Online solid-phase extraction liquid chromatography–electrospray-tandem mass spectrometry analysis of buprenorphine and three metabolites in human urine

Available online at www.sciencedirect.com

Talanta 75 (2008) 198–204

Online solid-phase extraction liquid chromatography–electrospray-tandem mass spectrometry analysis of buprenorphine and three metabolites in human urine An-Chan Liu a , Tzuen-Yeuan Lin a , Lien-Wen Su b , Ming-Ren Fuh a,∗ a

b

Department of Chemistry, Soochow University, 111 Shihlin, Taipei, Taiwan Taipei City Psychiatric Center, Department of Addiction Science, Taipei City Hospital, Taipei, Taiwan Received 6 September 2007; received in revised form 30 October 2007; accepted 30 October 2007 Available online 20 February 2008

Abstract An online solid-phase extraction (SPE) liquid chromatography–electrospray tandem mass spectrometry (LC–ESI-MS/MS) method for the determination of buprenorphine (Bup), norbuprenorphine (nBup), buprenorphie-3-␤-d-glucuronide (Bup-3-G) and norbuprenorphie-3-␤-d-glucuronide (nBup-3-G) in human urine was developed and validated. A mixed mode SPE column with both hydrophilic and lipophilic functions was used for online extraction. A C18 column was employed for LC separation and ESI-MS/MS was utilized for detection. Buprenorphine-D4 (Bup-D4 ) and norbuprenophine-D3 (nBup-D3 ) were used as internal standards for quantitative determination. The extraction, clean-up and analysis procedures were controlled by a fully automated six-port switch valve. Identification and quantification were based on the following transitions: m/z 468→414 for Bup, m/z 414→364 for nBup, m/z 644→468 for Bup-3-G and m/z 590→414 for nBup-3-G, respectively. Good recoveries from 93.6% to 102.2% were measured and satisfactory linear ranges for these analytical compounds were determined. Minimal ion suppression effect (∼7% response decrease) was determined. Intra-day and inter-day precision showed coefficients of variance, CV, ranged from 3.3% to 10.1% and 4.4% to 9.8%, respectively. Accuracy ranging from 97.0% to 104.0% was determined. The applicability of this newly developed method was demonstrated by analyzing human urine samples from the patients in Bup treatment program for therapeutic monitoring purpose. © 2007 Elsevier B.V. All rights reserved. Keywords: Buprenorphine; Urine analysis; Online solid-phase extraction; Liquid chromatography–electrospray-tandem mass spectrometry

1. Introduction Buprenorphine (Bup) is a synthetic opioid drug for the treatment of chronic pain. Bup is more potent than morphine and has been used for the treatment of heroin addiction [1,2]. In human, Bup is N-dealkylated to norbuprenorphine (nBup); furthermore, both Bup and nBup undergo extensive conjugation to glucuronides, buprenorphie-3-␤-d-glucuronide (Bup-3-G) and norbuprenorphie-3-␤-d-glucuronide (nBup-3-G), prior to urine excretion [3,4]. The structures of these chemicals are shown in Fig. 1. It has been reported the concentration of free Bup in urine is in low nanogram ranges and the glucuronides are

∗ Corresponding author at: Department of Chemistry, Soochow University, P.O. Box 86-72, Taipei, Taiwan. Tel.: +886 2 2881 9471x6821; fax: +886 2 2881 1053. E-mail address: [email protected] (M.-R. Fuh).

0039-9140/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.talanta.2007.10.050

the major metabolites found; in addition, extensive difference of inter-individual Bup metabolism in human was observed [5]. Therefore, accurate monitoring of Bup and its metabolites in biological fluids of patient is needed to effectively personalize the treatment program. The determination of drug and its metabolites in plasma provides the real indication of active amount of a drug; however, monitoring the concentrations of active component in urine is an useful complement to plasma measurement for therapeutic drug monitoring purpose. Several analytical methods have been developed for the analysis of Bup and nBup in biological matrices using gas chromatography–mass spectrometry (GC–MS) [6–8], liquid chromatography (LC) with electrochemical detector [9], UV detector [10], electrospray mass spectrometry (ESI-MS) [11–13], or atmospheric pressure chemical ionization mass spectrometry (APCI-MS) [14]. However, there are only few reported methods for the quantitative determination of Bup, nBup, Bup-3-G and nBup-3-G. Two reported methods

A.-C. Liu et al. / Talanta 75 (2008) 198–204

199

Fig. 1. Structure of analytical component.

utilized LC–ESI-tandem MS to quantify Bup, nBup and their glucuronides in plasma [15,16]. Liquid-liquid extraction combined with solid-phase extraction was employed as sample pre-treatment processes for these studies. The limit of detection of buprenophine and its metabolites were 0.1–0.3 ng/ml. In addition, LC–tandem MS method has been reported to approximate the concentrations of Bup-3-G and nBup-3-G

in human urine based on the quantification of Bup and nBup before and after hydrolysis [17]. In this investigation, authors estimated the cutoff of LC-ES assay was near 1 ng/ml. For LC–ESI-MS analysis of biological samples, ion suppression effect is frequently observed [18–20]. Solid-phase extraction (SPE) has been proven to be an effective sample pre-treatment technique to minimize ion suppression effect.

200

A.-C. Liu et al. / Talanta 75 (2008) 198–204

However, SPE generally requires laborious and time-consuming equilibrium, clean-up and evaporation procedures. Online SPE has been successfully applied to LC–ESI-MS analysis for the analysis of various drugs in biological fluid [21–23]. This paper describes the development and validation of an online SPE LC–ESI-MS/MS method to quantify Bup, nBup, Bup-3-G and nBup-3-G in human urine for therapeutic application. This fully automated method required little minimum sample pre-treatment and was applicable to high through-put analysis. In addition, ion suppression effect was examined and the short-term stability of analyte was evaluated. The applicability of this newly developed method to monitor the urine samples of Bup treatment patients was demonstrated. To our knowledge, this is the first study to direct quantitative measurement of Bup and three metabolites (nBup, Bup-3-G and nBup-3-G) in human urine. 2. Experimental 2.1. Chemicals and reagents Bup, nBup, Bup-3-G, nBup-3-G, Bup-D4 and nBup-D3 were purchased from Cerilliant Corp. (Austin, TX, USA). HPLCgrade acetonitrile and methanol were from Milinckrodt Baker, Paris, KY, USA. Reagent-grade trifluoroacetic acid (TFA) was obtained from Riedel-de Haen AG, Germany. Purified water from a Milli-Q system from Millipore Corp. (Beford, MA, USA) was used. Drug-free urine samples collected from five healthy volunteers were used for method development and preparation of calibration standards. 2.2. Equipments An Agilent 1100 LC system (Agilent Co., Palo Alto, CA, USA), consisting of a quaternary pump, an online degasser, an auto-sampler and a six-port switch valve was used. A Symmetry Shield RP18 column (2.1 mm × 50 mm, 3.5 ␮m; Waters Corp., Milford, MA, USA) was utilized for LC separation and an Oasis HLB column (2.1 mm × 20 mm, 3.5 ␮m) from Waters Corp. was employed for online SPE. All mass spectrometric measurements were performed on an Agilent LC/MSD SL ion trap mass spectrometer with an electrospray ionization source operating in positive ion mode. Agilent 1100 series LC/MSD Trap software (version 4.0) was utilized for system control, data acquisition and data analysis. The spray voltage was set at −3.5 kV and compressed nitrogen gas (50 psi) was used for nebulization. Heated nitrogen (350 ◦ C, 10 L/min) was for solvent evaporation. The data acquisition was performed under the following conditions: normal scan speed, molecular mass scan range 100–700, ion charge control target 30,000, maximum accumulation time 300 ms and the isolation width for precursor ions was 2. 2.3. Standard and sample preparation Individual stock solutions (100 ␮g/mL) of each compound were prepared in methanol and stored in a refrigerator (4 ◦ C)

Fig. 2. Configuration of online SPE LC–ESI-tandem MS: (a) extraction, (b) analysis and (c) clean-up.

when not in use. Spiked urine standards were prepared by spiking appropriate amounts of each analytical component into drug-free urine. Prior to online SPE LC–tandem MS analysis, each urine sample (2 mL) was diluted with 2 mL of 0.01% TFA solution with internal standards (Bup-D4 and nBup-D3 , 5 ng/ml each) and centrifuged for 5 min at 1500 × g. A 50-␮L aliquot of supernatant was used for analysis. 2.4. Online sample preparation and LC separation The experimental setup for online SPE LC–ESI-tandem MS is shown in Fig. 2 and the time program for extraction, washing and analysis is listed in Table 1. First, the SPE extraction column was equilibrated by 0.05% TFA aqueous solution for 2 min at 1 mL/min with the system in “extraction” configuration as shown in Fig. 2a. Urine sample was introduced by an auto-sampler and transported by 0.05% TFA aqueous solution at 1 mL/min. While the endogenous components of urine were washed to waste, the analytical compounds were withheld in the SPE column. From 1.1 to 5.0 min, the extracted compounds in SPE column were further washed by methanol/water (5/95, v/v) with 0.05% TFA mixture and then by acetonitrile/water mixture as described in Table 1. At 5.1 min after sample injection, a gradient elution (at flow rate of 0.2 mL/min) was employed for chromatographic separation of the analytes. At the same time, the system was turned to “analysis” configuration as shown in Fig. 2b and the eluted components were delivered to ESI-tandem MS for determination. At 12.1 min, the system was converted to “clean-up” position as shown in Fig. 2c. The system was washed

A.-C. Liu et al. / Talanta 75 (2008) 198–204 Table 1 Time program of extraction and LC separation (see Section 2 for detail) Time (min)

0.0 1.0 1.1 4.0 4.1 5.0 5.1e 11.0 12.0 12.1f 16.0 16.1g 19.0 a b c d e f g

201

bor, WA, USA) and a syringe pump (Harvard Apparatus, South Natick, MA, USA).

Solvent A (%)a

B (%)b

C (%)c

D (%)d

Flow rate (ml/min)

100 100 0 0 0 0 0 0 0 0 0 100 100

0 0 100 100 0 0 0 0 0 0 0 0 0

0 0 0 0 70 70 30 0 0 70 70 0 0

0 0 0 0 30 30 70 100 100 30 30 0 0

1.0 1.0 1.0 1.0 1.0 1.0 0.2 0.2 0.2 0.4 0.4 1.0 1.0

3. Results and discussion 3.1. ESI-MS/MS

0.05% TFA aqueous solution. Methanol/water (5/95, v/v) with 0.05% TFA. Acetonitrile/water (5/95, v/v) with 0.05% TFA. Acetonitrile/water (95/5, v/v) with 0.05% TFA. At 5.1 min, the system is turned to “analysis” configuration. At 12.1 min, the system is turned to “clean-up” configuration. At 16.1 min, the system is turned to “extraction” configuration.

with acetonitrile/water (95/5, v/v) with 0.05% TFA mixture and then equilibrated by the initial mobile phase of gradient elution. At 16.1 min, the system was switched to “extraction” mode and the SPE column was re-equilibrated with 0.05% TFA aqueous solution for 3 min at flow rate of 1.0 mL/min. Afterwards, the system was ready for next injection. 2.5. Investigation of ion suppression effect Ion suppression effect was examined by a previously reported post-column infusion setup [18]. Drug-free urine blank from five volunteers and water were injected into online SPE LC–ESItandem MS system while a mixture of four analytes (100 ng/mL each) in water was continuously infused in parallel at 20 ␮L/min through a PEEK tee-junction (Upchurch Scientific, Oak Har-

In order to optimize the condition for ES-MS/MS determination, 100 ng/mL standard solutions of each analyte were directly infused ESI-ion trap MS. In ESI-MS analysis, proton adduct ion was detected as base ion for all analytes and internal standards; therefore, [M + H]+ of each compound was selected as the precursor ion for the subsequent ESI-MS/MS measurement. The MS/MS results and tentative assignments of MS/MS fragmented ions are summarized in Table 2. The MS/MS for Bup gives the fragment 414 (m/z) resulted from the loss of cyclopropylmethylene group of [M + H]+ ; further cleavage of t-butyl group (–C(CH3 )3 ) and methoxy group (–OCH3 ) produces fragment 326 (m/z). For nBup, the MS/MS fragment ions are similar to those of Bup since it has comparable structure of Bup but without cyclopropylmethy group. The tetrahydropyranyl sixmembered rings in both Bup-3-G and nBup-3-G can easily be cleaved to give fragments 468 (m/z) and 414 (m/z), respectively. MS/MS results for Bup-D4 and nBup-D3 furthermore supported the identification of fragment ions. The MS/MS fragmentation pathways of these compounds in ion trap have previously proposed [12,15]. The results of this study are similar to those of the previous investigation. 3.2. Online SPE LC–ESI-MS/MS method development A HLB SPE column which is packed with micro-porous material consisting of both hydrophilic and lipohpilic groups was employed as an extraction column. Loading, washing and separation procedures were a modification of previously reported method [21]. Diluted urine sample was introduced into system by 0.05% TFA solution. TFA was used as an ion paring reagent to form an ion pair with analytical molecule and then retained in the SPE column.

Table 2 ESI-MS/MS results Compound

Frag. energy (V)

MS/MS product iona

Bup

1.55

nBup

1.30

Bup-3-G

1.35

nBup-3-G

1.00

Bup-D4

1.55

nBup-D3

1.30

[M + H]+ (468), [M + H–H2 O]+ (450), [M + H–C3 H6 ]+ (426), [M + H–C4 H6 ]+ (414), [M + H–C4 H6 –H2 O]+ (396), [M + H–C4 H6 –C(CH3 )3 –OCH3 ]+ (326) [M + H]+ (414), [M + H–H2 O]+ (396), [M + H–CH3 OH]+ (382), [M + H–H2 O–CH3 OH]+ (364), [M + H–H2 O–C4 H8 ]+ (340), [M + H–C(CH3 )3 –OCH3 ]+ (326) [M + H]+ (644), [M + H–C6 H8 O6 ]+ (468), [M + H–C6 H8 O6 –C4 H6 ]+ (414), [M + H–C6 H8 O6 –C4 H6 –H2 O]+ (396) [M + H]+ (590), [M + H–H2 O]+ (572), [M + H–C6 H8 O6 ]+ (414), [M + H–C6 H8 O6 –H2 O]+ (396), [M + H–C6 H8 O6 –CH3 OH]+ (382), [M + H–C6 H8 O6 –CH3 OH–H2 O]+ (364), [M + H–C6 H8 O6 –C(CH3 )3 OH]+ (340) [M + H]+ (472), [M + H–H2 O]+ (454), [M + H–C3 H4 D2 ]+ (428), [M + H–C4 H3 D3 ]+ (415), [M + H–C4 H3 D3 –CH3 ]+ (400), [M + H–C3 H3 D2 –C(CH3 )3 ]+ (372), [M + H–C4 H2 D3 –C(CH3 )3 –OCH3 ]+ (328) [M + H]+ (417), [M + H–H2 O]+ (399), [M + H–CH3 OH]+ (382), [M + H–H2 O–CH3 OH]+ (364), [M + H–H2 O–C4 H8 ]+ (343), [M + H–C(CH3 )3 –OCD3 ]+ (326)

a

m/z of each ion in parenthesis; quantitative ions are in bold.

202

A.-C. Liu et al. / Talanta 75 (2008) 198–204 Table 3 Recoverya Bup 1.0 ng/mL 5.0 ng/mL 30.0 ng/mL 75.0 ng/mL a

98.0 96.6 93.6 97.6

nBup ± ± ± ±

2.9% 4.0% 2.5% 5.7%

95.2 94.7 99.1 96.0

± ± ± ±

Bup-3-G 3.1% 4.4% 3.1% 3.7%

97.1 98.4 102.2 98.6

± ± ± ±

nBup-3-G 4.7% 7.4% 3.8% 5.4%

100.7 99.7 94.9 97.7

± ± ± ±

4.0% 3.5% 4.5% 5.5%

Average ± standard deviation, n = 5.

ated. There was less than 7% decline in response of each analyte comparing to the responses of water sample. This indicated that online SPE LC method effectively removes and separates the potential interfering components in diluted urine sample. Fig. 3. Extracted ion chromatograms of a spiked human urine sample (10 ng/ml for each analyte, 5 ng/ml for each ISTD).

We examined various compositions of methanol/water (3/97, 5/95, 10/90) with 0.05% TFA mixtures as washing solutions to remove endogenous components retained on the SPE column. The results indicated that analytical compounds were easily washed out from SPE column when the methanol content is greater than 5/95. Therefore, methanol/water (5/95, v/v) with 0.05% TFA solution was selected as the washing solution. During the investigation of gradient elution, broadened eluted peaks of four analytes were observed and there was little chromatographic separation of analytical compounds when LC column coupled to SPE column. As the result, a second washing step using acetonitrile/water with 0.05%TFA was introduced to minimize the peak broadening effect. A gradient elution was employed for LC separation. The flow rate was reduced to 0.2 mL/min to enhance the sensitivity of ESItandem MS. The online SPE LC–ESI-tandem MS results of a spiked human urine sample are shown in Fig. 3. The retention time for Bup, nBup, Bup-3-G and nBup-3-G was 9.5, 7.9, 7.4 and 6.8 min, respectively. There was no deuterium effect in chromatographic separation observed for the internal standards. Each analysis including post-separation clean-up and equilibrium was accomplished in less than 20 min. 3.3. Evaluation of ion suppressing effect and sample pre-treatment procedure Ion suppression often diminishes response of ESI-MS and the sensitivity of analytical method. In order to examine the effect of ion suppression of this newly developed method, drugfree urine samples without analytes from five volunteers were examined by the apparatus described in Section 2.5. First, TFA (5 ␮L) was added to drug-free urine or water sample (5 ml), vortex for approximately 1 min then centrifuged for 5 min at 1500 × g. A 50-␮L aliquot of supernatant was injected for ion suppression examination. The responses of analytical components of these drug-free urine samples were compared to TFA added water sample; there was more than 35% decrease in each analyte detected. The results indicated that considerable ion suppression was observed in TFA added urine sample. Secondly, diluted drug-free urine, as described in Section 2.3, was evalu-

3.4. Extraction recovery Spiked urine and water samples at four different concentrations (1.0, 5.0, 30.0 75.0 ng/mL) over the linear range were analyzed to evaluate the recovery of this newly developed method. The recovery was determined by the peak area of spiked urine as a fraction of the corresponding standard in water solution. The results of this study are summarized in Table 3. Good recovery ranged from 93.6% to 102.2% for all analytes at the tested concentrations. It suggested that this assay is adequate for the analyses of these compounds in urine. 3.5. Calibration and validation Isotope internal standards, Bup-D4 and nBup-D3 , were utilized for quantitative determination. Due to the lack of availability of the isotope standard of glucuronides, Bup-D4 and nBup-D3 were also used as the internal standard for Bup-3-G and nBup-3-G, respectively. A series of spiked urine solutions was used to evaluate the linearity of this newly developed assay and the results are summarized in Table 4. A 1/x weighting factor was applied to least squares regression of quantities of each analyte versus peak are ratio of analyte peak area to that of its internal standard. For the analytes, except for nBup, good linearity was determined from 0.5 to 100 ng/mL. For nBup, linear range of 1–100 ng/mL was measured. Base on the signal-to-noise ratio of 3, the limit of detection for the analytes ranged from 0.2–0.5 ng/mL. Comparable detection limits of these analytes, 0.1 to 0.3 ng/ml, in human plasma have been reported previTable 4 Linearity and detection limit Compound

Calibration curvea

r2

Range (ng/mL)b

Detection limit (ng/mL)

Bup nBup Bup-3-G nBup-3-G

Y = 0.4864X − 0.0023 Y = 0.3218X + 0.0201 Y = 0.2627X − 0.0103 Y = 0.4494X − 0.0207

0.999 0.999 0.999 0.999

0.5–100 1.0–100 0.5–100 0.5–100

0.2 0.5 0.2 0.2

a Y: peak area ratio of standard and internal standard, X: concentration (ng/mL). b Concentration of standards: Bup, Bup-3-G, nBup-3-G: 0.5, 1, 3, 5, 10, 30, 50, 75 and 100 ng/mL; nBup: 1, 3, 5, 10, 30, 50, 75 and 100 ng/mL.

A.-C. Liu et al. / Talanta 75 (2008) 198–204

203

Table 5 Precision and accuracy Intra-day (n = 5) 1.0 ng/mL

Inter-day (n = 5) 10.0 ng/mL

Bup Mean Accuracy (%) CV (%)a

1.04 104.0 8.8

9.89 98.9 4.2

nBup Mean Accuracy (%) CV (%)a

0.99 99.0 6.3

9.72 97.2 3.4

Bup-3-G Mean Accuracy (%) CV (%)a

0.99 99.0 3.3

9.86 98.6 3.4

nBup-3-G Mean Accuracy (%) CV (%)a

0.97 97.0 10.1

10.15 101.5 7.3

a

100.0 ng/mL 98.94 98.9 4.2

1.0ng/mL

10.0 ng/mL

100.0 ng/mL

1.02 102.0 9.8

9.96 99.6 6.4

99.44 99.4 7.4

1.01 101.2 7.4

9.78 97.8 4.9

99.0 99.0 5.9

99.42 99.4 5.1

0.99 99.0 5.7

10.68 100.7 4.4

101.15 101.2 6.8

98.06 98.1 3.9

0.98 98.0 9.6

10.06 100.6 6.8

98.75 98.8 6.2

100.4 100.4 3.9

CV: coefficient of variation, n = 5.

ously [15,16]. These reported methods utilized SPE cartridge as a sample pre-treatment and detected by LC–ESI-tandem MS. They required laborious and time-consuming sample preparation procedures. However, only minimal sample preparation was required for this newly developed method. Precision and accuracy of the method were appraised at three concentrations (1.0, 10.0 and 100.0 ng/mL) over the linear range and the results were summarized in Table 5. The intra-day and inter-day precision showed coefficients of variance, CV, ranged from 3.3% to10.1% and 4.4% to 9.8%, respectively. The accuracy was evaluated by [mean measured concentration/spiked concentration] × 100%. Accuracy ranging from 97.0% to 104.0% was determined. 3.6. Short-term stability study The short-term stabilities of all four analytes under various storage conditions were examined. Fortified human urine samples (30 ng/ml of each analyte) were utilized for this study. Short-term temperature stability was evaluated by urine samples stored for 12 and 24 h at room temperature and 4 ◦ C. The freeze–thaw stabilities of all analytes were determined after three freeze–thaw cycles of the fortified human urine samples. For all four analytes, the concentrations were within ±12% under all five storage conditions. 3.7. Analysis of urine samples from buprenorphine treatment patients This method has been applied to the analysis of urine samples from the patients in the buprenorphine treatment program at Taipei City hospital. Extracted ion chromatograms of a patient urine sample are shown in Fig. 4. All four analytical compounds were detected in all patients’ urine samples examined and the results are summarized in Table 6. For all samples studied,

Fig. 4. Extracted ion chromatograms of a patient urine sample (sample #11) with the addition of ISTD (5 ng/ml for each ISTD).

Table 6 Bup, nBup, Bup-3-G and nBup-3-G concentration in urine collected from buprenorphine treatment patients Sample

Bup (ng/ml)

nBup (ng/ml)

Bup-3-G (ng/ml)

nBup-3G (ng/ml)

1a 2b 3a 4b 5a 6a 7a 8a 9a 10a 11a

14.1 29.8 16.2 10.2 8.2 14.4 21.1 17.5 18.5 30.5 19.1

116.4 136.6 62.1 216.7 86.8 150.4 62.1 63.2 76.5 120.3 95.8

352.5 213.3 31.7 88.7 115.4 31.9 85.8 108.2 30.1 107.1 137.4

406.3 858.8 122.8 983.9 390.4 212.8 437.3 407.6 373.1 372.2 438.9

a b

Sample was diluted fivefold prior to analysis. Sample was diluted 10-fold prior to analysis.

204

A.-C. Liu et al. / Talanta 75 (2008) 198–204

concentration of Bup was lower than its metabolites and nBup-3-G exhibited highest concentration of the four analytical compounds. Additional examination of urine samples is needed for better understanding of urinary concentration of Bup and its metabolites of each patient and the effectiveness of treatment program. 4. Summary This newly developed online SPE LC–ESI-tandem MS method has proven to be an efficient and sensitive method for the determination of Bup, nBup, Bup-3-G and nBup-3-G in human urine. The assay was fully automated and minimal sample preparation was required. The detection limits of this method, 0.2 to 0.5 ng/mL, are comparable to other reported studies; however, no laborious and time-consuming sample pre-treatment was needed in this assay. There was minimal ion suppression effect observed and good recovery was obtained. This method will be used to support the clinical study of continuing buprenorphine treatment program. Acknowledgements This study was financially supported by National Science Council of Taiwan. Authors would like to thank Prof. Ray H. Liu and Miss Yu-Shan Wang of Fooyin University for much assistance in this project. References [1] D. Jansinski, J. Pevnick, J. Griffth, Arch. Gen. Pyschiatry 34 (1978) 501. [2] C.N. Chiang, R.L. Hawks, Drug Alcohol Dependence 70 (2003) S39.

[3] E.J. Cone, C.W. Gorodetzky, D. Yousefnejad, W.F. Buchwald, R.E. Johnson, Drug Metab. Dispos. 12 (1984) 577. [4] R.C. Baselt, R.H. Cravey, Disposition of Toxic Drugs and Chemicals in Man, 4th ed., Chemical Toxicology Institute, Foster City, CA, 1995. [5] A. Elkader, B. Sproule, Clin. Pharmacokinet. 44 (2005) 661. [6] S. Gopal, T.B. Tzeng, A. Cowan, Eur. J. Pharm. Biopharm. 51 (2001) 147. [7] J. Kuhlman Jr., J. Magluilo Jr., E.J. Cone, J. Anal. Toxicol. 20 (1996) 229. [8] D.E. Moody, J.D. Kaycock, A.C. Spanbauer, D.J. Crouch, R.L. Foltz, J.L. Josephs, J. Amass, W.K. Bickel, J. Anal. Toxicol. 21 (1997) 406. [9] E. Schleyer, R. Lohmann, C. Rolf, A. Gralow, C.C. Kaufmann, M. Unterhalt, W. Hiddemann, J. Chromatogr. 614 (1993) 275. [10] L. Mercolini, R. Mandrioli, M. Conti, C. Leonardi, G. Gerra, M.A. Raggi, J. Chromatogr. B 847 (2007) 95. [11] D.E. Moody, M.H. Slawson, E.C. Strain, J.D. Laycock, R.L. Foltz, Anal. Biochem. 306 (2002) 2002. [12] D. Favretto, G. Frison, S. Vogliardi, S.D. Ferrara, Rapid Commun. Mass Spectrom. 20 (2006) 1257. [13] S. Pirnay, F. Herve, S. Bouchonnet, B. Perrin, F.J. Baud, I. Ricordel, J. Pharm. Biomed. Anal. 41 (2006) 1135. [14] A. Ceccato, R. Klinkernberg, P. Hubert, B. Streel, J. Pharm. Biomed. Anal. 32 (2003) 619. [15] A. Polettini, M.A. Huestis, J. Chromatogr. B 754 (2001) 447. [16] C.M. Murphy, M.A. Huestis, J. Mass Spectrom. 40 (2005) 70. [17] R. Kronstrand, T.G. Selden, M. Josefsson, J. Anal. Toxicol. 27 (2003) 464. [18] R. King, R. Bonfiglio, C. Fernandez-Metzler, C. Miller-Stein, T. Olah, J. Am. Soc. Mass Spectrom. 11 (2000) 942. [19] D.L. Buhrman, P.I. Price, P.J. Rudewicz, J. Am. Soc. Mass Spectrom. 7 (1996) 1099. [20] H. Mei, Y. Hsieh, C. Nardo, X. Xu, S. Wang, K. Ng, W.A. Korfmacher, Rapid Commun. Mass Spectrom. 17 (2003) 97. [21] M.R. Fuh, S.W. Lin, L.L. Chen, T.Y. Lin, Talanta 72 (2007) 1329. [22] T.Y. Wu, M.R. Fuh, Rapid Commun. Mass Spectrom. 19 (2005) 775. [23] C. Mallet, Z. Lu, J. Mazzeo, U. Neue, Rapid Commun. Mass Spectrom. 16 (2002) 805.