Validation of a Solild Phase Extraction Procedure for the GC-MS Identification and Quantitation of Cocaine and Three Metabolites in Blood, Urine, and Milk

J. LIQ. CHROM. & REL. TECHNOL., 24(3), 401–414 (2001) VALIDATION OF A SOLID PHASE EXTRACTION PROCEDURE FOR THE GC-MS IDENTIFICATION AND QUANTITATION OF COCAINE AND THREE METABOLITES IN BLOOD, URINE, AND MILK Imad K. Abukhalaf,1,2,∗ Bryan A. Parks,1 Natalia A. Silvestrov,1 Daniel A. von Deutsch,3 Ashraf Mozayani,3 and Hassan Y. Aboul-Enein4 1 2 Department of Pharmacology & Toxicology and the Clinical Research Center, Morehouse School of Medicine, Atlanta, GA 30310, USA 3 Joseph A. Jachymczyk Forensic Center, Houston, TX 77033, USA 4 Pharmaceutical Analysis Laboratory, Biological and Medical Research Department (MBC-03) King Faisal Specialist Hospital and Research Centre, Riyadh, Saudi Arabia ABSTRACT A simple and widely used solid-phase extraction procedure (United Chemical Technologies Method Handbook) was applied for the GC-MS identification and quantitation of cocaine (COC), benzoylecgonine (BE), cocaethylene (COCE), and m-hydroxybenzoylecgonine (HBE) in blood, urine, and milk. The method 401 C 2001 by Marcel Dekker, Inc. Copyright  www.dekker.com 402 ABUKHALAF ET AL. which utilizes BSTFA as a derivatizing agent yielded abundant diagnostic ions with high m/z values. Linear quantitative response curves were generated for the analytes of interest over a concentration range of 5–1000 ng/mL. Linear regression analyses of the standard curve in the three specimen types exhibited correlation coefficients ranging from 0.997 to 1.000. The LOD values for COC, COCE, and derivatives of BE, and HBE in the three specimen types ranged from 2.5 to 5.0 ng/mL. The LOQ values, however, ranged from 5.0 to 10.0 ng/mL. Intra-assay and inter-assay precision studies reflected a high level of reliability and reproducibility of the method. The applicability of the method for the detection and quantitation of COC, BE, COCE, and HBE was demonstrated successfully in human blood and urine samples, as well as blood samples obtained from cocaine-treated (subcutaneously) rats. INTRODUCTION Cocaine (COC) is one of the most potent of the naturally-occurring central nervous system stimulants. It has been widely utilized in medicine as a local anesthetic (especially in opthalmological procedures) and increasingly by drug abusers for its stimulant properties. During the 1980s and 90s, cocaine abuse increased to epidemic proportions across the United States, and continues to be a major public concern. Cocaine is rapidly metabolized in man by the hydrolysis of one or both of its ester linkages. At slightly basic pH, the drug is readily hydrolyzed to its primary metabolite benzoylecgonine (BE) (1–6). BE, in turn, is converted to ecgonine (E) by enzymatic hydrolysis, to m-hydroxybenzoylecgonine (HBE) by hydroxylation, and to benzoylnorecgonine (BNE) by N-methylation (1–6). In blood, cocaine is also hydrolyzed to ecgonine methyl ester (EME) by cholinesterase (1–6). Cocaethylene (COCE) is a neurologically active compound that provides the same degree of euphoria as COC, but for longer periods of time (7). As reported by Hearn et al. (8) and others (9,10), COCE is a unique metabolite that occurs in blood as a result of concurrent use of cocaine and ethanol. Varying amounts of other metabolic products of cocaine, such as the active metabolite norcocaine (NCOC), and meta- hydroxycocaine (HCOC) have also been detected (11). Because of the adverse health consequences and legal implications of cocaine abuse, it has become increasingly more important for analytical toxicologists to continue to improve upon the methods for the detection and quantitation of cocaine and its metabolites in biological specimens, to provide evidence of cocaine use COCAINE AND THREE METABOLITES 403 and abuse. A number of methods for the detection and quantitation of cocaine in blood, urine, meconium, and hair, are documented in the literature. These methods range from thin layer chromatography (TLC) to high performance liquid chromatography (HPLC) and gas chromatography-mass spectrometry (GC-MS) (12–19). The Gas Chromatograph-Mass Spectrometer is readily available in most pharmaceutical and toxicology laboratories, and its main advantage over other instrumentation used in the laboratories, is that it provides appropriate sensitivity, specificity, and selectivity for the analytes of interest. This manuscript describes the validation and applicability of a simple and widely used solid-phase extraction procedure (20) for the extraction and GC-MS quantitation of COC, BE, HBE, and COCE in blood, urine, and milk, utilizing only one milliliter of sample. The applicability of the method to quantitate COC, BE, HBE, and COCE was demonstrated successfully in human whole blood specimens, human urine samples, as well as blood samples obtained from rats treated with cocaine subcutaneously. To our knowledge, this manuscript is the first to describe method parameters and validation for testing COC, BE, HBE, and COCE in milk specimens. EXPERIMENTAL Materials The COC, BE, HBE, and COCE standard materials and deuterated analogs were purchased from Radian Corporation (Austin, TX). Clean Screen extraction columns were purchased from World Wide Monitoring (Bristol, PA). bis (Trimethylsilyl) trifluoroacetamide, containing 1% trimethylchlorosilane, was purchased from Pierce Chemical Company (Rockford, IL). Monobasic and dibasic potassium phosphate, hydrochloric acid, methanol, dichloromethane, isopropanol, ammonium hydroxide, and ethyl acetate were purchased from Fisher Scientific (Suwanee, GA) and were of analytical grade or the highest purity available. Autosampler vials (12 × 32 mm; clear crimp-top) with 100 µL limited volume inserts were purchased from Alltech Associates Inc. (Deerfield, IL). Water used in this study met Type II water criteria (21) and was filtered through Nanopure System (Barnstead, Dubuque, IA). Methods Extraction and Derivatization Extraction and derivatization procedures were conducted essentially as described (20). Samples (1 mL) were spiked with 50 ng of deuterated (D3 ) internal 404 ABUKHALAF ET AL. standards, corresponding to the analytes of interest. Then, 4 mL of deionized water were added, and blood samples were centrifuged at 2000 rpm for 10 min. Pellets were discarded and 2 mL of 100 mM potassium phosphate buffer was added to all specimen types and the pH was adjusted to 6–6.5. The samples were then loaded on the extraction columns which had been pre-equilibrated by sequential treatment with 2 mL of methanol, 2 mL of deionized water, and 3 mL of 100 mM potassium phosphate buffer. After the samples passed through the bed of the columns, the columns were sequentially washed with 2 mL deionized water, 2 mL of 100 mM HCl, and 3 mL of methanol. COC, BE, HBE, and COCE were eluted with 3 mL of freshly prepared elution mixture composed of dichloromethane, isopropyl alcohol, and ammonium hydroxide (78/20/2,v/v/v). Eluates were dried at room temperature under a continuous stream of nitrogen. To ensure complete removal of water, 0.5 mL dichloromethane was added to the residue and evaporated to dryness. The analyte mixture was derivatized by the addition of 50 µL ethyl acetate and 50 µL BSTFA containing 1% TMCS. After 30 min incubation at 65◦ C, samples (2 µL) were injected onto the autosampler of an HP 5890 gas chromatograph (GC) equipped with a 5972 mass selective detector (MS) [Hewlett Packard (currently Agilent Technologies), Palo Alto, CA]. For the construction of standard curves, blank blood, urine, and milk samples were spiked with 50 ng internal standards and increasing concentrations of analytes (5, 50, 100, 500, and 1000 ng/mL). GC-MS Run Conditions Samples (2 µL) were injected from the autosample vials into the GC-MS. The analytes were resolved with an HP 1MS capillary column crosslinked with 1% phenylmethylsilicone (15 m × 0.25 mm with 0.25··m film thickness) using inlet pressure programming. The electron multiplier was operated at 200 V above the tune value. The carrier gas used was ultra high purity helium (99.99%). Splitless injection was used, and the splitless valve remained closed for 1 min and the initial inlet pressure was 25 psi. The column head pressure was held for 0.5 min, then decreased to 16 psi at a rate of 25 psi/min, and, finally, maintained at this pressure for the duration of the analysis time. This resulted in a flow rate of 1.1 mL/min during the run. Injector temperature was 280◦ C. The initial oven temperature was 150◦ C and was held for 1 min; the temperature was programmed to increase at 15◦ C/min to 215◦ C, and then increased to 300◦ C at 35◦ C/min, and held there for 2 min. The transfer line temperature was held at 280◦ C. Selected ion monitoring (SIM) mode was used in all runs. The entire run time was less than 8.0 min. COCAINE AND THREE METABOLITES 405 Selection of Ions The quantitation and qualifier ions for this study were selected by examining the full scan mass spectra after injecting approximately 100 ng of each analyte and its deuterated analog or their TMS derivatives, as appropriate, into the HP 5970 MSD. The mass spectra of the analytes and their deuterated analogs were compared and ion pairs were selected. Selections were based on the following criteria with decreasing order of importance: (A) The corresponding ion from the analog must not occur in the analyte, and the corresponding ion from the analyte must not occur in the analog. (B) The most abundant ions were selected. (C) The largest m/z ions were selected. Once the most abundant ions were selected for each analyte, their exact masses were determined periodically to ensure that the target ions remained the same. RESULTS AND DISCUSSION Solid-phase extraction has several advantages over liquid-liquid extraction. These advantages are: decreased solvent volumes, resulting in decreased solvent disposal costs; reduced operator time; high percent recovery of analytes; and low limits of detection and quantitation for the analytes (22). Additionally, solid-phase extraction yields clean extracts and minimizes the appearance of endogenous peaks, thereby increasing analyte selectivity (23). The major ion peaks selected are shown in Table 1, and they are in agreement with previous reports (2,20). As shown in Figure 1, COC, BE, HBE, and COCE are easily separated under the above-described GC-MS conditions. The retention times for the four analytes ranged from 5.74 to 7.12 min (Table 1). The calibration curves obtained for each analyte were constructed using linear regression analysis (24). For COC, BE, and COCE in blood, urine, and milk, linearity was achieved Table 1. Retention Times and Quantitative and Qualitative Ions Monitored for COC, COCE, and Derivatized BE and HBE Analyte COC TMS-BE TMS-HBE COCE Retention Time (min) Quantitative Ion (m/z) Qualitative Ion (m/z) Qualitative Ion (m/z) 5.74 6.08 7.12 5.99 182 240 240 196 198 256 210 82 303 361 449 317 406 ABUKHALAF ET AL. Figure 1. GC-MS Total Ion Chromatogram of 500 ng/mL COC, BE, HBE, and COCE. over a concentration range of 5 to 1000 ng/mL, whereas linearity for HBE in all specimen types (blood, urine, and milk) ranged from 10 to 1000 ng/mL. The data, upon the analysis of the corresponding regression lines (24), exhibited correlation coefficients ranging from 0.999 to 1.0 for blood, 0.998 to 0.999 for urine, and 0.997 to 0.999 for milk (Table 2). The formulas of the lines for each analyte in the three specimen types are also shown in Table 2. The limits of detection (LOD) and limits of quantitation (LOQ) for each analyte in each specimen type were determined experimentally (N = 5). COCAINE AND THREE METABOLITES 407 Table 2. Analysis of the Regression Line for Analytes in Blood, Urine, and Milk Analyte COC Category Blood Urine Milk RR∗ 1.70 × AR¶ + 4.86 × 10−2 0.999 5–1000 0.954 × AR + 1.39 × 10−2 0.999 5–1000 1.12 × AR + 8.09 × 10−3 0.999 5–1000 0.85 × AR + 8.69 × 10−2 0.999 5–1000 1.08 × AR + 5.01 × 10−2 0.998 5–1000 1.04 × AR + 6.23 × 10−2 0.999 5–1000 1.70 × AR + 4.9 × 10−2 1.000 10–1000 2.62 × AR + 9 × 10−2 0.999 10–1000 2.41 × AR + 9.8 × 10−1 0.997 10–1000 0.896 × AR + 3.1 × 10−2 1.000 5–1000 1.08 × AR + 1.97 × 10−2 0.999 5–1000 8.0 × AR + 1.39 × 10−2 0.999 5–1000 r2∗∗ Linearity (ng/mL) TMS-BE RR r2 Linearity (ng/mL) TMS-HBE RR r2 Linearity (ng/mL) COCE RR r2 Linearity (ng/mL) ∗ RR = Response Ratio; Area of Analyte/Area of Internal Standard. AR = Amount Ratio; Concentration of Analyte/Concentration of Internal Standard. ∗∗ 2 r = correlation coefficient. ¶ For LOQ, blank specimens were spiked with a series of decreasing concentrations of analytes and a constant concentration (50 ng/mL) of their corresponding internal standards. LOD was defined as the concentration corresponding to a signal to noise ratio of 3. LOQ was defined as the lowest quantitated concentration that was within 20% of the target concentration. As shown in Table 3, the LOD values obtained for COC, BE, HBE, and COCE extracted from blood were 2.5, 2.5, 4.0, and 2.5 ng/mL, respectively. These values were comparable to those obtained for the same analytes extracted from urine or milk. The LOQ values obtained for COC, BE, HBE, and COCE extracted from blood were 5.0, 5.0, 8.0, and 5.0 ng/mL, respectively. Similarly, LOQ values obtained for COC, BE, HBE, and COCE extracted from urine or milk ranged from 5.0 to 10 ng/mL (Table 3). Table 4 illustrates the extraction efficiency of various analytes at three concentrations (10, 500, and 1000 ng/mL) in blood, urine, and milk. These three concentration points represent the low, middle, and high portions of the standard 408 ABUKHALAF ET AL. Table 3. Limits of Detection and Quantitation, and Percent Recovery of Analytes in Blood, Urine, and Milk Blood∗∗ Analyte COC TMS-BE TMS-HBE COCE Urine∗∗ Milk∗∗ LOD* LOQ¶ LOD LOQ LOD LOQ 2.5 2.5 4.0 2.5 5.0 5.0 8.0 5.0 2.5 2.5 5.0 2.5 5.0 5.0 10.0 5.0 2.5 2.5 5.0 2.5 5.0 5.0 10.0 5.0 ∗∗ n = 5. LOD = Limit of Detection, defined as the concentration of analyte at which mean signal to noise ratio is 3:1. ¶ LOQ = Limit of Quantitation, defined as the lowest standard that was within 20% of the target concentration concentration of the analyte. ∗ curve. With the exception of HBE, extraction efficiencies (recoveries) of all analytes from the three biological matrices (blood, urine, and milk) were comparable (Table 4). At a low concentration (10 ng/mL), percent recoveries for COC, BE, and COCE ranged from 90.9 to 94.1%, whereas recoveries for HBE in the three matrices ranged from 80.4 to 86.6% (Table 4). At 500 ng/mL, the highest percent recovery achieved was for COC (98.2%), and the lowest recovery was for HBE (81.2%). At a concentration of 1000 ng/mL, recoveries of COC, BE, and COCE ranged from 92.7 to 97.8%, whereas recoveries for HBE ranged from 82.6 to 87.8% (Table 4). Generally, recoveries of HBE from blood, urine, and milk at the three concentration levels (10, 500, and 1000 ng/mL) were less than those observed with the other analytes (Table 4). This suggests that HBE either did not extract as efficiently as the other analytes, or underwent stereochemical configuration with Table 4. Percent Recoveries of Analytes in at Three Concentrations (ng/mL) in Blood, Urine, and Milk Blood∗ Analyte COC TMS-BE TMS-HBE COCE ∗ Milk∗ 10 500 1000 10 500 1000 10 500 1000 92.8 94.1 82.8 92.1 98.2 97.6 83.5 94.0 96.6 97.8 87.8 94.8 91.9 93.5 80.4 93.3 96.1 95.6 81.2 97.1 97.2 96.1 82.6 94.6 90.9 93.3 86.6 91.9 95.8 NA¶ 87.9 93.2 96.4 93.5 84.8 92.7 n = 5. NA = Not Analyzed. ¶ Urine∗ COCAINE AND THREE METABOLITES 409 a different spatial configuration. Whether HBE was converted from the metahydroxy configuration to para-hydroxy configuration, or not, is currently under intensive investigation in our laboratory. Preliminary data (not shown) suggest that conversion of the meta-hydroxy configuration of HBE to para-hydroxy configuration (para-HBE) did not occur during the extraction or the GC-MS run conditions for there was no evidence for the occurrence of the m/z 82, usually abundant in the para-HBE mass spectrum (2). Intra- and inter-assay precision of the analytical procedure in blood, urine, and milk, as represented by percent correlation of variance (% C.V.), is illustrated in Table 5 (A,B,C). Precision was determined experimentally (n = 5) by spiking negative blood, urine, and milk samples with COC, BE, HBE, and COCE at concentrations of 10, 500, and 1000 ng/mL. For all the analytes, at concentrations of 500 and 1000 ng/mL, intra-assay precision (% C.V.) values were less than 10 (Table 5 A,B,C). At a concentration near the limit of quantitation (10 ng/mL), % C.V. values reflected a greater variation of precision for COC, BE, HBE, and COCE which ranged between 2.3 to 14.1 (Table 5 A,B,C). Inter-assay precision of the method was determined experimentally in a manner similar to that of intra-assay precision. Spiked blood, urine, and milk samples were analyzed on a daily basis for two weeks. With the exception of HBE, inter-assay precision ranged from 2.6 to 12. Table 5. Intra- and Inter-Assay Precision for Cocaine (COC), Cocaethylene (COCE), Derivatized Benzoylecgonine (TMS-BE), and Derivatized m-Hydroxybenzoylecgonine (TMS-HBE) in (A) Blood, (B) Urine, and (C) Milk Target Conc. (ng/mL) A. Blood 10 Analyte COC TMS-BE TMS-HBE COCE Intra-Assay (mean conc)* 9.3 9.4 8.8 9.6 Precision (% C.V.) 2.3 2.8 14.1 7.4 Inter-Assay (mean conc.)* 9.1 9.2 7.9 9.3 Precision (% C.V.) 5.5 4.3 18.3 8.7 500 COC TMS-BE TMS-HBE COCE 491 488 416 470 2.2 2.1 9.8 4.3 472 465 381 411 3.1 2.6 11.6 6.7 1000 COC TMS-BE TMS-HBE COCE 966 978 878 948 3.4 5.2 9.9 4.6 921 908 816 919 7.4 8.8 12.9 7.1 (continued) 410 ABUKHALAF ET AL. Table 5. Continued Target Conc. (ng/mL) Analyte Intra-Assay (mean conc)* Precision (% C.V.) Inter-Assay (mean conc.)* Precision (% C.V.) COC TMS-BE TMS-HBE COCE 9.2 NA¶ 8.0 9.3 4.3 — 12.3 6.7 9.1 NA¶ 7.6 8.9 8.9 — 14.9 12.0 500 COC TMS-BE TMS-HBE COCE 481 478 406 NA¶ 4.1 4.9 8.8 — 448 449 373 NA¶ 7.9 8.3 13.1 — 1000 COC TMS-BE TMS-HBE COCE 972 961 826 946 4.2 4.7 9.3 5.2 921 908 781 889 9.9 9.8 12.5 6.4 B. Urine 10 C. Milk 10 COC TMS-BE TMS-HBE COCE 8.7 9.3 9.0 9.5 5.1 4.7 12.4 6.1 9.0 9.1 7.9 8.6 8.9 7.6 15.7 9.8 500 COC TMS-BE TMS-HBE COCE 479 481 440 466 5.3 6.1 9.0 4.5 427 428 383 461 8.4 9.3 10.8 9.7 1000 COC TMS-BE TMS-HBE COCE 964 935 848 927 5.1 5.2 9.7 6.0 NA¶ 931 801 902 — 7.8 11.2 10.0 ∗ n = 5. NA = Not Analyzed. ¶ The coefficients of variation (% CV) for HBE in all specimen types were higher than those obtained for COC, BE, and COCE (Table 5). This may be due to the lower abundance of the quantitation ion of TMS-HBE and the relatively lower recovery. The higher % C.V. values observed with the 10.0 ng/mL samples are because this concentration is near the limit of quantitation where greater variation should be expected. The applicability of this validated method was tested on human and rat blood samples, as well as on human urine samples obtained from alcohol and/or COCAINE AND THREE METABOLITES 411 Table 6. Analysis of Rat Blood for COC, BE, HBE, and COCE Specimen I.D. Analyte Concentration (ng/mL) 1 COCE BE HBE COC <LOQ∗ 0.00 0.00 34.8 2 COCE BE HBE COC 3 COCE BE HBE COC <LOQ∗ 0.00 0.00 31 4 COCE BE HBE COC <LOQ∗ 0.00 0.00 20.0 ∗ 11.4 113.6 0.00 28 <LOQ = the concentration was less than the limit of quantitation. cocaine-abusing individuals. Unfortunately, we were unable to obtain human milk samples from cocaine-abusing nursing mothers. As shown in Table 6, four control rat blood samples obtained from rats not treated with cocaine did not reveal any presence of cocaine analytes. On the other hand, analysis of samples obtained from rats treated with cocaine (10 µg/kg) subcutaneously revealed the presence of COC and BE. Because the rats were not administered alcohol concomitantly with cocaine, cocaethylene, a metabolite produced as a result of a concomitant administration of cocaine and alcohol (7), was not detected. HBE was detected in rat blood samples spiked with the compound. The reason for the absence of HBE in samples obtained from cocaine-treated rats may have been due to the fact that blood samples were drawn one week (longer than the half-lives of most cocaine metabolites including HBE) after cocaine treatment. Another possibility is the possible inability of rats to form HBE. Comprehensive literature searches conducted at the time of submitting this manuscript failed to produce any literature on the presence of HBE in rats. As shown in Table 7, all five human urine samples tested contained all four cocaine analytes in various concentrations. The corresponding blood samples, however, contained BE and traces of HBE. This was not surprising for COC’s half-life is significantly shorter than BE. 412 ABUKHALAF ET AL. Table 7. Analysis of Human Blood and Urine Samples for COC, BE, HBE, and COCE Specimen I.D. Analyte Blood Concentration (ng/mL) Urine Concentration (ng/mL) COCE BE OH-BE COC COCE 18.4 921 218 1930 49 2 BE OH-BE COC COCE 714 78 557 171 998 13 719 21 3 BE OH-BE COC COCE 391 30 57 7 433 2.9∗ 44 13 4 BE OH-BE COC COCE 914 111 257 34 2013 17 57 66 5 BE OH-BE COC 502 49 53 871 152 128 1 ∗ 11.1 801 0.4∗ 1272 210 = less than the limit of quantitation. Oyler et al. (2) have shown that both the p- and m-HBE are excreted in adult urine. The presence of HBE in adult urine and blood (trace amounts) indicates that our data further supports Oyler et al. (2) findings that p- and m-HBE are not unique fetal metabolites of COC as initially reported by other investigators (25). In conclusion, a practical and reliable analytical method for the detection and quantitation of COC, BE, HBE, and COCE is validated. The applicability of the method to extract COC, BE, HBE, and COCE from blood, urine, and milk specimens was demonstrated successfully. The method can be utilized for the detection of these cocaine analytes in a research setting, as well as in forensic drug testing laboratories. ACKNOWLEDGMENTS The authors wish to thank Dr. Mohammad Bayorh for providing rat blood samples, and Dr. Robert Oster for conducting the statistical analyses. This work COCAINE AND THREE METABOLITES 413 was partially supported by a NIH grant P20 RR11104-05 and by a Morehouse School of Medicine Faculty Development award. Bryan Parks current address is the Department of Chemistry, University of Georgia, Athens, GA 30602. REFERENCES 1. Jufer, R.A.; Walsh, S.L.; Cone, E.J. J. Anal. Toxicol. 1998, 22, 435–444. 2. Oyler, J.; Darwin, W.D.; Preston, K.L.; Suess, P.; Cone, E.J. J. Anal. Toxicol. 1996, 20, 453–462. 3. Baselt, R.C.; Cravey, R.H. Cocaine. In Disposition of Toxic Drugs and Chemicals in Man; Chemical Toxicology Institute: Foster City, 1995; 186–190. 4. Jatlow, P. Yale J. Biol. Med. 1988, 61, 105–113. 5. Stewart, D.J.; Inaba, T.; Lucassen, M.; Kalow, W. Clin. Pharmacol. Ther. 1979, 25, 464–468. 6. Zhang, J.Y.; Foltz, R.L. J. Anal. Toxicol. 1990, 14, 201–205. 7. Hearn, W.L.; Rose, S.; Wagner, J.; Ciarleglio, A.; Mash, S. Pharm. Biochem. Behav. 1991, 39, 531–533. 8. Hearn, W.L.; Flynn, D.D.; Hime, G.W.; Rose, S.; Cofino, J.C.; ManteroAtienza, E.; Wetli, C.V.; Mash, D.C. J. Neurochem. 1991, 56, 698–701. 9. Jatlow, P.; Elsworth, J.D.; Bradberry, C.W.; Winger, G.; Taylor, J.R.; Russell, R.; Roth, R.H. Life Sci. 1991, 48, 1787–1794. 10. Boyer, C.S.; Petersen, D.R. J. Pharmacol. Exp. Ther. 1992, 260, 939–946. 11. Steele, B.W.; Bandstra, E.S.; Wu, N.-C.; Hime, G.W.; Hearn, W.L. J. Anal. Toxicol. 1993, 17, 353–358. 12. Bailey, D.N. Am. J. Clin. Pathol. 1994, 101, 342–345. 13. Kogan, M.J.; Pierson, D.J.; Durkin, M.M.; Wilson, N.J. J. Chromatogr. 1989, 490, 236–242. 14. Jenkins, A.J.; Goldberger, B.A. J. Forensic Sci. 1997, 42, 824–827. 15. Ferrara, S.D.; Tedeschi, L.; Frison, G.; Brusini, G.; Castagna, F.; Bernardelli, B.; Soregaroli, D. J. Anal. Toxicol. 1994, 18, 278–291. 16. Glass, R.L.; Johnson, E.L. J. Liquid Chromatogr. 1993, 16, 3543–3555. 17. Tagliaro, F.; Valentini, R.; Manetto, G.; Crivellente, F.; Carli, G.; Marigo, M. Forensic Sci. Int. 2000, 107, 121–128. 18. Girod, C.; Staub, C. Forensic Sci. Int. 2000, 107, 261–271. 19. Abusada, G.M.; Abukhalaf, I.K.; Alford, D.D.; Vinzon-Bautista, I.; Pramanik, A.K.; Ansari, N.A.; Manno, J.E.; Manno, B.R. J. Anal. Toxicol. 1993, 17, 353–358. 20. Clean Screen Extraction-Columns. Applications Manual. Worldwide Monitoring: Bristol, PA, 1997. 21. Preparation and Testing of Reagent Water in the Clinical Laboratory, 2nd Ed.; National Committee for Laboratory Standards, NCCLS Document C3-A2: Villanova, PA, 1991; Vol. 11, No. 13. 414 ABUKHALAF ET AL. 22. Pocci, R.; Dixit, V.; Dixit, V.M. J. Anal. Toxicol. 1992, 16, 45–47. 23. Lillsunde, P.; Korte, T. J. Anal. Toxicol. 1991, 17, 71–81. 24. Tallarida, R.J.; Murray, R.B. Manual of Pharmacologic Calculations with Computer Programs, 2nd Ed.; Springer-Verlag: New York, NY, 1987. 25. Steele, B.W.; Bandstra, E.S.; Wu, N.-C.; Hime, G.W.; Hearn, W.L. J. Anal. Toxicol. 1993, 17, 348–352. Received August 2, 2000 Accepted August 30, 2000 Manuscript 5359 Copyright of Journal of Liquid Chromatography & Related Technologies is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.