Anal. Chem. 2003, 75, 5710-5718
Toxicological Screening with Formula-Based Metabolite Identification by Liquid Chromatography/Time-of-Flight Mass Spectrometry Anna Pelander,*,† Ilkka Ojanpera 1 ,† Suvi Laks,†,‡ Ilpo Rasanen,† and Erkki Vuori†
Department of Forensic Medicine, P.O. Box 40, and Department of Analytical Chemistry, P.O. Box 55, FIN-00014, University of Helsinki, Finland
An analytical procedure was evaluated for the comprehensive toxicological screening of drugs, metabolites, and pesticides in 1-mL urine samples by TurboIon spray liquid chromatography/time-of-flight mass spectrometry (LC/TOFMS) in the positive ionization mode and continuous mass measurement. The substance database consisted of exact monoisotopic masses for 637 compounds, of which an LC retention time was available for 392. A macroprogram was refined for extracting the data into a legible report, utilizing metabolic patterns and preset identification criteria. These criteria included (30 ppm mass tolerance, a (0.2-min window for absolute retention time, if available, and a minimum area count of 500. The limit of detection, determined for 90 compounds, was 1.0 mg/L for 6% of the compounds. For method comparisons, 50 successive autopsy urine samples were analyzed by this method, and the results confirmed by gas chromatography/mass spectrometry (GC/MS). Findings for parent drugs were consistent with both methods; in addition, LC/TOFMS regularly revealed apparently correct findings for metabolites not shown by GC/MS. Mean and median mass accuracy by LC/TOFMS was 7.6 and 5.4 ppm, respectively. The procedure proved well-suited for tentative identification without reference substances. The few false positives emphasized the fact that all three parameters, exact mass, retention time, and metabolite pattern, are required for unequivocal identification. Forensic toxicology involves the analysis of drugs and poisons in biological specimens and interpretation of the results to be applied in a court of law. Despite the diversified instrumentation in use, availability of analytical reference standards becomes a critical factor when a novel target substance is encountered or a comprehensive screening procedure is updated. In this busy setting, commercial drugs can be acquired within a reasonable period of time; their metabolites generally cannot. The situation is even more complicated in the rapidly changing scene of designer drugs. Identification of low-dose substances in bioma* Corresponding author. Fax: +358-9-19127518. E-mail: anna.pelander@ helsinki.fi. † Department of Forensic Medicine. ‡ Department of Analytical Chemistry.
5710 Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
trixes without reference standards is a challenge to any wellequipped research laboratory, and the forensic analyst in charge usually has to be satisfied with comparing sample mass spectra to those published in electronic libraries for electron impact gas chromatography/mass spectrometry (GC/MS). The number of liquid chromatography/mass spectrometry (LC/MS) applications in forensic toxicological analysis has increased markedly during the past decade.1 In comprehensive drug screening, identification has been based on fragment ions or comparison of full mass spectra, which necessitates reference substances for the construction of spectra libraries. Moreover, the interlaboratory reproducibility of mass spectral libraries obtained by these techniques has been questioned,2 which hinders the creation of universal reference libraries. Accurate mass has been utilized in the monitoring of specific compounds in environmental and biological samples already in combination with packed column gas chromatography3,4 and since the introduction of glass capillary gas chromatography/highresolution mass spectrometry a few decades ago.5,6 However, this approach is limited by expensive instrumentation. Orthogonal acceleration time-of-flight mass spectrometry (OATOFMS) allows continuous mass measurement with moderate resolution (5000) and high mass accuracy (5 ppm). Several affordable benchtop liquid chromatography/time-of-flight mass spectrometry (LC/ TOFMS) instruments were recently launched onto the market. The accurate mass measurement enables formulation of candidate elemental compositions for a particular mass, thus allowing tentative characterization of substances. In addition, predefined exact masses can be searched for identification. Applications by OATOFMS have covered identification and characterization of unknown drug metabolites,7-9 glucuronide conjugates,10 pesticides,11,12 anabolic steroids,13 and quantitative drug analysis.8,14,15 (1) Marquet, P. Ther. Drug Monit. 2002, 24, 255-276. (2) Marquet P. Ther. Drug Monit. 2002, 24, 125-133. (3) Kimble, B. J.; Cox, R. E.; McPherron, R. V.; Olsen R. W.; Roitman E.; Walls, F. C.; Burlingame A. L. J. Chromatogr. Sci. 1974, 12, 647-655. (4) Ehrenthal, W.; Pfleger, K.; Stuebing, G. Acta Pharmacol. Toxicol., Suppl. 1977, 41, 199-211. (5) Burlingame, A. L. Ecotoxicol. Environ. Saf. 1977, 1, 111-150. (6) Lewis, S.; Kenyon, C. N.; Meili, J.; Burlingame, A. L. Anal. Chem. 1979, 51, 1275-1285. (7) Michelsen, P.; Karlsson, Å. Rapid Commun. Mass Spectrom. 1999, 13, 2146-2150. (8) Zhang, N.; Fountain, S. T.; Honggang, B.; Rossi, D. D. Anal. Chem. 2000, 72, 800-806. 10.1021/ac030162o CCC: $25.00
© 2003 American Chemical Society Published on Web 09/24/2003
A preliminary communication from this laboratory introduced the concept of urine drug screening by positive pneumatically assisted electrospray ionization LC/TOFMS with an automated target library search based on elemental formulas.16 This novel approach was based on the assumption that tentative identification of drugs in urine is viable without reference standards by use of exact monoisotopic masses and metabolite patterns from the literature. The present study evaluates this screening methodology to the full with a series of urine samples taken at autopsy and shows its scope and limitations in forensic toxicology practice. EXPERIMENTAL SECTION Materials. Standard substances were obtained from various pharmaceutical companies and were of pharmaceutical purity. All reagents were analytical grade and purchased from Merck (Darmstadt, Germany), except for Jeffamine D-230 (Fluka, Buchs, Switzerland) and β-glucuronidase (Roche, Mannheim, Germany). Acetonitrile and methanol were HPLC grade and purchased from Rathburn (Walkerburn, U.K.); other solvents were from Merck and of analytical grade. Water was Alpha-Q purified (Millipore, Bedford, MA). Isolute HCX-5 (100 mg) mixed-mode solid-phase extraction cartridges were purchased from International Sorbent Technology (Hengoed, U.K.). Urine samples were collected at autopsies. LC/TOFMS. The liquid chromatograph was an Agilent (Waldbronn, Germany) 1100 series system consisting of vacuum degasser, autosampler, binary pump, column oven, and diode array detector. Separation was performed in gradient mode with a Phenomenex (Torrance, CA) Luna C-18(2) 100 × 2 mm (3 µm) column and a 4 × 2 mm precolumn. The column oven was kept at 40 °C. Eluent components were 5 mM ammonium acetate in 0.1% formic acid and acetonitrile. Flow rate was 0.3 mL/min. The proportion of acetonitrile was increased from 10 to 40% in 10 min, to 75% in 13.50 min, to 80% in 16 min, and held at 80% for 3 min. Post-time was 5 min, and injection volume 10 µL. The mass analyzer was an Applied Biosystems (Framingham, MA) Mariner TOF mass spectrometer equipped with a PE Sciex (Concord, ON, Canada) TurboIon Spray source and a 10-port switching valve. The instrument was operated in the positive ion mode. The eluent flow was carried to the ion source without splitting. The nebulizer gas (N2) flow was 0.7 L/min, the curtain gas (N2) flow 1.2 L/min, and the heater gas (N2) flow 8 L/min. The spray tip potential of the ion source was 5.5 kV, and the heater temperature was 350 °C. Interface settings were as follows: nozzle potential 70 V, quadrupole rf voltage 800 V, and quadrupole temperature 140 °C. Skimmer 1 potential, quadrupole dc potential, deflection voltage, and Einzel lens potential varied depending on the daily tuning of the instrument. Analyzer settings were as (9) Zhang, H.; Henion, J.; Yang, Y.; Spooner, N. Anal. Chem. 2000, 72, 33423348. (10) Sundstro ¨m, I.; Hedeland, M.; Bondesson, U.; Andre´n, P. E. J. Mass Spectrom. 2002, 37, 414-420. (11) Thurman, E. M.; Ferrer, I.; Parry, R. J. Chromatogr., A 2002, 957, 3-9. (12) Maizels, M.; Budde, W. L. Anal. Chem. 2001, 73, 5436-5440. (13) Nielen, M. W. F.; Vissers, J. P. C.; Fuchs, R. E. M.; van Velde, J. W.; Lommen, A. Rapid Commun. Mass Spectrom. 2001, 15, 1577-1585. (14) Zhang, H.; Heinig, K.; Henion, J. J. Mass Spectrom. 2000, 35, 423-431. (15) Zhang, H.; Henion, J. J. Chromatogr., B 2001, 757, 151-159. (16) Gergov, M.; Boucher, B.; Ojanpera¨, I.; Vuori E. Rapid Commun. Mass Spectrom. 2001, 15, 521-526.
follows: push pulse potential 492 V, pull pulse potential 225 V, acceleration potential 4.0 kV, reflector potential 1.55 kV, and detector voltage 1.9 kV. Pull bias potential varied depending on daily tuning. Spectrum acquisition time was 2 s, and a m/z range from 100 to 750 was recorded. Daily instrument tuning and three-ion mass scale calibration was performed with 0.5 µg/mL Jeffamine D-230 solution in acetonitrile-0.1% formic acid (1:1) by infusion injection at a flow rate of 50 µL/min. The theoretical exact m/z of the calibration ions were 191.17 544, 249.14 731, and 317.25 917, and a minimum resolution of 5000 was used in the calibration. Automated postrun internal mass scale calibration of individual samples was enabled by injecting the calibration solution in the beginning of each run 10 s after sample injection via a 10-port switching valve equipped with a 20-µL loop. Calibrator ions were same as in the instrument tune. Mass Library and Data Analysis. Theoretical monoisotopic exact masses of protonated compounds based on molecular formula were calculated with the Data Explorer software of the TOF instrument. The library was constructed to include also the molecular formula, retention time (if known), and a 3-5-digit numerical code for each compound. The code expressed the compound group, number of compounds within the group, and the ordinal number of the compound in the group. MS data were analyzed with the screening application macro created with Microsoft Visual Basic for Applications, as included in Data Explorer software. The application performed automatic mass scale calibration, searched for target masses included in the library, and generated an Excel-based results report of positive findings. Search criteria included ppm mass tolerance, retention time window, and minimum area count selected by the operator. Compounds with unknown retention time were reported according to mass tolerance and area count only. The report included theoretical mass, measured mass, ppm difference of these masses, reference retention time, measured retention time, retention time difference, compound name, its molecular formula, peak area, and compound code of each substance identified within the criteria. The compounds identified were arranged according to compound group coding in the report. This code-based reporting order ensured that parent compounds and metabolites were reported together, facilitating report interpretation. Chromatographic Library. The chromatographic library was created by running standard mixtures containing 8-10 drugs, metabolites, or pesticides in acetonitrile-0.1% formic acid (1 + 9, v/v). These mixtures were diluted from 1 mg/mL stock solutions and 0.1 mg/mL working solutions made in methanol with the concentration of each substance in the mixtures being 1 µg/mL. Reproducibility was monitored by running three repetitions during one week and by repeating this three times at one-month intervals. Sample Preparation. Urine samples (1 mL) were hydrolyzed with β-glucuronidase for 2 h in a water bath at 56 °C, and 10 µL of dibenzepin internal standard solution (10 µg/mL in methanol) was added. The extraction was performed according to IST application note IST 1044 A17 with minor modifications. The pH of the samples was adjusted between 5 and 7 by adding 2 mL of 0.1 M pH 6 phosphate buffer. The SPE cartridges were solvated (17) SPE Application note for extraction of acidic, neutral and basic drugs from urine International Sorbent Technology, Hengoed, U.K., 1997; IST 1044 A.
Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
5711
and equilibrated with 2 mL of methanol, 2 mL of water, and 3 mL of 0.1 M pH 6 phosphate buffer. After sample addition, the cartridges were rinsed with 1 mL of 0.1 M pH 6 phosphate buffer and dried under full vacuum for 5 min. The cartridges were further rinsed with 1 mL of 1 M acetic acid and again dried for 5 min. The acidic-neutral fraction was eluted with 3 mL of ethyl acetatehexane (25:75 v/v). The cartridges were dried for 2 min, rinsed with 3 mL of methanol, and again dried for 2 min. Basic drugs were eluted with 3 mL of ethyl acetate-ammonium hydroxide (98:2, v/v). After extraction, the eluates (acidic-neutral fraction and basic fraction) were combined, evaporated to dryness at 40 °C under nitrogen, and reconstituted in 150 µL of acetonitrile0.1% formic acid (1 + 9, v/v). If any HPLC column overload was observed by distorted peak shape, the extracts were diluted 1:1 or 1:10. Limit of Detection. The limit of detection (LOD) was determined in spiked blank urine collected from living volunteers, previously tested to be free of interference. The initial concentration in the LOD studies was 0.1 mg/L. If the compound was detectable, the concentration was reduced to a nondetectable level, and three repetitions were made at the LOD level. If the compound was not detectable at 0.1 mg/L, the concentration was increased until to the detectable level, and the analysis was repeated at the LOD level three times. A compound was regarded as detected when identified and reported by the analysis macro. GC/MS. GC/MS was performed with a 5973 mass-selective detector coupled to a 6890 Plus gas chromatograph, equipped with a 7683 injector and a HP-5MS (12 m × 0.20 mm i.d. with 0.33-µm film) capillary column (Agilent Technologies, Palo Alto, CA). GC/ MS was operated by Chemstation software. The gas chromatograph was used in the splitless and constant-flow mode with a 2-µL injection volume and a column flow of 1.5 mL/min. The injector port temperature was 250 °C, and the transfer line temperature 300 °C. The oven temperature was initially held at 90 °C for 0.5 min and then increased by 20 °C/min to 300 °C, which was held for 3 min. The mass spectrometer was operated in scan mode in the range m/z 50-550 (2.94 scans/s). The identification of unknown peaks was carried out by comparison of the sample mass spectra to the standard collections of library spectra (PMW_TOX3, Nist98 and Wiley275 obtained through Agilent Technologies). Samples positive by LC/TOFMS but negative by GC/MS for benzodiazepines or drugs containing hydroxyl, carboxyl, or primary or secondary amino groups were reanalyzed by the selected ion monitoring (SIM) technique. The sample was evaporated to dryness, and hydroxyl and carboxyl groups were derivatized with BSTFA containing 1% TMCS, and amino groups with trifluoroacetic acid anhydride. GC/MS analysis of the trimethylsilyl and trifluoroacetyl derivatives was performed with SIM of three ions for each compound. Safety considerations: when working with BSTFA appropriate protective clothing should be worn. Analysis of Authentic Samples and Evaluation of Results. Randomly selected autopsy urine samples were analyzed with LC/ TOFMS and full-scan GC/MS and the results compared. Further confirmation of GC/MS results was performed in SIM mode as described above. Two parallel 1-mL samples were extracted as described in the Sample Preparation section, except that for GC/ 5712
Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
MS analysis the extract was reconstituted with 300 µL of ethyl acetate. RESULTS AND DISCUSSION Substance Library. The mass library comprised a total of 637 theoretical monoisotopic masses of drugs, metabolites, and pesticides, of which 392 compounds were available to the authors as reference substances, and their retention times were determined. The remaining 245 substances were metabolites, for which only theoretical mass was included in the library. Both absolute (RT) and relative retention times (RRT) were studied. RRT were calculated using dibenzepin as an internal standard. The complete list of compounds with their retention time data is available as Supporting Information. The precision of RT and RRT was calculated from nine parallel measurements. For RT, the mean and median standard deviations (STDV) were both 0.04 min, and the relative standard deviations (RSTDV) was 0.50 and 0.35%, respectively. For RRT, the mean and median STDV were 0.004 and 0.003, and RSTDV were 0.65 and 0.32%. As RT and RRT proved to be equally reproducible, RT was chosen as chromatographic parameter. The selection of compounds covered a wide spectrum of chemical properties, but the majority were basic drugs. As most of the reference substances ionized readily in the positive ion mode, it is likely that under identical conditions their metabolites would behave in a similar way. The following substances were not ionized under the conditions used: ethyl parathion, methyl parathion, cyclothiazide, dichlorprop, MCPA, ibuprofen, and γ-hydroxy butyrate. Nine substances, acetazolamide, apronalide, chlorpropamide, chlorthalidone, felodipine, phenytoin, primidone, salicylamide, and sulthiame, had poor ionization efficiency, with 10-100-fold concentration required for sufficient peak intensity. Limit of Detection and Library Search Criteria. LOD was determined for 90 drugs or metabolites in urine (Table 1). LOD was e0.1 mg/L for 66 (73%) of the 90 compounds studied. Only five compounds had a LOD g 1 mg/L. Two of these, paracetamol and phenytoin, were undetectable at 50 mg/L. In the case of paracetamol, the majority of the drug was lost in the acidic wash fraction of the SPE procedure. In the case of phenytoin, a combination of poor ionization and possible substance loss in the sample preparation resulted in poor detection. Negative ionization could provide lower LODs for some compounds, e.g., those with carboxylic functions (indomethacin and ketoprofen), and will be further studied in the future. The LODs obtained in this study for drugs and metabolites in urine were higher than those determined for various target analyses by LC/MS18 but at the same level as those obtained for screening-type applications.19,20 Analysis report criteria were set 30 ppm for mass tolerance, 0.2 min for the retention time window, and 500 for the minimum area count. In our previous study, reasonable mass tolerance was defined as 20 ppm for biological samples;16 this study showed that the 20 ppm window was indeed sufficient for molecules with a theoretical exact mass higher than 200, but smaller molecules (18) Van Boxclaer, J. F.; Clauwaert, K. M.; Lambert, W. E.; Deforce, D. L.; Van den Eeckhout, E. G.; De Leenheer, A. P. Mass Spectrom. Rev. 2000, 19, 165-214. (19) Gergov, M.; Robson, J. N.; Duchoslav, E.; Ojanpera¨, I. J. Mass Spectrom. 2000, 35, 912-918. (20) Thevis, M.; Opfermann, G.; Scha¨nzer, W. Biomed. Chromatogr. 2001, 15, 393-402.
Table 1. Limit of Detection (LOD) and Retention Time (RT) of 90 Drugs and Metabolites in Urine by LC/TOFMS compound
RT (min)
LOD (mg/L)
compound
RT (min)
LOD (mg/L)
acebutolol alprenolol amiodarone amitriptyline amphetamine atenolol benzoylecgonine betaxolol bisoprolol buprenorphine carbamazepine carvedilol celiprolol chloroquine chlorpromazine chlorprothixene cisapride citalopram clomipramine clonazepam clonidine clozapine cocaine codeine dextropropoxyphene diazepam diltiazem dixyrazine doxepin ethylmorphine flunitrazepam fluoxetine flupentixol fluvoxamine glibenclamide glipizide hydrocodone 10-hydroxycarbamazepine indomethacin ketoprofen levomepromazine lorazepam LSD maprotiline MDMA (ecstacy)
7.42 10.51 16.33 12.99 3.12 1.59 6.63 10.78 9.61 11.64 13.12 12.36 8.88 2.94 13.59 13.81 11.98 11.44 13.97 14.43 2.63 10.73 8.36 2.47 12.99 16.04 11.74 13.55 11.51 5.11 15.04 13.6 14.61 12.83 16.8 15.01 3.94 9.78 16.79 15.44 13.06 14.29 9.25 12.86 4.18
0.01 0.01 10.0 0.02 0.1 0.1 0.5 0.01 0.02 0.2 0.02 0.1 0.02 0.2 0.1 0.5 0.5 0.02 0.05 0.2 0.01 0.05 0.02 0.05 0.02 0.01 0.02 0.1 0.02 0.05 0.05 0.2 2.0 0.05 0.01 0.1 0.1 0.1 0.5 1.0 0.05 0.2 0.02 0.1 0.05
methadone methamphetamine metoprolol mianserine midazolam moclobemide 6-monoacetylmorphine (MAM) morphine nicotine nizatidine norcitalopram norclomipramine nordiazepam norflunitrazepam normianserine nortriptyline olanzapine orphenadrine oxazepam oxprenolol paracetamol paroxetine perphenazine phenazone phencyclidine phenytoin pindolol practolol promazine propranolol quinine ranitidine risperidone selegiline sotalol sulpiride thioridazine timolol tramadol trimipramine venlafaxine verapamil warfarin zolpidem zopiclone
13.11 3.78 7.52 11.12 11.17 5.89 3.63 1.49 1.07 1.47 11.25 13.78 15.04 13.81 11.00 12.77 3.29 12.10 13.91 9.06 2.44 12.36 13.62 7.32 10.00 13.27 4.47 1.91 12.17 10.35 7.36 1.73 9.63 6.27 1.69 1.96 14.51 7.11 7.45 13.25 9.42 13.03 16.13 9.02 7.34
0.01 0.01 0.02 0.02 0.05 0.01 0.05 0.2 0.5 0.02 0.05 0.1 0.1 0.2 0.05 0.05 0.05 0.05 0.1 0.1 >50 0.5 0.5 0.05 0.01 >50 0.5 0.05 0.1 0.01 0.1 0.05 0.02 0.1 0.5 0.01 0.5 0.02 0.01 0.1 0.01 0.02 0.02 0.01 0.2
were missed. It would be ideal to have a possibility to set the mass tolerance level according to molecule size. The retention time window of 0.2 min was large compared to the precision of the chromatographic study, which indicated that the majority of the compounds would allow a retention time window of 0.1 min. The 0.2-min window was selected to ensure the identification of individual compounds with poorer chromatographic properties. The use of accurate-mass mass chromatograms in searching for targeted compounds efficiently eliminated matrix effects, and sample quality had practically no influence on LOD, so long as chromatography was not overloaded. The present principle allows search for an unlimited number of compounds. However, total data analysis time may become a limiting factor, as the macro goes through each mass sequentially. With 637 library compounds, total data analysis time was 5-15 min depending on sample quality. Application of the LC/TOFMS Method To Authentic Urine Samples. Fifty successive autopsy urine samples were analyzed by the LC/TOFMS method and by GC/MS against commercial
libraries (Table 2). The number of compounds identified was higher by LC/TOFMS than by GC/MS. The overall mass accuracy of true positive findings was excellent: mean and median ppm absolute value differences were 7.6 and 5.4, respectively. With the exception of caffeine, no false negatives were observable by LC/TOFMS, within the present in-house library. The reporting criteria by LC/TOFMS were as described above, and the report was further interpreted as follows: Results for compounds with molecular mass of g200 Da and ppm difference of 20-30 were rejected (see Limit of Detection). The remaining findings within the three criteria (retention time, mass tolerance, minimum area) were considered true positives if a parent drug and at least one metabolite were present. A false positive was defined when a substance fell within the criteria, but no metabolites were identifiable, and GC/MS was negative for the particular compound. These apparent false positives, such as several meclozine, amitriptyline, and dextropropoxyphene findings, were possibly due to components present in postmortem urine. That a total of seven caffeine findings by GC/MS were missed by LC/TOFMS is most likely due to the small molecular mass and poor peak Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
5713
Table 2. Comparison of Screening Analysis Results Obtained by LC/TOFMS and GC/MS for Autopsy Urine Samples case no.
findings by LC/TOFMS
mass error (ppm)
findings by GC/MS
4401
caffeine carvedilol temazepam carbamazepine carbamazepine-10-11-epoxidea (3 candidates) dihydroxycarbamazepinea nicotine cotinine hydroxycotininea nicotine cotinine hydroxycotininea lamotrigine caffeine zopiclone norzopiclonea nicotine cotinine hydroxycotininea mirtazapine normirtazapinea lorazepam caffeine nicotine cotinine hydroxycotininea nicotine cotinine hydroxycotininea diazepam nordiazepam temazepam oxazepam amphetamine caffeine quinine 2-hydroxyquininea/ 3-hydroxyquininea diltiazem nordiltiazema/O-desmethyldiltiazema (3 candidates) deacetyldiltiazema deacetyldiltiazem N-oxidea (4 candidates) deacetylnordiltiazema (3 candidates) O-desmethyldeacetyl-nordiltiaza (3 candidates) nicotine cotinine hydroxycotininea citalopram norcitalopram 7-aminoclonazepam nicotine cotinine hydroxycotininea demoxepam nordiazepam temazepam oxazepam promazine hydroxypromazinea/promazine sulfoxidea (3 candidates) norpromazine dinorpromazinea chlorpromazine hydroxychlorpromazinea/chlorpromazine sulfoxidea norchlorpromazinea dinorchlorpromazinea caffeine levomepromazine norlevomepromazine hydroxylevomepromazinea/levomepromazine sulfoxidea
-29 5.9 -3.3 -2.6 -14/4.5/-14 -2.1 -6.5 -15 -7.5 2.9 3.7 -9.8 -0.2
caffeine carvedilol temazepam carbamazepine -d dihydroxycarbamazepine nicotine cotinine nicotine cotinine lamotrigine ibuprofenc caffeine zopiclone cotinine mirtazapine lorazepam caffeine nicotine cotinine hydroxycotinine cotinine diazepam nordiazepam temazepam oxazepam amphetamine caffeine quinine diltiazem O-desmethyldiltiazem -
4402
4403b
4404
4405
4407b 4408
4409 4410
4411
4412b
5714 Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
-23 5.2 -3.4 -8.8 -14 -2.8 -0.7 -7.7 -4.4 -23 -6.6 -17 -2.9 0.6 3.2 -18 -8.0 -12 -14 0.9 -2.6 -15 -18 -15 -6.2 4.3/-2.0/2.2 -4 -12/-11/-12/-13 -10/-2.3/-5.7 -8.5/-4.3/-19 -3.9 -7.0 -2.3 -1.1 -3.3 0.4 -3.8 -7.6 -0.3 0.6 -17 -4.8 -2.8 -4.5 1.2/0.2/1.4
citalopram 7-aminoclonazepam nicotine cotinine nordiazepam temazepam oxazepam promazine hydroxypromazine
1.0 -8.6 3.5 -2.5/-2.1
chlorpromazine hydroxychlorpromazine
-8.5 8.7 -25 2.7 -0.9 -1.4/-5.3
chlorpromazine-metab. 2 caffeine levomepromazine -
false positives by LC/TOFMS
meclozine cotinine
meclozine meclozine caffeine
meclozine
nicotine
Table 2 (Continued) case no.
findings by LC/TOFMS
mass error (ppm)
findings by GC/MS
4412b
zopiclone norzopiclonea paroxetine caffeine nicotine cotinine hydroxycotininea caffeine citalopram norcitalopram dinorcitaloprama diazepam temazepam oxazepam zopiclone norzopiclonea 4-methylaminoantipyrine trimethoprim fluvoxamine celiprolol nicotine cotinine quinine 2-hydroxyquininea/3-hydroxyquininea caffeine diltiazem nordiltiazema (3 candidates) deacetyldiltiazema deacetyldiltiazem N-oxidea deacetylnordiltiazema (3 candidates) O-desmethyldeacetylnordiltiaza (2 candidates) -
3.4 -5.0 -9.6 -14 -5.9 -3.1 0.0 -13 1.2 5.0 1.3 7.7 4.7 -9.5 -3.1 -14 -5.3
zopiclone paroxetine caffeine nicotine cotinine caffeine citalopram norcitalopram diazepam temazepam oxazepam zopiclone 4-methylaminoantipyrine acetamidoantipyrinec trimethoprim fluvoxamine celiprolol nicotine cotinine quinine caffeine diltiazem -
4413
4414
4415b
4416
4417
4418 4419b
4420
4421
4423b
4425
oxazepam carbamazepine nicotine cotinine hydroxycotininea venlafaxine norvenlafaxinea/O-desmethylvenlafaxinea nicotine cotinine hydroxycotininea lidocaine caffeine amitriptyline amitriptyline N-oxidea (3 candidates) nortriptyline (E)-10-hydroxyamitriptyline (Z)-10-hydroxyamitriptyline (E)-10-hydroxynortriptyline (Z)-10-hydroxynortriptyline desmethylnortriptylin zopiclone norzopiclonea nicotine cotinine hydroxycotininea lorazepam thioridazine thioridazine 5-sulfoxide northioridazinea nicotine cotinine fluoxetine norfluoxetinea nicotine cotinine -
10 -2.1 9.3 -7.0 -8.8 -18 -11 -26 9.1 16/15/13 9.1 4.7 -5.0/0.0/2.9 -2.1/5.1 -1.3 -0.9 1.1 1.4 -14 -7.8 -6.4/2.0 -4.9 -14 -2.7 2.7 -27 -13 2.2/-4.4/0.3 -1.7 -5.2 1.1 -7.5 -2.4 -7.7 12 6.6 -9.0 -17 -5.2 4.4 0.3 0.9 -11 9.7 4.6 -4.8 -3.5 -9.8 -17
oxazepam carbamazepine nicotine cotinine venlafaxine norvenlafaxine nicotine cotinine caffeine lidocaine caffeine amitriptyline nortriptyline hydroxyamitriptyline zopiclone nicotine cotinine lorazepam thioridazine mesoridazine cotinine fluoxetine nicotine cotinine caffeine
false positives by LC/TOFMS
phenazone metoclopramide
meclozine
amitriptyline dextropropoxyphene
amitriptyline dextropropoxyphene
dextropropoxyphene
diazepam dextropropoxyphene
dextropropoxyphene lidocaine ephedrine
Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
5715
Table 2 (Continued) case no.
findings by LC/TOFMS
mass error (ppm)
4426
nicotine cotinine hydroxycotininea nicotine cotinine hydroxycotininea nicotine cotinine chlorprotixene chlorprotixene sulfoxidea/hydroxychlorprotixenea (3 candidates) norchlorprotixene norchlorprotixene sulfoxidea (4 candidates) chlorprotixene N-oxide sulfoxidea (6 candidates) nicotine cotinine hydroxycotininea pindolol hydroxypindolola olanzapine norolanzapinea paroxetine fluoxetine norfluoxetinea propranolol mianserine nicotine cotinine mirtazapine normirtazapinea nicotine cotinine hydroxycotininea atenolol temazepam oxazepam carbamazepine carbamazepine-10-11-epoxidea (2 candidates) dihydroxycarbamazepina oxcarbazepine 10-hydroxycarbamazepine oxycodone fluoxetine caffeine nicotine cotinine hydroxycotininea caffeine cotinine hydroxycotininea oxazepam -
-4.2 -13 -5.0
4428 4429 4430b
4431b
4432
4433b
4434 4436 4463
4464 4465
4466 4467 4468 4469 4470
-6.5 -17 -4.1 -19 -22 3.1 -0.8/0.3/1/3 -6.8 3.7/3.7/4.2/0.4 -2.4/5.6/-1.5/-2.6/-7.3/-3.7 3.0 2.4 -15 -2.8 -8.0 3.4 5.4 1.9
caffeine nordiazepam temazepam oxazepam codeine norcodeine nicotine
5716 Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
-5.1 -3.3 -7.4 -1.9 5.4 4.3 -1.4 0.1 -9.1 -16 -1.3 2.6 -5.3 -3.7 -1.5 -8.7/-2.2 0.4 -2.2 -3.2 -8.9 -5.5 -18 -12 -5.5 -7.1 -17 -17 -19
findings by GC/MS nicotine cotinine caffeine nicotine cotinine -
false positives by LC/TOFMS dextropropoxyphene
amitriptyline dextropropoxyphene amitriptyline dextropropoxyphene
caffeine chlorprotixene chlorprotixene metabolite nicotine cotinine pindolol olanzapine paroxetine valproic acidc fluoxetine norfluoxetine propranolol propranolol metabolite mianserine nicotine cotinine mirtazapine hydroxymirtazapinec caffeine nicotine cotinine atenolol caffeine temazepam oxazepam carbamazepine carbamazepine metab. 1 dihydroxycarbamazepin oxcarbazepine oxycodone ibuprofenc fluoxetine caffeine nicotine cotinine caffeine nicotine cotinine
-1.2
oxazepam ibuprofenc
-29 -7.4 -11 -3.8 4.7 1.7 -21
caffeine nordiazepam oxazepam codeine nicotine
amitriptyline dextropropoxyphene
ketoprofen
lidocaine
mepyramine
caffeine cotinine thiethylperazin mesoridazine
Table 2 (Continued) case no.
findings by LC/TOFMS
mass error (ppm)
findings by GC/MS
4470
cotinine hydroxycotininea venlafaxine norvenlafaxinea/O-desmethylvenlafaxinea caffeine nicotine cotinine hydroxycotininea oxazepam diazepam temazepam oxazepam amitriptyline amitriptyline N-oxidea nortriptyline (E)-10-hydroxyamitriptyline (Z)-10-hydroxyamitriptyline (E)-10-hydroxynortriptyline (Z)-10-hydroxynortriptyline levomepromazine norlevomepromazine hydroxylevomepromazinea/levomepromazine sulfoxidea nicotine cotinine hydroxycotininea mirtazapine normirtazapinea venlafaxine norvenlafaxinea O-desmethylvenlafaxinea perphenazine hydroxyperphenazinea/sulfoxyperphenazinea nordiazepam temazepam oxazepam ranitidine ranitidine N-oxidea/ranitidine S-oxidea norranitidinea fluoxetine norfluoxetinea caffeine zopiclone norzopiclonea nicotine cotinine hydroxycotininea nicotine cotinine hydroxycotininea caffeine nicotine cotinine hydroxycotininea nicotine cotinine hydroxycotininea nicotine cotinine hydroxycotininea caffeine levomepromazine norlevomepromazine hydroxylevomepromazinea/levomepromazine sulfoxidea nicotine cotinine hydroxycotininea
-16 -13 -14 -13/-7.1
cotinine venlafaxine O-desmethylvenlafaxine ibuprofenc caffeine nicotine cotinine oxazepam caffeine temazepam oxazepam amitriptyline nortriptyline hydroxyamitriptyline
4471
4472 4473b
4474 4475
4490 4491
4492 4493
4494
-29 -20 -19 -14 3.7 -19 -2.0 -9.9 -14 5.8 7.6 6.5 5.1 4.9 7.3 -6.8 -3.8 2.9/3.6 12 8.2 -6.8 -4.6 6.0 -4.7 4.2/3.4 4.2/3.4 1.3 -2.7/-6.5 -9.8 3.1 2.9 1.4 0.5/1.7 15 -6.4 0.3 -21 11 -1.3 -11 -11 -5.7 -15 -14 -12 -28 -14 -19 -11 -9.0 -15 -4.5 -11 -14 -12 -15 9.9 4.1 5.9/-1.3 -2.2 -9.4 -3.8
levomepromazine hydroxylevomepromazine nicotine cotinine mirtazapine hydroxymirtazapinec venlafaxine norvenlafaxine O-desmethylvenlafaxine perphenazine nordiazepam temazepam oxazepam ranitidine fluoxetine caffeine zopiclone nicotine cotinine nicotine cotinine caffeine nicotine cotinine nicotine cotinine nicotine cotinine caffeine caffeine levomepromazine -
false positives by LC/TOFMS
meclozine caffeine ephedrine alprazolam
nicotine cotinine -
Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
5717
Table 2 (Continued) case no.
findings by LC/TOFMS
mass error (ppm)
findings by GC/MS
false positives by LC/TOFMS
4495
caffeine propranolol nicotine cotinine
-22 1.0 -21 -16
caffeine propranolol nicotine -
demoxepam alprazolam
a Metabolite identified by exact mass only. b Sample diluted for LC/TOFMS analysis. c Compound not included in LC/TOFMS library. d Not detected/no drug detected.
Table 3. Original Results Report of Sample 4402 Generated by the Analysis Macroa
mass found
reference mass
ppm error
retention time
reference retention time
retention time error
compound
formula
peak area
compd code
179.1143 301.0748 237.1029 253.1007 253.0983 253.1007 271.1083 163.124 177.1048 193.0986 193.0985 207.1162 207.1163 296.1754
179.1179 301.0738 237.1022 253.0972 253.0972 253.0972 271.1077 163.123 177.1022 193.0972 193.0972 207.1128 207.1128 296.1757
20.3 -3.3 -2.6 -13.9 -4.5 -13.9 -2.1 -6.5 -14.6 -7.5 -6.9 -16.5 -17 1
1.38 15.07 13.11 8.89 10.34 10.91 8.89 1.13 1.38 1.2 2.06 1.6 2.27 9.39
0 15.12 13.12 0 0 0 0 1.07 1.42 0 0 0 0 9.34
1.38 0.05 0.01 8.89 10.34 10.91 8.89 0.06 0.04 1.2 2.06 1.6 2.27 0.05
glycinexylidine temazepam carbamazepine carbamazepine-10-11-epoxide carbamazepine-10-11-epoxide carbamazepine-10-11-epoxide dihydroxycarbamazepine nicotine cotinine hydroxycotinine hydroxycotinine phenylethylmalonamide phenylethylmalonamide dibenzepin
C10H14N2O C16H13N2O2Cl C15H12N2O C15H12N2O2 C15H12N2O2 C15H12N2O2 C15H14N2O3 C10H14N2 C10H12N2O C10H12N2O2 C10H12N2O2 C11H14N2O2 C11H14N2O2 C18H21N3O
704.8 2145.1 4673.7 736 1236.3 2453.6 5727.2 39030.6 31866.4 4796.4 617.6 617.7 7427.8 18812
533 1343 1541 1542 1542 1542 1543 4931 4932 4933 4933 5742 5742 n/a
a
Lines in boldface type considered true positives.
shape of caffeine, which resulted in weak signals and ppm differences above 30. Metabolites without retention time were included in the original report according to the mass tolerance and area count criteria only. The finding was considered a true positive if supported by the presence of a parent compound with reference retention time. If no other substances belonging to the same metabolite pattern were reported, the result was rejected. Consequently, identification based solely on the present mass accuracy was insufficiently reliable, but the combination of exact mass, metabolite patterns, and available retention times proved feasible. In routine screening, it will be possible to collect confirming retention data to assign metabolites. For some metabolites, several candidates appeared, and their confirmation will be more complicated. UV diode array detector data may provide additional confirmatory information for some metabolites, because the UV spectra of parent drug and metabolites are often nearly identical. The current Data Explorer software does not, however, include UV spectral library properties. Table 3 shows the original results report for a autopsy urine sample generated by the screening application macro. The lines printed in boldface type were interpreted as true positive findings. Glycinexylidine, phenylethylmalonamide, and the second proposal for hydroxycotinine were rejected as false positives. Assigning hydroxycotinine to RT 1.20 was based on 26 reported hydroxycotinine findings with RT varying from 1.13 to 1.31 min, with a mean of 1.23 min. In the compound code column, the first 1-3
5718
Analytical Chemistry, Vol. 75, No. 21, November 1, 2003
digit part of the code defines the compound group number, e.g., 15 for carbamazepine. The second digit from the right, 4 for carbamazepine, defines the total number of compounds in the group, and the last digit, 1 for carbamazepine, defines the ordinal number of the compound in the group. This facilitates correct interpretation of results, as compounds belonging to the same metabolic pattern are reported together. CONCLUSIONS Screening urine samples for drugs, metabolites, and pesticides by positive ion LC/TOFMS, based on a library of exact monoisotopic masses, metabolic patterns, and LC retention times, if available, proved to be feasible in forensic toxicological practice. This approach facilitates rapid examination of cases involving new or nonavailable compounds. It is anticipated that the evolution of affordable MS techniques with high mass accuracy will direct toxicological screening toward formula-based identification without immediate need for reference substances. SUPPORTING INFORMATION AVAILABLE Retention time data for 392 compounds included in the chromatographic study. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review April 22, 2003. Accepted July 18, 2003. AC030162O
