4.) MS MS analysis The detection and quantitation of the drugs was carried out using an Applied Biosystems 3200 QTRAP system. The turbo spray conditions were identical for analysis of all the drugs (curtain gas: 45 psi; collision gas: 8 psi; ion spray voltage: 5500 V; temperature: 65 C; ion source gas: 65 psi; ion source gas2: 70 psi; dwell time: 100 ms), and two multiple reaction monitoring (MRM) transitions were monitored for each drug. The details of MRM transitions, retention times, and collision energy used for each drug are given in Table I. The cutoff concentrations for AMPs, opiates, PCP, cocaine, and BE were set at 50, 40, 10, 20, and 20 ng/ml oral fluid according to the suggestions from RTI. Data analysis was carried out using Applied Biosystems Analyst Software (version 1.4.2). Ion-suppressive effects of oral fluid matrix were studied by analyzing water and oral fluid samples spiked with different drugs by LC MS MS. No ion suppressive effects of oral fluid matrix were found on any of the analytes studied.
Results and Discussion Linearity studies were initially conducted to verify the range at which AMPs (AMP, MAMP, MDA, MDMA), opiates (codeine, morphine, hydrocodone, hydromorphone, oxycodone, oxymorphone), PCP, cocaine, and BE can be detected using LC MS MS. Saliva or oral fluid is generally collected into a collection device containing a buffer solution. Immunalysis collection devices Figure 1. Ion chromatograms and linear regression curves of AMP, MAMP, MDA, and MDMA. Extracted ion chromatogram of 10 ng/ml amphetamines mixture in oral fluids generated by LC MS MS (A). Note that the signal-to-noise ratio for each of the compounds is > 4. The deuterated internal standard concentrations were 50 ng/ml. Linear regression curves were generated by plotting the data obtained from LC MS MS analysis of negative oral fluids spiked with different concentrations of AMPs (10 to 2000 ng/ml) (B). The linear regression coefficient (r) for AMPs in oral fluids was found to be > 0.993 over the range of 10 to 2000 ng/ml. AMPs were separated on a Pinnacle II C18 column, using gradient of 5 mm ammonium acetate and methanol containing 0.1% formic acid. The flow rate was set at 1.2 ml/min and the run time was 5 min. Figure 2. Ion chromatograms and linear regression curves of opiates (codeine, morphine, hydrocodone, hydromorphone, oxycodone, and oxymorphone). Extracted ion chromatogram of 10 ng/ml opiate mixture in oral fluids generated by LC MS MS (A). Note that the signal-to-noise ratio for each of the compounds is > 4. The deuterated internal standard concentrations were 40 ng/ml. Linear regression curves were generated by plotting the data obtained from LC MS MS analysis of negative oral fluids spiked with different concentrations of opiates (1 to 500 ng/ml) (B). The linear regression coefficient (r) of opiates in oral fluids was found to be > 0.998 over the range of 10 to 500 ng/ml. Opiates were separated on an Allure PFP propyl column using gradient of 0.1% formic acid in water and 80:20 methanol/acetonitrile containing 0.1% formic acid. The flow rate was set at 1.0 ml/min and the run time was 6 min. 607
were used for the validation of this method as well as to analyze the oral fluid samples from real donors. The Immunalysis oral fluid collection device (Quantisal) allows the collection of 1 ml of oral fluid and this is collected into 3 ml of buffer solution. During this process, the drug in the oral fluid is diluted 4 times. A In order to maintain the same dilution, diluted oral fluid samples were prepared and fortified with concentrations that would give a final concentration of the drug in oral fluid at near and around cutoff levels. The results are expressed as concentration of the drug in the original (undiluted) oral fluid. B Figure 3. Ion chromatograms and linear regression curves of cocaine and BE. Extracted ion chromatogram of 2 ng/ml cocaine and BE mixture in oral fluids generated by LC MS MS (A). Note that the signal-to-noise ratio for each of the compounds is > 4. The deuterated internal standard concentrations were 80 ng/ml. Linear regression curves were generated by plotting the data obtained from LC MS MS analysis of negative oral fluids spiked with different concentrations of cocaine and BE (1 to 1000 ng/ml) (B). The linear regression coefficient (r) of cocaine and BE in oral fluids was found to be > 0.998 over the range of 1 to 1000 ng/ml. Cocaine and BE were separated on an Allure PFP propyl column using a gradient of 0.1% formic acid in water and 80:20 methanol/acetonitrile containing 0.1% formic acid. The flow rate was set at 1.0 ml/min, and the run time was 6 min. Figure 4. Ion chromatograms and linear regression curves of PCP. Extracted ion chromatogram of 2 ng/ml PCP in oral fluids generated by LC MS MS (A). Note the signal-to-noise ratio for each of the compounds is > 4. The deuterated internal standard concentration was 100 ng/ml. Linear regression curves were generated by plotting the data obtained from LC MS MS analysis of negative oral fluids spiked with different concentrations of PCP (1 to 500 ng/ml) (B). The linear regression coefficient (r) of PCP in oral fluids was found to be > 0.998 over the range of 1 to 500 ng/ml. PCP was separated on an Allure PFP propyl column using gradient of 0.1% formic acid in water and 80:20 methanol/acetonitrile containing 0.1% formic acid. The flow rate was set at 1.2 ml/min, and the run time was 6 min. 608
Table I. MRM Transitions and Retention Times of Drugs of Abuse* Drug RT (min) MRM CE (V) Amphetamine 2.24 136 91; 136 119 23, 13 Amphetamine-d 11 2.22 147 98; 147 130 25, 13 Methamphetamine 2.38 150 91; 150 119 27, 15 Methamphetamine-d 11 2.33 161 97; 161 127 25, 17 MDA 2.19 180 163; 180 105 13, 31 MDA-d 5 2.19 185 168; 185 110 15, 31 MDMA 2.27 194 163; 194 105 17, 33 MDMA-d 5 2.24 199 165 199 107 17, 35 Codeine 4.28 300 152; 300 115 77, 79 Codeine-d 3 4.28 303 199; 303 115 41, 93 Morphine 3.12 286 152; 286 165 79, 51 Morphine-d 3 3.12 289 185; 289 128 41, 79 Hydrocodone 5.20 300 199; 300 128 43, 73 Hydrocodone-d 3 5.16 303 199; 303 115 41, 97 Hydromorphone 3.74 286 185; 286 157 41, 57 Hydromorphone-d 3 3.70 289 185; 289 128 41,75 Oxycodone 4.93 316 298; 316 241 23, 39 Oxycodone-d 6 4.89 322 304; 322 247 25, 41 Oxymorphone 2.38 302 284; 302 227 23, 41 Oxycodone-d 6 4.89 322 304; 322 247 25, 41 Cocaine 2.44 304 182; 304 82 25, 41 Cocaine-d 3 2.46 307 185; 307 85 27, 43 Benzoylecgonine 0.97 290 168; 290 105 27, 41 Benzoylecgonine-d 3 0.97 293 171; 293 105 27, 43 Phencyclidine 1.80 244 91; 244 86 47, 17 Phencyclidine-d 5 1.81 249 96; 244 86 49, 17 * MRM: multiple reaction monitoring; CE: collision energy; internal standards used in this study are shown in italics. For oxymorphone, oxycodone-d 6 was used as internal standard. Electrospray positive ionization has been used for the detection of all drugs of abuse in oral fluids in this study. Initially standard drug solutions of 200 ng/ml were infused into the MS MS system to optimize the conditions for each drug. Two product ions were selected for each drug and an MRM method was set up to monitor the MRM transitions of each drug during LC MS MS analysis (Table I). Oral fluids containing various concentrations of AMPs (amphetamine, MAMP, MDA, and MDMA) ranging from 10 to 2000 ng/ml final oral fluid concentration were prepared and extracted by liquid liquid extraction and the final extracts were analyzed by LC MS MS. Amphetamine-d 11, MAMP-d 11, MDAd 5, and MDMA-d 5 at a concentration of 50 ng/ml were used as the internal standards. The extracted ion chromatogram of 10 ng/ml AMP mixture is presented in Figure 1A. The linear regression for AMPs was found to be > 0.993. The LOD/LOQ for all the AMPs was determined to be 10 ng/ml oral fluid. The details of HPLC separation of AMPs are given in the legend for Figure 1. Similarly, oral fluid samples containing various concentrations of opiates (codeine, morphine, hydrocodone, hydromorphone, oxycodone, oxymorphone) ranging from 10 to 500 ng/ml were extracted by solid-phase extraction and analyzed by LC MS MS. Codeine-d 3, morphine-d 3, hydrocodoned 3, hydromorphone-d 3, and oxycodone-d 6 at a concentration of 40 ng/ml were used as the internal standards. The extracted ion chromatogram and linear regression curves for each opiate is presented in Figure 2. The linear regression for opiates was > 0.999 and LOD/LOQ for opiates was determined to be 10 ng/ml oral fluid. (Please see the caption for Figure 2 for the details of separation of opiates by HPLC). Oral fluid samples containing cocaine, BE (1 1000 ng/ml), and PCP (1 500 ng/ml) were analyzed by LC MS MS, and the data are presented in Figures 3 and 4. Cocaine-d 3 and BE-d 3 at a concentration of 80 ng/ml and PCP-d 5 at a concentration of 100 ng/ml were used Table II. LOD/LOQ, Precision, and Accuracy Study of Drugs of Abuse in Oral Fluids by LC MS MS* Name LOD/LOQ Precision (% CV) Accuracy (%) Amphetamine 10 6.1 94 Methamphetamine 10 3.1 98 MDA 10 8.2 90 MDMA 10 5.1 96 Codeine 10 3.1 87 Morphine 10 7.6 90 Hydrocodone 10 3.9 96 Hydromorphone 10 2.9 95 Oxycodone 10 4.9 87 Oxymorphone 10 3.7 82 Cocaine 2 4.6 85 Benzoylecgonine 2 2.8 93 Phencyclidine 2 5.0 95 * LOD/LOQ values are expressed as ng/ml of oral fluid. Precision and accuracy studies were carried out by analyzing 10 samples for each drug at 40% of cutoff (20 ng/ml for AMPs, 16 ng/ml for opiates, 8 ng/ml for cocaine and BE, and 4 ng/ml for PCP). Table III. Verification of the Study by Participation in North American Oral Fluid Proficiency Testing* Target Mean of Conc. Participating Sample ID Analyte (ng/ml) Laboratories Range Result OFU-1 Amphetamine 50 60.8 47 84.8 57 OFU-1 Methamphetamine 50 58.7 40 70.8 62 OFU-2 Codeine 40 44.7 33.1 63.1 49 OFU-2 Morphine 60 68.5 45.1 101.6 76 OFU-4 Benzoylecgonine 60 61.3 33 79 60 OFU-5 Cocaine 40 40.7 20.8 51.3 46 OFU-5 PCP 15 14.6 4.4 19 14.6 * North American Oral fluid Proficiency Testing was administered by RTI. The mean value represents the average of results from 17 to 22 laboratories. The results from this study were within 10% of the mean of the all laboratories participated in the proficiency program except for cocaine, where the results came within the 20% of the mean. 609
Figure 5. Analysis of oral fluid samples from real donors and proficiency samples. The ion chromatograms of AMP, cocaine, and BE were from real donor samples whereas opiates and PCP are from proficiency samples. as the internal standards. The linear regressions for these drugs were found to be > 0.998 and the LOD/LOQ for cocaine, BE, and PCP were determined to be 2 ng/ml oral fluid (see details of HPLC separation in the corresponding figure legends). The precision and accuracy studies were conducted using analysis of 10 individual oral fluid samples spiked with 40% of respective cutoffs of drugs of interest, and the data are presented in Table II along with summary of LOD/LOQ. The percentage variation of precision for the drugs ranged between 3 and 10%, and accuracy was found to be within 20% of the target value. Figure 5 represents ion chromatograms of the analysis of either real samples or proficiency samples obtained from RTI. The ion chromatograms of positive samples for amphetamines, cocaine, and BE are presented as Figures 5A and B. For PCP and opiates, the ion chromatograms of the RTI proficiency samples are presented in Figures 5C and D. In this study, to date there were no positive oral fluid samples for PCP and opiates from real donors. The authors have participated in the voluntary proficiency program offered by RTI, and the results are presented in Table III. All the drugs in the proficiency samples were identified by this method, and the quantitative results were found to be within 20% of the target value, suggesting that the methods developed in our laboratory are valid and accurate for the detection and quantitation of drugs of abuse in oral fluids. In conclusion, the authors have developed an LC MS MS method for the detection and quantitation of AMPs, opiates, PCP, cocaine, and BE in oral fluids. This study represents the analysis of different classes of drugs in oral fluids by LC MS MS. Additionally, the analysis of donor samples along with the proficiency samples supplied by RTI confirms and verifies the validity of this method. Presently, studies of identification of other drugs of abuse including barbiturates and benzodiazapines by LC MS MS are in progress. Acknowledgments The authors thank Mr. Jeorge Reyna for his technical assistance. References 1. E.J. Cone, J. Clarke, and L. Tsanaclis. Prevalence and disposition of drugs and opioid treatment drugs in oral fluid. J. Anal. Toxicol. 31(8): 424 433 (2007). 2. S. George. A snapshot of workplace drug testing in the UK. Occup. Med. (Lond.) 55: 69 71 (2005). 3. E.J. Cone and A.J. Jenkins. Saliva drug analysis. In Handbook of Analytical Therapeutic Drug Monitoring and Toxicology. S.H.Y. Wong and I. Sunshine, Eds. CRC Press, Boca Raton, FL, 1997, pp 303 333. 4. M. Concheiro, A. de Castro, O. Quintela, and C.M. Lopez-Rivadulla. Confirmation by LC MS of drugs in oral fluid obtained from roadside testing. Forensic Sci. Int. 170: 156 162 (2007). 610
5. E.J. Cone. Saliva testing of drugs of abuse. Ann. N.Y. Acad. Sci. 694: 91 127 (1993). 6. E.J. Cone and M.A. Huestis. Interpretation of oral fluid tests for drugs of abuse. Ann. N.Y. Acad. Sci. 1098: 51 103 (2007). 7. O.H. Drummer. Drug testing in oral fluid. Clin. Biochem. Rev. 27(3): 147 159 (2006). 8. L. Kadehjian. Legal issues in oral fluid testing. Forensic Sci. Int. 150: 151 160 (2005). 9. Y.H. Caplan and B.A. Goldberger. Alternative specimens for workplace drug testing. J. Anal. Toxicol. 25(5): 396 399 (2001). 10. F.M. Wylie, H. Torrance, R.A. Anderson, and J.S. Oliver. Drugs in oral fluid Part I. Validation of an analytical procedure for licit and illicit drugs in oral fluid. Forensic Sci. Int. 150: 191 198 (2005). 11. M. Mueller, F.T. Peters, G.A. Ricaurte, and H.H. Maurer. Validated liquid chromatographic electrospray ionization mass spectrometric assay for simultaneous determination of 3,4-methylenedioxymethamphetamine and its metabolites 3,4-methylenedioxyamphetamine, 3,4-dihydroxymethamphetamine and 4-hydroxy-3-methoxymethamphetamine in squirrel monkey plasma. J. Chromatogr. 855: 262 270 (2007). 12. K.A. Mortier, K.E. Maudens, W.E. Lambert, K.M. Clauwaert, J.F. Van Bocxlaer, D.L. Deforce, C.H. Van Peteghem, and A.P. Leenheer. Simultaneous, quantitative determination of opiates, amphetamines, cocaine and benzoylecgonine in oral fluid by liquid chromatography quadrupole-time-of-flight mass spectrometry. J. Chromatogr. 779: 321 330 (2002). 13. M. Wood, M. Laloup, M. del Mar Ramirez Fernandez, K.M. Jenkins, M.S. Young, J.G. Ramaekers, G. De Boeck, and N. Samyn. Quantitative analysis of multiple illicit drugs in preserved oral fluid by solid-phase extraction and liquid chromatography tandem mass spectrometry. Forensic Sci. Int. 150: 227 238 (2005). 611