High throughput screening of sub-ppb levels of basic drugs in equine plasma by liquid chromatography–tandem mass spectrometry

Journal of Chromatography A, 1156 (2007) 271–279

High throughput screening of sub-ppb levels of basic drugs in equine plasma by liquid chromatography–tandem mass spectrometry Gary N.W. Leung ∗ , David K.K. Leung, Terence S.M. Wan ∗∗ , Colton H.F. Wong Racing Laboratory, The Hong Kong Jockey Club, Sha Tin Racecourse, Sha Tin, N.T., Hong Kong, China Available online 19 October 2006

Abstract This paper describes a high throughput LC–MS–MS method for the screening of 75 basic drugs in equine plasma at sub-ppb levels. The test scope covers diversified classes of drugs including some ␣- and ␤-blockers, ␣- and ␤-agonists, antihypotensives, antihypertensives, analgesics, antiarrhythmics, antidepressants, antidiabetics, antipsychotics, antiulcers, anxiolytics, bronchodilators, CNS stimulants, decongestants, sedatives, tranquilizers and vasodilators. A plasma sample was first deproteinated by addition of trichloroacetic acid. Basic drugs were then extracted by solidphase extraction (SPE) using a Bond Elut Certify® cartridge, and analysed by LC–MS–MS in positive electrospray ionization (+ESI) and multiple reaction monitoring (MRM) mode. Liquid chromatography was performed using a short C8 column (3.3 cm L × 2.1 mm ID with 3 ␮m particles) to provide fast analysis time. The overall instrument turnaround time was 8 min, inclusive of post-run and equilibration time. No interference from the matrices at the expected retention times of the targeted masses was observed. Over 60% of the drugs studied gave limits of detection (LoD) at or below 25 pg/mL, with some LoDs reaching down to 0.5 pg/mL. The inter-day precision for the relative retention times ranged from 0.01 to 0.54%, and that for the relative peak area ratios (relative to the internal standard) ranged from 4 to 37%. The results indicated that the method has acceptable precision to be used on a day-to-day basis for qualitative identification. © 2006 Elsevier B.V. All rights reserved. Keywords: Basic drugs; Horse blood; Plasma; High throughput screening; Liquid chromatography–mass spectrometry

1. Introduction Urine has long been the preferred matrix over blood in equine sports drug testing because a large volume of urine can usually be obtained, and the concentrations of drugs and metabolites are generally much higher. Blood however, offers an advantage over urine in that it can be collected on demand, thus, assuring samples are always available from the subject selected for testing. In some cases, blood may be the only sample available if the horses selected for testing fail to provide a urine sample. In addition, parent drugs can usually be found in blood to serve as good target analytes, whereas in urine, analysts might have to resort to the detection of metabolites, particularly for drugs that are extensively metabolized. The lack of reference materials for some metabolites and unknown drug metabolism in ∗

Corresponding author. Tel.: +852 2966 6469; fax: +852 2601 6564. Corresponding author. Tel.: +852 2966 6296; fax: +852 2601 6564. E-mail addresses: [email protected] (G.N.W. Leung), [email protected] (T.S.M. Wan). ∗∗

0021-9673/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2006.10.006

the horse further, complicate the problem. It is therefore, highly desirable if a mass-spectrometry based method for detecting a large variety of drugs in equine blood can be established. Gas chromatography–mass spectrometry (GC–MS) has long been the gold standard for testing of drugs in biological matrices. Unfortunately, it often lacks the required sensitivity for detecting drugs in blood, particularly for many basic drugs with exceptionally low concentrations in blood. The rapid development of liquid chromatography–tandem mass spectrometry (LC–MS–MS) in the past decade has provided analysts with a powerful tool for detecting and confirming the presence of drugs in complex biological matrices. Several workers have reported their applications for the detection of drugs in equine or human plasma samples in recent years [1–6]. However, their test scopes were limited to a single or only a few analytes of the same drug class. Recently, Herrin et al. [7] and Mueller et al. [8] have reported separately the application of a hybrid triple-quadruplole linear ion trap mass spectrometer (QTrap) to comprehensive screening of respectively, over 100 and 300 drugs in human blood samples. The strategy basically employed multiple reaction monitoring

272

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

(MRM) as survey scans for target detection, with automatic triggering of an enhanced product ion (EPI) scan in an information dependent acquisition (IDA) experiment. Drug identification was performed by library search with an in-house MS/MS library of EPI spectra collected at three different collision energies. This approach is quite appealing as both drug screening and identification can be done in the same LC–MS–MS run. However, most of their reported LoDs were only applicable to forensic investigation of toxicology cases, and not for doping control purpose where detection for evidence of prior exposure is often required. This paper describes a high throughput LC–MS–MS method for the simultaneous screening of 75 basic drugs of diversified drug classes in equine plasma at sub-ppb levels. 2. Experimental 2.1. Materials Anileridine hydrochloride, butorphanol tartrate, cimetidine, clonidine hydrochloride, cocaine hydrochloride, desipramine hydrochloride, droperidol, famotidine, guanabenz acetate, labetalol hydrochloride, lignocaine, mazindol, mephentermine sulphate, methadone hydrochloride, methoxamine hydrochloride, methoxyphenamine hydrochloride, methylphenidate hydrochloride, nadolol, naphazoline hydrochloride, nizatidine, nordazepam, nylidrin hydrochloride, oxycodone hydrochloride, oxymetazoline hydrochloride, oxymorphone, pindolol, prazosin hydrochloride, ranitidine hydrochloride, ritodrine hydrochloride, terbutaline sulphate, tuaminoheptane sulphate, and xylometazoline hydrochloride were obtained from USP (Rockville, MD, USA). Clenbuterol, salmeterol xinafoate, spiperone, and tetrahydrozoline hydrochloride were acquired from Sigma (St. Louis, MO, USA). Benzoylecgonine and N-norpropoxyphene maleate were obtained from Alltech (Deerfield, IL, USA). Anhydrous acepromazine, atenolol, haloperidol, nortriptyline hydrochloride, perphenazine, propylhexedrine, sotalol hydrochloride and thebaine were obtained from BP (Middlesex, UK). Potassium losartan and bisoprolol fumarate were obtained from Merck (Darmstadt, Germany), trifluperidol from Janssen Pharmacetica (NJ, USA), benperidol from Janssen-Cilag (Buckinghamshire, UK), and bambuterol hydrochloride from ASTRA (S¨odert¨alje, Sweden). Romifidine and telmisartan were from Boehringer Ingelheim (Ingelheim, Germany); hydroxydetomidine hydrochloride and detomidine hydrochloride were from Farmos (Turku, Finland). Buspirone hydrochloride was obtained from Bristol-Myers Squibb (NY, USA), bromocriptine mesylate from Apotex (Auckland, New Zealand), practolol from ICI (Now Zeneca Plc, UK), etafedrine from Merrel Dow Research (OH, USA), carteolol hydrochloride from Otsuka (Tianjin, China), sildenafil citrate from Pfizer (NY, USA), ␣-hydroxyalprazolam from Cerilliant (Austin, TX, USA), nalbuphine hydrochloride from Research Biochemicals Incorporated (MA, USA), buprenorphine hydrochloride from Schering-Plough (Hull, UK), carvedilol from Roche (Mannheim, Germany), irbesartan from Sanofi (Paris, France), pioglitazone from Takeda Chemical Industries (Osaka, Japan), and esmolol hydrochloride from The Boots

(Isando, South Africa). Amisulpride was obtained from Lab Synthelabo (Kuwait), flupentixol from Lundbeck (Lumsas, Denmark), midodrine hydrochloride from Hafslund Nycomed (Linz, Austria), isometheptene mucate from Manx (Kent, UK), repaglinide from Novo Nordisk (Bagsvaerd, Denmark), sulpiride from Sanofi-synthelabo (NY, USA), rilmenidine from Servier (France), and rosiglitazone maleate was from Smithkline Beecham (PA, USA). Acetic acid, glacial (100%), acetonitrile (LiChrosolv® ), ammonium acetate (10 mM, pH 3.8), ammonia solution (25%, GR grade), dichloromethane (GR grade), ethyl acetate (GR grade), isopropanol (GR grade), methanol (LiChrosolv® ), potassium hydroxide (pellets), potassium dihydrogen phosphate buffer (0.1 M, pH 6.0) and trichloroacetic acid (GR grade) were obtained from Merck (Darmstadt, Germany). Bond Elut Certify® cartridges (130 mg, 3 mL) were purchased from Varian (CA, USA). Deionized water was generated from an in-house water purification system (Milli-Q, Molsheim, France). 2.2. Sample preparation and extraction procedures Blood samples were centrifuged at 2100 g for 30 min. The plasma fraction (3 mL) was deproteinated by the addition of trichloroacetic acid (10% in deionized water, w/v, 200 ␮L). The deproteinated plasma was left standing at room temperature for 10 min and then centrifuged at 2100 g for 10 min. The supernatant was pipetted out and placed in another centrifuge tube. Nadolol (15 ng) was added as an internal standard (I.S.), followed by addition of potassium dihydrogen phosphate buffer (pH 6.0, 0.1 M, 2 mL). The pH was further adjusted, if necessary, to 6.0 using either potassium hydroxide (0.1 M) or hydrochloric acid (0.1 M). The sample was loaded onto a Bond Elut Certify® cartridge that had been pre-conditioned with methanol (2 mL), deionized water (2 mL), and potassium dihydrogen phosphate buffer (pH 6.0, 0.1 M, 2 mL). The cartridge was then washed with phosphate buffer (pH 6.0, 0.1 M, 2 mL), followed by acetic acid (1.0 M, 2 mL); dried for 5 min with nitrogen at 20 psi, and then eluted with dichloromethane/ethyl acetate (4:1, v/v, 3 mL) to collect the neutral and acidic fraction (this fraction can be used for the screening of neutral and acidic drugs if desired). The SPE cartridge was further washed with methanol (2 mL), dried for 5 min with nitrogen at 20 psi, and eluted with ethyl acetate/dichloromethane/isopropanol (5:4:1, v/v/v, 3 mL) containing 2% of concentrated aqueous ammonia to collect the basic fraction. The eluate was then evaporated to dryness under nitrogen at room temperature, and the residue was reconstituted in methanol (50 ␮L). The content was transferred to a conical insert in a Chrompack autosampler vial for LC–MS–MS analysis. 2.3. Instrumentation All LC–MS–MS analyses, except those described under “Method Applicability”, were performed on an Applied Biosystems 4000 Q Trap mass spectrometer (Applied Biosystems, Foster City, CA, USA) equipped with an Agilent 1100 series HPLC system consisting of a quaternary gradient pump (Agilent Tech-

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

273

Table 1 MS parameters for the detection of the 75 basic drugs Target drugs

Precursor ion (m/z)

Product ion (m/z)

CE (eV)

DP (V)

EP (V)

CXP (V)

Acepromazine Amisulpride Anileridine Atenolol Bambuterol Benperidol Benzoylecgonine Bisoprolol Bromocriptine Buprenorphine Buspirone Butorphanol Carteolol Carvedilol Cimetidine Clenbuterol Clonidine Cocaine Desipramine Detomidine Droperidol Esmolol Etafedrine Famotidine Flupentixol Guanabenz Haloperidol Hydroxydetomidine Irbesartan Isometheptene Labetalol Lignocaine Lorsartan Mazindol Mephentermine Methadone Methoxamine Methoxyphenamine Methylphenidate Midodrine Nadolol (I.S.) Nalbuphine Naphazoline Nizatidine Nordazepam Norproxyphene Nortriptyline Nylidrin Oxycodone Oxymetazoline Oxymorphone Perphenazine Pindolol Pioglitazone Practolol Prazosin Propylhexedrine Ranitidine Repaglinide Rilmenidine Ritodrine Romifidine Rosiglitazone

327.2 369.9 353.2 267.2 368.3 382.2 290.0 326.2 654.2 468.3 386.3 328.0 293.0 407.2 253.2 277.0 230.0 304.1 267.2 187.1 380.2 296.2 194.0 338.3 435.2 231.1 376.1 203.1 429.2 142.2 329.2 235.2 423.2 285.1 164.2 310.2 212.2 180.2 234.2 255.1 310.0 358.0 211.2 332.2 271.1 326.2 264.2 300.2 316.0 261.2 302.2 404.2 249.0 357.1 267.2 384.0 156.2 315.2 453.2 181.1 288.2 258.0 358.1

86.1 242.1 120.1 190.1 294.2 165.1 168.0 116.1 636.4 55.1 122.1 310.0 237.0 100.1 159.2 203.0 44.0 182.2 208.1 81.0 165.1 145.1 176.0 188.9 305.0 172.0 165.1 185.0 207.2 69.2 311.3 86.2 207.1 44.1 91.1 265.2 162.1 121.0 84.2 237.2 254.0 340.0 115.1 155.1 140.3 143.1 233.0 150.1 298.0 205.2 284.1 171.2 116.0 134.1 190.2 247.0 69.1 176.1 230.3 95.1 270.2 160.1 135.1

29 39 39 27 29 37 29 27 25 95 45 35 23 43 21 25 53 29 31 29 39 29 21 29 43 32 35 35 35 21 21 27 35 55 33 23 29 29 29 13 25 33 69 28 41 35 35 31 27 37 27 35 27 41 27 41 25 26 41 21 21 47 39

71 111 96 80 81 96 66 96 91 131 121 96 71 86 66 61 91 76 61 66 66 41 66 66 106 90 76 56 96 51 66 71 86 91 46 66 50 51 61 51 81 71 91 60 81 46 76 66 66 101 80 51 76 111 81 116 56 60 96 46 61 111 106

10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

6 12 8 10 8 8 10 8 6 8 6 14 12 6 8 10 6 10 4 6 8 8 10 10 16 8 8 6 10 2 10 4 12 6 6 14 8 8 6 12 14 10 6 10 4 10 6 8 8 10 10 10 6 8 10 14 2 10 10 6 8 8 8

274

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

Table 1 (Continued ) Target drugs

Precursor ion (m/z)

Product ion (m/z)

CE (eV)

DP (V)

EP (V)

CXP (V)

Salmeterol Sildenafil Sotalol Spiperone Sulpiride Telmisartan Terbutaline Tetrahydrozoline Thebaine Trifluperidol Tuaminoheptane Xylometazoline ␣-Hydroxyalprazolam

416.3 475.2 273.1 396.0 342.1 515.2 226.1 201.2 312.2 410.1 116.2 245.2 325.1

398.4 100.2 255.2 165.1 112.1 276.1 152.1 131.0 58.0 165.1 57.1 189.1 297.1

23 41 19 37 39 65 23 39 39 37 19 37 37

86 116 61 86 101 116 61 91 66 106 46 96 106

10 10 10 10 10 10 10 10 10 10 10 10 10

14 6 14 8 6 14 8 8 10 8 10 10 16

CE, collision energy; DP, declustering potential; EP, entrance potential; CXP, collision cell exit potential.

nologies, Palo Alto, CA, USA). The HPLC system was replaced with an Agilent 1100 series capillary HPLC system consisting of a binary gradient pump when performing the LC–MS–MS experiments described under “Method Applicability”. For both HPLC systems, the same LC settings were used. Solid-phase extraction (SPE) was carried out using a RapidTrace® SPE workstation (Zymark Corporation, Hopkinton, MA, USA). 2.4. LC conditions for the screening of basic drugs A reversed-phase SupelcosilTM LC-8-DB column (3.3 cm × 2.1 mm ID, 3 ␮m; Supelco, Bellefonte, PA, USA) was used. The mobile phase was composed of 10 mM ammonium acetate (pH 3.8) as solvent A and acetonitrile as solvent B. A linear gradient was run at 300 ␮L/min, with 100% solvent A and 0% solvent B for the first 0.5 min (t = 0–0.5 min), decreasing to 0% solvent A and increasing to 100% solvent B from t = 0.5 to 2 min, and hold for 1 min (until t = 3 min). The gradient was then returned to 100% solvent A and 0% solvent B from t = 3 to 3.5 min, and stabilised until t = 8 min when the next injection would start. The injection volume was 5 ␮L. 2.5. MS conditions for the screening of basic drugs Detection of the 75 basic drugs was performed in MRM mode with a single time segment. The MRM transitions, collision energies and other target-dependent parameters for each target were optimized by infusion of the corresponding reference material (Table 1) into the 4000 Q Trap. The dwell time for each transition was 8 ms with a 5 ms pause time. The overall cycle time for 76 MRM transitions (including one for the internal standard) was 0.99 s. The source was operated in positive ESI mode at 400 ◦ C with the nebulizer gas and heater gas set to 60 and 30 psi, respectively. IonSpray voltage was set to 5500 V, curtain gas was set to 13 psi, and collision gas was set to medium. The resolution for the selection of the precursor ions in Q1 and the product ions in Q3 was each set to unit mass. Data processing was performed using the Analyst (Version 1.4.1) software.

2.6. Settings for information dependent acquisition and enhanced product ion scan experiments The IDA criteria were set to acquire an EPI spectrum using dynamic fill time (2–250 ms) when the signal was greater than 3000 counts per second (cps). The collision energy used for the EPI scan was the rolling collision energy, which was set automatically by the software based on the molecular weight of the precursor ion. Dynamic exclusion was set to 30 s to exclude the acquisition of the EPI scan for the same transition, allowing the detection of co-eluting targets. EPI scan range was m/z 100–680, and the overall duty cycle for one complete IDA experiment (76 MRM + 1 EPI scan) was 1.8 s. 2.7. Semi-quantification of target compounds For each batch of plasma samples a calibrator, containing the targeted drugs spiked in negative horse plasma, was processed in parallel. The spiked concentrations of individual targets in the calibrator were four times higher than the corresponding quality control (QC) sample shown in Table 2. A one-point calibration curve was prepared for each analyte using the Analyst (Version 1.4.1) software. The concentrations of the target drugs in plasma samples were calculated automatically by the software. 3. Results and discussion 3.1. Solid-phase extraction The solid-phase extraction protocol used in this study was based on the work by Wynne et al. [9] for equine urine and with modification to reduce sample blockage in the SPE cartridges. This modified SPE protocol has been applied in the authors’ laboratory to process equine urine samples since 2004 for the screening of a large variety of drugs by both GC–MS and LC–MS–MS [10]. For plasma samples, a protein precipitation step using trichloroacetic acid was found necessary to minimize column blockage in subsequent solid-phase extraction. Two SPE fractions were collected from each C8-SCX SPE cartridge. The first fraction would contain acidic and neutral drugs and was used

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

275

Table 2 Inter-day precision data on relative retention times and relative peak area ratios for the 75 basic drugs using nadolol as the internal standard Drug

Spike conc. (QC sample) (pg/mL)

Anileridine Buspirone Haloperidol Lignocaine Methylphenidate Naphazoline Tetrahydrozoline Bambuterol Bisoprolol Butorphanol Rosiglitazone Xylometazoline Cocaine Prazosin Acepromazine Pioglitazone Methadone Nylidrin Oxymetazoline Trifluperidol Amisulpride Carvedilol Detomidine Guanabenz Methoxyphenamine Midodrine Rilmenidine Spiperone Telmisartan Mazindol Benperidol Droperidol Isometheptene Mephentermine Nordazepam Irbesartan Perphenazine Sildenafil Atenolol Clenbuterol Flupentixol Labetalol Desipramine Esmolol Ranitidine Repaglinide Thebaine ␣-Hydroxyalprazolam Benzoylecgonine Carteolol Cimetidine Famotidine Lorsartan Methoxamine Nizatidine Nortriptyline Salmeterol Pindolol Propylhexedrine Buprenorphine Etafedrine Nalbuphine Practolol

100 100 100 100 100 100 100 100 100 100 100 100 250 250 250 100 500 500 750 750 1000 100 100 1000 100 1000 100 1000 1000 1000 500 100 100 100 500 250 250 250 1000 100 1000 100 250 500 500 250 500 500 1000 1000 1000 1000 1000 1000 1000 500 100 750 750 1000 1000 500 1000

LoD (pg/mL) 0.5 0.5 0.5 0.5 0.5 0.5 0.7 1 1 1 1 1 1.25 1.25 1.7 2 2.5 2.5 3.75 3.75 5 5 5 5 5 5 5 5 5 6.7 10 10 10 10 10 12.5 12.5 12.5 20 20 20 20 25 25 25 25 25 25 50 50 50 50 50 50 50 50 50 75 75 100 100 100 100

Relative RT ratio (% RSD) n = 4

Relative peak area ratio (% RSD) n = 4

0.02 0.48 0.06 0.01 0.04 0.49 0.09 0.52 0.29 0.13 0.50 0.43 0.18 0.31 0.50 0.27 0.42 0.48 0.07 0.49 0.15 0.06 0.19 0.54 0.09 0.01 0.20 0.17 0.30 0.08 0.18 0.08 0.49 0.47 0.43 0.46 0.41 0.08 0.19 0.07 0.09 0.03 0.11 0.14 0.13 0.09 0.45 0.13 0.06 0.22 0.10 0.16 0.48 0.47 0.13 0.15 0.26 0.21 0.02 0.27 0.10 0.27 0.18

14 12 21 13 23 20 15 10 8 19 37 14 13 22 4 28 14 20 20 22 17 8 5 28 15 22 12 31 36 8 16 5 13 14 10 37 17 10 20 19 13 34 19 15 27 20 23 29 7 18 13 17 26 20 27 15 25 28 12 30 22 20 11

276

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

Table 2 (Continued ) Drug

Spike conc. (QC sample) (pg/mL)

Sulpiride Tuaminoheptane Bromocriptine Clonidine Hydroxydetomidine Norproxyphene Oxymorphone Ritodrine Romifidine Terbutaline Oxycodone Sotalol

1000 1000 750 1000 1000 1000 1000 1000 1000 1000 500 1000

LoD (pg/mL) 100 100 150 200 200 200 200 200 200 200 500 500

to screen acidic drugs by GC–MS and selected neutral drugs (such as steroids) by LC–MS–MS. The second or basic fraction was used to screen basic drugs using the LC–MS–MS procedures described in this study. The target basic drugs selected for this study were mainly drugs with high potential of abuse in equine sports, which require low levels of detection (sub-ppb levels) in plasma. Attempts were made to use GC–MS to screen basic drugs in the second fraction. However, the limits of detection for many target drugs were found to be too high for doping control purpose. Nevertheless of this, GC–MS is used in the authors’ laboratory as a complementary technique for detecting other basic drugs in equine plasma (at or above low ppm levels), which are not covered by the LC–MS–MS method. Drug detections by GC–MS were achieved using two automatic library search algorithms, namely Probability-Based Match (PBM) and Automatic Mass Spectral Deconvolution and Identification System (AMDIS). 3.2. Method characteristics The concentrations of the target analytes in the QC samples are shown in Table 2. All 75 drugs could be easily detected at the QC levels, which ranged from 100 to 1000 pg/mL. The overall LC–MS turnaround time was 8 min inclusive of post-run and solvent equilibration times. For every batch of equine plasma samples (about 40–60), cleaning of the curtain plate with methanol was performed before sample analysis. A QC sample was injected at the beginning and end of the analytical sequence to verify that the analytical process was in control. No significant change in sensitivity was observed between the two QC sample injections throughout this study, indicating that the method is robust for routine use. Limits of detection were estimated by analysing the targeted analytes at different concentrations (down to 0.5% of the concentrations in the QC sample) in spiked plasma samples. LoD for a target analyte here represents the lowest spiked concentration of the analyte evaluated that can still consistently give a signal-to-noise ratio (S/N) greater than 3. Over 60% of the targets have LoDs less than or equal to 25 pg/mL, with some LoDs down to 0.5 pg/mL. Method specificity of this LC–MS screening method was assessed by testing a total of 25 negative post-race horse plasma

Relative RT ratio (% RSD) n = 4

Relative peak area ratio (% RSD) n = 4

0.09 0.04 0.50 0.06 0.17 0.11 0.23 0.11 0.17 0.30 0.16 0.22

28 17 32 16 27 29 24 29 13 36 14 17

samples. Interferences from the different plasma matrices at the expected retention times of the target transitions were not observed. The effect of ion suppression was evaluated by comparing the signals from an equine plasma extract spiked with the 75 target drugs with those from a standard mixture of the 75 drugs. Signal suppression ranged from 95 to −96% (i.e., enhancement), with the majority (65 drugs) observed between 0 and 80%. Despite significant suppressions for most targets, the method could still detect their presence as demonstrated from the results obtained with the QC samples. The effect of dwell time on sensitivity has been reported by Herrin et al. [7]. As expected, longer dwell times usually led to better sensitivity, although the gain in sensitivity was moderate beyond 25 ms. The down side of longer dwell times is obviously a longer duty cycle time that can have a negative impact on chromatographic data points, particularly for the narrow peaks. No thorough investigation was performed in this study to identify the optimum dwell times for individual targets. However, a quick evaluation was done at the initial stage using dwell time settings at 8 ms (0.99 s cycle time) and 20 ms (1.9 s cycle time). All 75 targets could be easily detected with sufficient sensitivity for both settings. However, using dwell time of 20 ms did not result in enhanced peak areas (55 targets decreased and 20 targets increased), probably because of insufficient data points across the chromatographic peaks (∼11 s base to base peak width). A dwell time of 8 ms was therefore, used for all targets in subsequent MRM experiments. The inter-day precision of the relative peak area ratios (analyte-to-I.S.) and the relative retention time ratios for the 75 basic drugs at the QC levels were examined by analysing a spiked plasma sample for 4 different days. The inter-day precision for the relative retention times ranged from 0.01 to 0.54%, and that for the relative peak area ratios from 4 to 37% (Table 2). The results indicate that the method has adequate precision to be used on a day-to-day basis for qualitative identification. Background limits were established for each target analyte and were used as criteria for identifying suspicious samples. Samples with a target analyte at or above its background limit are considered suspicious and will be investigated further using MRM or Q Trap full-scan mode. A number (about 100) of negative equine plasma samples was analysed using the LC–MS–MS

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

method, and the background limit for each target analyte was derived by rounding up the mean plus three standard deviations of its apparent concentration. 3.3. Method applicability Horse-in-Training (‘HIT’) sample testing is performed routinely in Hong Kong. This doping control program is similar to the out-of-competition testing in human sports. The objective is to ensure unauthorized medications are not used during training. Both urine and blood are collected from horses selected for ‘HIT’ testing. In order to demonstrate that the above screening method is effective in detecting trace level of basic drugs in equine plasma, it is applied to five ‘HIT’ blood samples of which their corresponding urine samples had been reported positive. The drugs reported in the urine samples were butorphanol (a common veterinary analgesic), clenbuterol (a ␤-agonist) and 3-hydroxylignocaine (a metabolite of the local anaesthetic lignocaine). Fig. 1 shows the screening results for the five plasma samples. All showed the expected MRM transitions at the expected retention times. The changes in the estimated plasma concentrations of the target analytes in Samples 1–4 were roughly in line with those in the corresponding urine samples (Table 3). For Sample 5, only the parent drug lignocaine was detected in plasma. This example demonstrates one of the advantages of using blood over urine as the test matrix, as exposure can usually be detected by

277

targeting the parent drugs rather than its metabolites, reference standards of which may not be readily available. IDA experiments were also performed on the five blood samples using an IDA intensity threshold of 3000 cps. Only Blood Sample 3, which contained clenbuterol at around 1800 pg/mL, could trigger an EPI full-scan mass spectrum (Fig. 2). The results indicate that a lower IDA intensity threshold is necessary in order to obtain automatic EPI scans for samples with low analyte concentrations. However, if the threshold is set too low, it runs the risk of triggering too many unnecessary IDA events. This can have a negative impact on the quality of the MRM data acquisitions. The normal cycle time for the 76 MRM transitions (75 targets + 1 internal standard) is 0.99 s, which generates around 15 data points for a typical LC peak of 15 s in this study. One additional EPI scan would increase the total duty cycle to 1.8 s. Therefore, multiple EPI scans triggered within a short retention time window would reduce the number of data points to an unacceptable level. Indeed, broadening of chromatographic peaks, due to fewer data points and the effect of data smoothing, was observed when IDA experiments were performed. In addition, data smoothing tends to exaggerate the peak area slightly, resulting in a higher calculated amount on the target analyte. The use of a longer column would resolve the 76 targets better, and should help to alleviate the problem by reducing the total number of EPI scans at a particular time window, but productivity would decrease due to a longer turnaround time. Another draw-

Fig. 1. MRM results for blood samples suspected to contain (a) butorphanol; (b) clenbuterol; and (c) lignocaine.

278

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

Table 3 Approximate drug concentrations in the ‘HIT’ blood samples and their corresponding urine samples Blood sample

Drug detected

Estimated plasma concentration (pg/mL)

Estimated conc. in the corresponding urine samples (pg/mL)

Time difference in sample collection (urine preceded by blood)

Sample 1 Sample 2 Sample 3 Sample 4 Sample 5

Butorphanol Butorphanol Clenbuterol Clenbuterol Lignocaine

18 4 1,800 23 56

23,000 500 37,000 500 9,000 (as 3-hydroxy-lignocaine)

∼2 h ∼2 h <0.2 h ∼5 h ∼8 h

Fig. 2. Selected product-ion chromatograms of m/z 203 obtained from (a) Sample 3, and (b) a clenbuterol reference standard, and their corresponding EPI full-scan mass spectra of m/z 277, as shown in (c) and (d), respectively, triggered by IDA.

back of using IDA in multi-drug screening is that both the IDA intensity threshold and the collision energy are global settings, which do not allow optimal conditions to be applied to individual targets. In view of these limitations, the strategy adopted by the authors’ laboratory is to use MRM only for screening. Any sample found with a target concentration above its corresponding background limit would be subjected to a second injection, either in MRM or Q Trap full-scan mode, for further identification using optimal MS/MS settings for the suspicious target. If evidence for the presence of the target analyte is also obtained from the second injection, a new aliquot of the sample in question will be re-tested under optimised full scan or MRM settings. Confirmation for the presence of a prohibited substance is achieved by direct comparison of the full scan or MRM data with those from a reference standard analysed in series with the sample.

4. Conclusion A high throughput LC–MS–MS method was developed for the simultaneous screening of 75 basic drugs using a short LC column coupled with a fast scanning triple-quadrupole mass spectrometer. The 75 targets could all be detected at parts per trillion levels within an 8 min LC–MRM cycle. The method had adequate inter-day precision to be used as a qualitative screening method on a regular basis. Matrix interference was not observed at the expected retention times of the target transitions. The applicability of the method has been demonstrated by the detection of clenbuterol, butorphanol and lignocaine in ‘horse-in-training’ blood samples. As detections were by a fast scanning triple quadrupole mass spectrometry, the method can easily be expanded to accommodate additional target analytes.

G.N.W. Leung et al. / J. Chromatogr. A 1156 (2007) 271–279

Acknowledgements The authors wish to thank Stephen Cheung, Pauly Chan, Christina Tang and Sandra Chan for their technical assistance. References [1] J.Y. Kim, M.H. Choi, S.J. Kim, B.C. Chung, Rapid Commun. Mass Spectrom. 14 (2000) 1835. [2] M. Gergov, J.N. Robson, I. Ojanpera, O.P. Heinonen, E. Vuori, Forensic Sci. Int. 121 (2001) 108. [3] H.H. Maurer, O. Tenberken, C. Kratzsch, A.A. Weber, F.T. Peters, J. Chromatogr. A 1058 (2004) 169.

279

[4] B.E. Smink, J.E. Brandsma, A. Dijkhuizen, K.J. Lusthof, J.J. de Gier, A.C. Egberts, D.R. Uges, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 811 (2004) 13. [5] F. Guan, C.E. Uboh, L.R. Soma, Y. Luo, J.S. Jahr, B. Driessen, Anal. Chem. 76 (2004) 5127. [6] Y. Luo, C.E. Uboh, L.R. Soma, F.Y. Guan, J.A. Rudy, D.S. Tsang, Rapid Commun. Mass Spectrom. 19 (2005) 1245. [7] G.L. Herrin, H.H. McCurdy, W.H. Wall, J. Anal. Toxicol. 29 (2005) 599. [8] C.A. Mueller, W. Weinmann, S. Dresen, A. Schreiber, M. Gergov, Rapid Commun. Mass Spectrom. 19 (2005) 1332. [9] P.W. Wynne, D.C. Batty, J.H. Vine, N.J.K. Simpson, Chromatographia 59 (2004) S51. [10] E.N. Ho, D.K. Leung, T.S. Wan, N.H. Yu, J. Chromatogr. A 1120 (2006) 38.