Assessment of the Ion-Trap Mass Spectrometer for Routine Qualitative and Quantitative Analysis of Drugs of Abuse Extracted from Urine*"

Transcription

1 Assessment of the Ion-Trap Mass Spectrometer for Routine Qualitative and Quantitative Analysis of Drugs of Abuse Extracted from Urine*" Shawn P. Vorce ~, Jason H. Sklerov, and Kathryn S. Kalasinsky Office of the Armed Forces Medical Examiner, Division of Forensic Toxicology, Armed Forces Institute of Pathology, 1413 Research Boulevard, Building 102, Rockville, Maryland Abstract The ion-trap mass spectrometer (MS) has been available as a detector for gas chromatography (GC) for nearly two decades. However, it still occupies a minor role in forensic toxicology drugtesting laboratories. Quadrupole MS instruments make up the majority of GC detectors used in drug confirmation. This work addresses the use of these two MS detectors, comparing the ion ratio precision and quantitative accuracy for the analysis of different classes of abused drugs extracted from urine. Urine specimens were prepared at five concentrations each for amphetamine (AMP), methamphetamine (METH), benzoylecgonine (BZE), ~9-carboxy-tetrahydrocannabinol (A9-THCCOOH), phencyclidine (PCP), morphine (MOR), codeine (COD), and 6-acetylmorphine (6-AM). Concentration ranges for AMP, METH, BZE, A9-THCCOOH, PCP, MOR, COD, and 6-AM were , , , , 1-250, , , and ng/ml, respectively. Sample extracts were injected into a GC-quadrupole MS operating in selected ion monitoring (SIM) mode and a GC-ion-trap MS operating in either selected ion storage (SIS) or full scan (FS) mode. Precision was assessed by the evaluation of five ion ratios for n injections at each concentration using a single-point calibration. Precision measurements for SJM ion ratios provided coefficients of variation (CV) between 2.6 and 9.8% for all drugs. By comparison, the SIS and FS data yielded CV ranges of % and %, respectively. The total ion ratio failure rates were 0.2% (SIM), 0.7% (SIS), and 1.2% (FS) for the eight drugs analyzed. Overall, the SIS mode produced stable, comparable mean ratios over the concentration ranges examined, but had greater variance within batch runs. Examination of postmortem and quality-control samples produced forensically accurate quantitation by SIS when compared to SIM. Furthermore, sensitivity of FS was equivalent to SIM for all compounds examined except for 6-AM. * The opinions and assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense. ~' This work was presented in part at the annual meeting of the Society of Forensic Toxicologists, San Juan, Puerto Rico, Author to whom correspondence should be addressed. Introduction Over the last two decades gas chromatography-mass spectrometry (GC-MS) has been considered the gold standard for qualitative and quantitative analysis for drugs of abuse in forensic toxicology drug-testing laboratories (FTDTL) (1-3). The Department of Defense (DOD) and most federal drugtesting programs mandate the confirmation of illicit substances by GC-MS. The overwhelming majority of these confirmatory procedures involve the quadrupole MS operating in selected ion monitoring (SIM) mode (4). Although a full spectral scan (FS) from an MS yields the most conclusive identification of target analytes, it suffers from lack of sensitivity, especially with extracts of biological samples. Alternatively, SIM improves sensitivity, but produces less specific identification. DOD defines the criteria for GC-SIM-MS identification of drugs using retention time and relative ion ratios (4). DOD guidelines further state that a minimum of three ions must be monitored for the analyte and two for the internal standard. The Department of Health and Human Services states that quantitative values must be within + 20% or + 2 standard deviations (whichever is larger) of the calculated group reference mean value in order to be valid (5). In addition, there are minimum criteria for acceptance of retention times, peak shapes, and signal-to-noise ratio before a sample is designated as positive. An alternative to the quadrupole MS is the ion-trap MS (6-9). Since the introduction of the first commercial GC-ion-trap instrument in 1983, the ion trap has been in the forefront of ion-molecule reaction (10) and ion frequency (11) studies. However, the ion-trap MS still has a relatively minor role in the FTDTL. Historically, the ion trap has been criticized for its perceived spectral instability. Problems such as space charging, mass misassignments, self-protonation, and poor reproducibility have been noted (12,13). These problems are inherent in the physical boundaries associated with ion trap design (14). The quadrupole MS produces and detects ions in a continuous time Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission. 595

2 Journal of Analytical Toxicology, Vo]. 24, October 2000 frame. Ion traps, on the other hand, ionize, store, and detect ions in cycles operating in the millisecond scale. Although this storage ability provides the ion trap with the time necessary to conduct further ion manipulations such as tandem MS, it can also create problems when the storage capacity of the detector is exceeded. The work reported here was conducted to evaluate the performance of an internal ionization ion-trap GC-MS system for urinary drug confirmations. The particular focus was the instrument's ability to produce reproducible ion ratios across broad concentration ranges and accurate quantitations. The performance was judged against that of a quadrupole MS GC-MS system. The quadrupole instrument was operated in SIM mode. The ion trap was used in both full scan (FS) and selected-ion-storage (SIS) mode. The most prevalent drugs of abuse, as identified by the Substance Abuse and Mental Health Services Administration (SAMHSA), were selected as model compounds for this work. Experimental Chemicals and reagents All drug standards and their deuterated internal standards (Table I) were obtained in methanolic solutions from Radian (Austin, TX). Pentafluoropropionic anhydride (PFPA) was obtained from Aldrich Chemical Co. (Milwaukee, WI), and hep- Table I. Sample Extraction Parameters Extraction Calibrators volume Concenlrations (DOD cutoffs) Drug (ml) Derivative (ng/mt) (ng/ml) AMP 2 HFBA 50,220,500,900, AMP-d8 METH 2 HFBA 50,220,500,900, METH-dH BZE 3 butyl 15,40,100,160, BZE-d~ THCCOOH 3 TMS 1.5,6,15,24,65 15 THCCOOH-d~ PCP 3 N/A 1,10,25,40, PCP-d5 MOR 1 TMS 500,1600,4000,6400, MOR-d~ COD 1 TMS 250,800,2000,3200, COD-d~ 6-AM 3 PFPA 1.5,4,10,25, AM-d Table II. Gas Chromatography Parameters Drug Injector temp (~ AMP 150~ METH 150~ BZE 270~ THCCOOH 270~ PCP 250~ MOR 270~ COD 270~ 6-AM 270~ tafluorobutyric anhydride (HFBA) was purchased from Pierce Chemical (Rockford, IL). Iodobutane, dimethylsufoxide, and N,O-bis(trimethylsilyl)trifluoro-acetamide with 1% trimethylchlorosilane (BSTFA) were obtained from Sigma Chemicals (St. Louis, MO). All other chemicals were obtained from Fisher Scientific (Pittsburgh, PA) or Fluka (Switzerland). Clean Screen CSDAU020 solid-phase extraction columns were obtained from United Chemical Technologies (Bristol, PA). Certified-negative urine, containing 0.1% sodium azide as a preservative, was obtained from obtained from Bio-Rad (Irvine, CA). Samples Table I lists the urine concentration of each drug or metabolite prepared for the study. Each sample was prepared as a 25-mL aliquot using drug-free urine buffered to ph 5 with the exception of Ag-carboxy-tetrahydrocannabinoi (Ag-THCCOOH), morphine (MOR), codeine (COD), and 6-acetylmorphine (6-AM), which were prepared in ph 7 urine. Comparative instrumental accuracy was assessed using urine samples obtained from the DOD Drug Detection Quality Assurance Laboratory, Armed Forces Institute of Pathology (AFIP, Rockville, MD). Postmortem samples were obtained from the Postmortem and Human Performance Branch of the Division of Forensic Toxicology, Office of the Armed Forces Medical Examiner (OAFME, Rockville, MD). Extraction All extractions were performed using 10-mL Clean Screen CSDAU020 solid-phase extraction columns. The extraction procedures were adopted from the Clean Screen application manual (15). The benzoylecgonine (BZE) samples required a Internal standard Hold Column conditions pre post J&W DB-5MS 15 m x 0.25mm x 0.25 pm (min) (rain) 65~ to 10~ to 40~ ~ to 10~ to 40~ ~ to 20~ to 40~ ~ to 30~ ~ to 30~ ~ to 15~ ~ to 15~ 1 2 ] 70~ to 10~ to 40~ 1 1 liquid-liquid acid back extraction subsequent to solid-phase extraction and formation of the butyl derivative. Table I lists the extraction volumes, derivatives, and internal standards for all eight compounds studied. Instrumentation Quadrupole MSD GC-MS. A Hewlett- Packard (HP, Palo Alto, CA) 5890 GC equipped with an HP 7673 autosampler was used to introduce the sample into an HP 5972 MSD. Injections for all drugs were in the splitless mode except for MOR and COD. The GC injector and oven temperatures for all drug assays are listed in Table II. The septum purge flow rate was maintained at 10 ml/min. The split vent was set to open at 0.8 rain post injection at 50 ml/min. The MSD was operated in SIM mode. Instrumental control and data integrations were controlled by HP Chemstation software version B (HP, Palo Alto, CA). A standard autotune with PFTBA was performed on each day of operation. The mass-to-charge ratio values monitored for each drug assay are listed in Table III. 596

3 lon-trap GC-MS. A Varian 3400 GC (Varian, Sugarland, TX) equipped with a Varian 8200 autosampler was used to inject samples into a temperature-programmable Varian 1078 split/splitless injector. All samples were injected splitless except for MOR and COD. The septum purge flow was maintained at 12 ml/min. The split vent was programmed open at 0.8 min after injection with a 55 ml/min flow rate. Table II lists the GC parameters used for each drug analysis. Analyte detection was performed with a Varian Saturn 2000 (Varian, Sugar Land, TX) internal ionization ion-trap MS equipped with a waveboard for SIS applications. Endcap and ring electrodes were pretreated with Silcosteel (Restek, Bellafonte, PA). The MS manifold and ion trap temperatures were set to 50~ and 200~ respectively. Instrument control and data analysis were performed using Saturn GC-MS Workstation software version (Varian, Sugarland, TX). Table III lists the SIS and FS ranges used for the ion trap experiments. A standard autotune with PFTBA was performed on each day of operation. Ion ratio precision The concentrations used for each drug and metabolite are listed in Table I. These values were selected based upon the limits of linearity and quantitation measured using the authors' existing quadrupole GC-MS procedures. The DOD cutoff concentrations for each drug were designated as calibrators (Table I). Following extraction and derivatization, each sample was sequentially injected 15 times into both the quadrupole and iontrap GC-MS instruments. Integrated peak areas of analyte (three ions) and internal standard (two ions) were used to calculate four ion ratios for each sample. A fifth ion ratio was used to generate a multipoint linearity curve for each method. Ion ratios for the replicates at each concentration were compared to those of the cutoff calibrator. Ion ratio failures were defined as those values that lay outside the 20% tolerances defined by the calibrator. The mean, standard deviation, and percent coefficient of variation (%CV) for each of the four ion ratios at each drug concentration were calculated (n = 75). For each drug analyzed with each instrument, the mean relative standard deviation was calculated as the average %CV from the four ion ratios n = 300 (4 ion ratios; 75 injections). The F test was employed to determine which detector produced more stable ion ratios over each concentration range by comparing the mean %CV values. The F test is a statistical analysis used to compare the precision of two sets of measurements. The F test was used to determine if (1) detector A was more precise than detector B and (2) there was a significant difference between the precision of the two detectors. The mean %CVs of the SIS and FS ion trap results were compared to the SIM mean %CV to determine which instrument was more precise for the measurement of each drug. The SIS and FS mean %CVs were also compared (Table IV). Quantitative accuracy Quantitative accuracy was evaluated for each instrument by the analysis of AFIP proficiency samples. Accuracy for the proficiency samples was defined by the mean concentrations resulting from the analysis of the specimens by seven participating DOD laboratories. All of the laboratories employed quadrupole SIM for their analyses. Three separate months of proficiency samples were obtained for each drug. Five aliquots of each month's proficiency samples were analyzed using ion trap SIS. Five-point calibration curves were used for quantita- Table III. Mass Spectrometry Parameters Ion Trap (SIS) MSD (SIM) Target TIC Mass range Drug ions (m/z) (counts) (m/z) SIS ranges (m/z) Ion Trap (full scan) Target TiC Mass range (counts) (m/z) AMP-d~ 243,126 AMP 240,118,117 METH-du 260,213 METH 254,210, , , , , , , , BZE-d3 348,275 BZE 345,272, , , , THCCOOH-d3 476,374 THCCOOH 488,473, , , , PCP-ds 205,246 PCP 243,242, , , , MOR-d 3 432,417 MOR 429,414,401 COD-d3 374,346 COD 371,343, O , , , O-402, , AM-d3 417,364 6-AM 473,414, , ,

4 tion. The samples were acceptable if their retention times and ion ratios fell within _+ 2% and 20% of the calibrators, respectively. Results were defined as accurate if they fell within 20% of the group mean concentration. batches. Coefficients of variation for the SIS samples were < 10%. All of the results acquired in SIS mode demonstrated acceptable accuracy. Results Table IV lists the mean %CV, standard deviations, and the number of ion ratio failures for each drug analyzed by each of the three MS modes. The ion-trap SIS mode demonstrated the highest ion ratio precision for the analysis of BZE and COD (99% confidence level according to the F test). The best precision for amphetamine, methamphetamine, MOR, phencylidine, and 6-AM was obtained using quadrupole SIM. There was no significant difference in the precision for Ag-THCCOOH between the three MS modes. The total number of ion ratio failures for 600 injections made in both SIM and SIS modes were 1 and 4, respectively. This equates to a total ion ratio failure rate of 0.2% for SIM and 0.7% for SIS. The ion trap FS mode had more failed ratios than either SIM or SIS. There was an 80% ratio failure at the lowest concentration of 6-AM (1.5 ng/ml) when analyzed by FS. When the 6-AM values were excluded, the FS mode produced seven failed ratios or a failure rate of 1.2%. The results of the SIS analysis of AFIP proficiency samples are shown in Table V. There were no ratio failures in any of the Table IV. Mean Relative Standard Deviations and Ion Ratio Failures Mean %CV SD Failures Drug SIM SIS FS SIM SIS FS SIM SIS AMP METH BZE , THCCOOH PCP MOR ,5 4,1 0 COD AM * Twelve failures at 1.5 mg/ml; precision based on zero failures with the four highest concentrations. Discussion An important benefit of the ion trap is its ability to acquire full scan spectra at concentrations requiring SIM for quadrupole instruments (16). Figure 1 shows the SIM and FS spectra from a urine extract containing 1 ng/ml of A9-THC- COOH. The sensitivity of the FS analysis was complimented by the abundant fragmentation data acquired. Full scan ion-trap detection matched the sensitivity of the SIS scan function for all compounds except 6-AM. The increased signal-to-noise ratio achieved using SIS would be applicable to analyzing samples or matrices requiring lower limits of detection, such as hair. In preliminary analysis, the major cause of ion ratio failures for the ion trap was mass misassignment. One example of this was the rn/z 340 for the TMS derivative of 6-AM. The ion ratios for the FS analyses all passed except for two ratios at lower concentrations. However, in SIS mode there was an approximate 40% failure rate at all concentrations. This was attributed to the mass resolution of the ion trap. The actual mass for 6-AM-TMS is m/z The ion trap analyzer assigned an integer mass based upon its capacity for unit mass resolution. The consequence of this was a misassignment of the fragment ion as m/z 341. Figure 2 is the extracted ion chromatograph FS (EIC) of a 6-AM-TMS m/z 340 fragment ion. The mass misassignment occurred in the third scan across (scan 457) the chromatographic 0 o peak. Scans 456 and 458 exhibited the correct o o mass assignment for the fragment ion (Figure 1 4 2). The consequence of the misassignment was 0 o ion ratio failures. Assigning a mass defect 1 2 correction factor resulted in correction of 0 1 the assignment for m/z 340 but altered other o o fragment ion masses. This problem was solved 2 0/12' by selection of an anhydride derivative for 6-AM. Although this solution worked in the case of 6-AM, it is obviously not amenable Table V. Comparison of Ion Trap Accuracy with AFIP (SIM) Proficiency Samples Drug mean SD %CV Group mean +20% Group mean -20% sample(n= 5) psl ps2 ps3 psl ps2 ps3 psl ps2 ps3 psl ps2 ps3 psl ps2 ps3 AMP ,1 574, METH , ZE A THCCOOH 6, PCP , MOR ,5 107, COD AM ,

5 Journal of Analytical Toxicology, Vol, 24, October 2000 to all compounds that are monitored by ion trap selected ion storage. The sensitivity of the ion-trap MS demands that careful 7S'~ i +, ~ Time I00~ 7[dt Z~t 0% Ion Trap Full Sou T (rain) 473 4M J I qu-arepole S~ il +l+ 6 T. ~ J.+ m/z Time (rnin) Figure 1. Comparison of ion-trap full scan and quadrupole SIM spectra of 1 ng/ml A9-THCCOOH extracted from urine. 2~t "+ / 287 ne ~10~em-too~) ~ '~ te~t,..,,,,~,,+.,~ Tlmo(mJ~ o.. "" Figure 2. Mass missassignment of the m/z 340 ion of the 6-AM-TMS derivative, 234 mcz h,h 1f !4 I, '~ 'ao '~IS '300 '3~S '3CO m/z Figure 3. Unwanted m/z 356 ion in the ion trap SIS codeine-tms spectrum I 371 attention be paid to the mass-to-charge ratio ranges selected for SIS and chromatographic separation. Figure 3 shows the SIS spectrum of codeine-tms. The m/z 356, while an identifiable COD-TMS fragment ion, has an elevated in- IiT rain. Scan: rain. Scan: 457 EIC Ions: 340 I, ~3T$ ~Q tensity because of the presence of m/z 356 in the bleed spectrum of the capillary column. The degree of interference from the bleed m/z 356 would fluctuate depending on the concentration of the analyte. The higher analyte concentrations were less affected than the lower concentrations. Using m/z 356 as a qualifier for SIS would result in a high ratio failure rate at lower concentrations. Additionally, selection of low-bleed capillary columns with thinner stationary phases would reduce these problems. The SIS scan function operates by periodic modulation of the RF field to bring unwanted ions into resonance with a multifrequency ejection waveform. Mass ranges to be stored are treated as frequency "notches" in the waveform and are retained for later detection. Ions with masses near the selected ranges may maintain sufficient stability in the ion trap and avoid ejection. An example of this can be seen in the ion trap SIS spectrum of butyl-bze (Figure 4) which contains the m/z 243 internal standard ion that lies outside the m/z storage range. In SIS mode, a mass range for storage, typically 3, is assigned (Table IIl). The smaller the mass range, the more selective the analysis. Figure 4 illustrates a major difference between SIS and SIM spectra. Selected ion monitoring only filters the ions selected, whereas SIS produces spectra with additional masses near the mass of interest. The rn/z 243 present in the SIS spectrum originates from the tri-deuterated internal standard that is stabilized in the ion trap although it lies outside the SIS range selected. If other large abundance ions are inadvertently stored during the SIS scan, the ionization time will be reduced by the automatic gain control (AGC). There will be loss of sensitivity and an increased probability for interference. The target total ion current (TIC) is a userdefined parameter that affects the number of ions stored in the ion trap during acquisition. The target TIC sets a limit on the number of ions that automatic gain control will produce during ionization. Target TIC values were found to be analyte dependent (Table III). Target TIC values above those needed for a particular compound caused loss of mass resolution. As the target TIC value was increased, the ionization time increased, thus more ions 599

6 Journal of Analytical Toxicology, Vo[. 24, October 2000 were formed in the trap and a greater chance of adverse space charging occurred. As analyte concentrations increased, the ionization time was decreased (by the AGC) to limit the ion population. The lower the analyte concentration, the less effect the target value has on the response of the analyte (Figure 5). Figure 5 shows the relative base peak response of 6-AM plotted against the target TIC values for the cutoff concentration (10 ng/ml) and 10 times the cutoff. At higher concentrations, using a lower target value decreased the ionization time and the base peak response. Target value optimization was necessary to produce acceptable linearity and quantitation using SIS. The higher ion ratio failure rate (Table IV) of the ion trap was, in part, a consequence of the limited volume bounded by the trap electrodes. As ion populations were increased with rising sample concentrations, space charging occurred. For this reason, successful ion trap experiments were more closely tied to diligent sample preparation than that needed with quadrupole detection. Because the extraction procedures were developed for use with the MSD, ion trap performance could be improved by optimizing the extraction for ion trap analysis. Alternatively, an ion-trap instrument that employs external ionization prior to SIS could improve ion ratio stability by elimination of neutral interference (17). Conclusions The ion trap produced stable mean ion ratios over the entire concentration range but exhibited higher variance within sample batches. There was no difference in ion ratio precision between SIS and FS, except at the lowest concentrations monitored. Ion ratio failure rates within batches were highest in FS and lowest in SIM. The quantitative accuracy for selected ion storage ion-trap MS was comparable to selected ion monitoring MS. The ion trap demonstrated reproducible spectra across the concentration ranges of five SAMHSA drugs with sensitivities that were generally equivalent to quadrupole SIM results. However, ion ratios failure rates were higher for the ion trap. Diligent sample preparation and greater attention to chromatographic separation is necessary for a reproducible and accurate ion trap confirmation method. Also, attention to parameters such as target TIC, mass defect, and AGC are necessary to avoid the potential problems not experienced with quadrupole MS. Knowledge of the restriction regarding overloading the ion trap, and the requirements of ion manipulations are requisites for its successful inclusion as a confirmatory technique for forensic urine drug testing. 100% 75% 5O% 25% 700 6O( OO I00 o Ion-trap SIS ] m/z m/z Figure 4. Comparison of SIS and SIM spectra of gze-butyl. 90% 80% 70% 60% 50~ 40% S-AM 6-AM I0>./ / J Quadrupole SIM I BOO0 70QQ ~ % 94% 87% 91% 100% 72% 76% 55% 74% 70% 76% 100% 72% 72% Target value :igure 5. Effects of analyte concentration on the target and base peak responses il 4-6-AM 35O 9 "4-6-AM 10 concentration Acknowledgment This work was funded in part by the American Registry of Pathology, Washington, D.C References 1. D.W. Hoyt, R.E. Finnigan, T. Nee, T.F. Shults, and T.J. Butler. Drug testing in the workplace-- are methods legally defensible? J. Am. Med. Assoc. 258: (] 987). 2. A.H.B. Wu, D.W. Hill, D. Crouch, C.N. Hodnett, and H.H. McCurdy. Minimal standards for the performance and interpretation of toxicology test in legal proceedings. J. Forensic Sci. 44: (1999). 3. M. Lehrer. Application of gas chromatograph/mass spectrometry instrument techniques to forensic urine drug testing. Clin. Lab. Med. 10' (1990). 4. Department of Defense Instruction Number Technical procedures for the military personnel drug abuse testing program (1994). 5. Department of Health and Human Services. Mandatory guidelines for federal workplace drug testing programs. Substance Abuse and Mental Health Services Administration Publication 53 FR (1994). 6. W. Paul and H. Steinwedel. German Patent , 1956; U.S. Patent , M.F. Raymond, An introduction to quadrupole ion trap mass spectrometry. J. Mass Spectrom. 32: (1997). 600

7 8.. P.S.H. Wong and R.G. Cooks. Ion trap mass spectrometry. Cur. Sep. 16:85-92 (1997). 9. C.S. Crease, and J.W. Stygall. Recent developments in analytical ion trap mass spectrometry. Trends Anal. Chem. 17: (1998). 10. S.A. McLuckey and J.L. Stephenson, Jr. Ion/ion chemistry of highmass multiply charged ions. Mass. Spectrom. Rev. 17: (1998). 11. D.J. Berkeland, J.D. Miller, B.C. Young, J.C. Bergquist, W.M. Itano, and D.J. Wineland. High-accuracy frequency standards using lasercooled Hg + ions. SPIE conference proceedings: Methods for ultrasensitive detection (1998). 12. D.J. Botts and L.D. Bowers. Mass-to-charge ratio misassignments of molecular ions in ion-trap mass spectrometry. J. Anal. Toxicol. 22: (1998). 13. L.D. Bowers and D.J. Borts. Evaluation of selected-ion-storage ion-trap mass spectrometry for detecting urinary anabolic agents. Clin. Chem. 43: (1997). 14. F.A. Londry and R.E. March. Systematic factors affecting high mass resolution and accurate mass assignment in a quadrupole ion trap. Int. J. Mass Spectrom. Ion Processes. 144: (1995). 15. Clean screen extraction columns application manual. United Chemical Technologies, Inc. Revision number A.H.B. Wu, T.A. Onigbinde, S.S. Wong, and K.G. Johnson. Evaluations of full scanning GC/ion trap MS analysis of NIDA drugs of abuse testing in urine. J. Anal. Toxicol. 16" (1992). 17. S.C. McLuckey, G.J. Van Berkel, D.E. Goeringer, and G.L. Glish. Ion trap mass spectrometry of externally generated ions. Anal. Chem. 66: (1994). Manuscript received March 13, 2000; revision received May 23,