Wc

Na

b

c

a

ARRAA

KVSUUOQ

1

pgmf[

aegfhtcbR

0h

Journal of Chromatography A, 1258 (2012) 55– 65

Contents lists available at SciVerse ScienceDirect

Journal of Chromatography A

jou rn al h om epage: www.elsev ier .com/ locat e/chroma

ide-range screening of banned veterinary drugs in urine by ultra high liquidhromatography coupled to high-resolution mass spectrometry

uria Leóna, Marta Rocaa, Carmen Igualadaa, Claudia P.B. Martinsb, Agustín Pastorc, Vicent Yusáa,∗

Center for Public Health Research (CSISP), Avda de Cataluna 21, 46020 Valencia, SpainThermo Fisher Scientific, Barcelona, SpainAnalytical Chemistry Department, Universidad de Valencia, Edifici Jeroni Munoz, 50, Dr. Moliner, 46100 Burjassot, Valencia, Spain

r t i c l e i n f o

rticle history:eceived 21 May 2012eceived in revised form 6 August 2012ccepted 9 August 2012vailable online 18 August 2012

eywords:eterinary drugscreeningrine

a b s t r a c t

In this work, an ultra high performance liquid chromatography–high resolution mass spectrometry(UHPLC–HRMS) methodology is proposed for the multi-class multi-residue screening of banned andunauthorized veterinary drugs in bovine urine, using an Orbitrap ExactiveTM analyzer working at a resolv-ing power of 50,000 FWHM in full scan, both in positive and negative mode. The method currently covers87 analytes belonging to different families such as steroid hormones, �-agonists, resorcylic acid lactones(RAL), stilbens, tranquillizers, nitroimidazoles, corticosteroids, NSAIDs, amphenicoles, thyreostatics andother substances such as dapsone. A database including the elemental composition, the polarity of acqui-sition, retention time and expected adducts was built for the targeted analysis, and a high mass accuracy(<5 ppm) was set as one of the identification criteria. After comparing different sample preparation proce-

HPLC–HRMSrbitrapuEChERS

dures, QuEChERS was selected as the most appropriate methodology. An efficient separation of analyteswas achieved using ultra high performance liquid chromatography with a column packed with sub-2 �mparticles. The performance of the method has been evaluated in accordance with the EU guidelines for thevalidation of screening methods for the analysis of veterinary drugs residues. The screening target concen-trations were established between 0.2 �g/l and 20 �g/l, demonstrating the usefulness of UHPLC–HRMSas an ideal tool for compliance monitoring in regulatory laboratories.

. Introduction

The use of veterinary drugs is heavily regulated in the Euro-ean Union (EU) by different Regulations and Directives [1–4]. Theroup of analytes concerning grow-promoting agents (e.g., hor-ones and �-agonists) and those not regulated or unauthorized

or some matrices/species combinations are of particular interest1,3].

Analytical methods are essential in order to develop humanssessment exposure to these substances and to support thenforcement of law and regulations. Generally, analytical strate-ies include a two tiered approach: the screening as a first stepollowed by a confirmatory step of a positive result [5]. Nowadays,igh-throughput screening methods are necessary to optimizehe cost-effectiveness of analytical procedures and the low per-entage of non-compliant samples obtained (from 0 to 0.46% for

anned substances in the EU, according to the data of the Nationalesidue Monitoring Plan, 2009) [6]. A way to achieve this is to

∗ Corresponding author. Tel.:+34 96 1925865; fax: +34 96 1925704.E-mail address: yusa vic@gva.es (V. Yusá).

021-9673/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.chroma.2012.08.031

© 2012 Elsevier B.V. All rights reserved.

maximize the number of analytes checked through the develop-ment of multiclass–multiresidue methods.

Although growth-promoting agents and veterinary drugs gen-erally show high clearance rates in urine, this matrix is widelyused to monitor the illegal use of these substances [7–9]. Evenif direct analysis of samples (dilute and shot) is optimal [10],extracting chemical residues from the sample and a clean upstep is usually necessary [7]. Liquid–liquid extraction [11] andsolid-phase extraction [12] are the most important sample prepa-ration approaches to extract and purify veterinary drugs residuesfrom urine. However, for real multiclass–multiresidue analy-sis there is a trend in using more generic sample preparationprocedures since extraction of substances with very differentphysico-chemical properties is requested. The QuEChERS (stand-ing for Quick, Easy, Cheap, Effective, Rugged and Safe) procedurehas been widely used in the pesticide field [13,14]. This is a veryflexible method that permits modifications depending on analytes,matrices or analyst preferences. It has also been used for veteri-nary residues in different matrices [15]. However, to the best of

our knowledge, it has not been applied to veterinary drugs inurine.

Currently, multianalyte methods, both for screening and confir-mation purposes, are carried out using liquid chromatography (LC)

5 atogr

crstatip

tdorusutnshtstrtst

angcTfeart

2

2

S9w(

ickPp9S

5E(

2

f

resolving power of 50,000 FWHM. The value of mass toler-

6 N. León et al. / J. Chrom

oupled to a triple quadrupole analyser (QqQ). There are severalecent examples of methods developed for the analysis of theseubstances in urine samples using LC–MS/MS [16–19]. Althoughhis technique presents a high sensitivity and selectivity, it requiresn extensive method set-up. Also this technique presents limita-ions with regard to the number of compounds that can be analysedn one run and, only targeted analytes can be detected without theossibility of retrospective data analysis [20,21].

The full scan approach using high resolution mass spectrome-ry (HRMS) has emerged as a promising alternative that allows theevelopment of sensitive wide-range screening of veterinary drugsvercoming the limitations of SRM analysis [22,23]. Currently, highesolution and accurate mass measurements can be achieved bysing time of flight (TOF) mass spectrometers [5], or by using singletage Orbitrap (Exactive) technology, which provides resolutionsp to 100,000 full width at half maximum (FWHM) [24,25]. Theseechniques make it possible to simultaneously analyse an unlimitedumber of compounds, and also to detect non preselected and pos-ibly unknown analytes through the retrospective assessment ofistorical data (“post-targeted” analysis) [21,26,27]. Additionally,he accurate mass measurement of HRMS ensures an enhancedelectivity compared with tandem mass spectrometry, and also,he spectra produced by HRMS can provide further information ofesolved isotope peaks and isomeric ions. All these features makehis new approach a reliable and powerful tool for the wide-rangecreening of veterinary drug residues in the official control labora-ories (Fig. 1).

To our knowledge, no work has previously been reported,nalysing a complete set of banned or not regulated veteri-ary drugs in urine using the combination of a simple andeneric extraction procedure and ultra high performance liquidhromatography–high resolution mass spectrometry detection.herefore the aim of this study was to develop a new approachor the multi-residue screening of 87 banned or not regulated vet-rinary drugs in urine by making use of an Orbitrap ExactiveTM

nalyser. This new approach results in a simplification of the cur-ent laboratory methodologies and a considerable improvement inhe analytical control strategy.

. Materials and methods

.1. Chemicals and reagents

Acetonitrile and methanol were LC–MS grade and supplied bycharlab (Barcelona, Spain). Acetic acid and formic acid (purity8–100%), water, ammonium formiate and ammonium acetateere hypergrade quality grade and were purchased from Merck

KGaA, Darmstadt, Germany).QuEChERS extraction kit (EN extraction salt packet contain-

ng 4 g MgSO4, 1 g NaCl, 1 g sodium citrate, 0.5 g disodiumitrate sesquihydrate) and two QuEChERS fatty dispersive-SPEits (a) AOAC kit, 15 ml polypropylene tube containing 400 mgSA, 400 mg C18 and 1200 mg MgSO4 and (b) EN kit, 15 mlolypropylene tube containing 150 mg PSA, 150 mg C18 and00 mg MgSO4 were obtained from Agilent Technologies (Madrid,pain).

Solid phase extraction (SPE) cartridges Oasis-HLB® (6 cm3,00 mg) were obtained from Waters (Mildford, MA, USA) and Bondlut Certify® cartridges (130 mg, 3 ml) were purchased from VarianHarbor City, CA, USA).

.2. Standards and stock solutions

All commercial standards were of high purity and were obtainedrom Sigma–Aldrich (Barcelona, Spain), Witega (Berlin, Germany),

. A 1258 (2012) 55– 65

USP Reference Standards (MD, USA), Medical Isotopes (Waltham,MA, USA), RIKILT (Community Reference Laboratory, Wageningen,The Netherlands), LGC Standards S.L.U. (Barcelona, Spain) or Biop-ure (Tulln, Austria).

Individual stock solutions containing 10–500 �g/ml of analyteswere prepared by weighing each compound and dissolving it inmethanol. They were stored at −20 ◦C for a maximum of 6 months.10 �g/ml multianalyte intermediate standard solutions wereprepared by diluting with acetonitrile the individual stock solu-tions, except for 4�-hydroxystanozolol, 4�-hydroxystanozolol,3′-hydroxystanozolol, 16�-hydroxystanozolol, �-boldenone, �-testosterone, �-nortestosterone, zearalenone, metaproterenol,chlorpromazine and all the nitroimidazoles, thyreostats and nonsteroideal anti-inflammatory drugs (NSAIDs) which were addeddirectly to the working solution. The working mixed standardsolution was prepared by diluting the multi-analyte intermediatestandard solutions or stock solutions with a mixture of acetoni-trile:water (20:80, v/v). The concentration of the analytes in thisworking solution ranged from 50 to 5000 ng/ml depending on thecompound and to reach the screening target level when adding20 �l to the blank urine sample. A working solution of 250 ng/mlwas also prepared containing all of the analytes used as internalstandards.

2.3. Instrumentation and UHPLC–HRMS parameters

Chromatographic separation was performed on an Accela liquidchromatography UHPLC system equipped with a Hypersil Gold aQcolumn (100 mm × 2.1 mm, 1.9 �m) both from ThermoFisher Sci-entific (Bremen, Germany). The flow rate used was 400 �l/min andthe injection volume was 10 �l. Separations were performed usinga binary gradient. Mobile phase components were: solvent A: 0.1%acetic acid aqueous solution and solvent B: acetonitrile contain-ing 0.1% acetic acid. The initial mobile phase proportion, 100% (A)decreased linearly to 50% in 5 min and held at 50% for 3 min. Afterthat, solvent A decreased linearly to 5% in 0.5 min and maintainedfor 1.5 min. Finally, solvent A changed to the initial percentage(100%) in 0.5 min and was kept at 100% for 3 min to equilibrate thecolumn before the next injection. The total run time was 13.5 min.Data acquisition was performed by the Thermofisher Scientific’sXcalibur 2.1.0 software. The stability of the retention times usingthe optimized chromatographic conditions, was tested through20-fold measurements of a standard solution (100 �g/l) overtwo days.

Mass analysis has been performed on the Orbitrap mass spec-trometer ExactiveTM analyser (Thermofisher Scientific, Bremen,Germany). The system was equipped with a heated electro-spray ionization interface (HESI-II) and operated simultaneously inpositive and negative modes. The ESI source parameters were opti-mized by direct infusion of standard solutions of all the analytes(10 �g/ml) at a flow rate of 10 �l/min. The different parameterswere manually varied to obtain the maximum total ion current sig-nal (TIC) both in positive and negative operation mode within themass range of 65–500m/z.

Parameters of the ion source were as follows: spray voltage:3.5 kV (positive mode) and 2.5 kV (negative mode); sheath gas flow-rate: 50; auxiliary gas flow-rate: 10; skimmer voltage: 25 V; heatertemperature: 250 ◦C; capillary temperature: 270 ◦C; capillary volt-age: 40 V and tube lens voltage: 120 V.

The system operated in full-scan mode (65–500 Da) at a

ance was set at 5 ppm and no specific lock mass was usedfor internal mass axis correction (external mass calibration). Forthe Automatic Gain Control (AGC) the “Balanced” setting wasselected.

N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65 57

esidue

2

2

Stst

2

sqtoia

2

diet

2wwts

Fig. 1. Wide-range r

.4. Sample preparation/sample processing

.4.1. SamplingBovine urine samples from adult animals were collected by the

panish Veterinary Inspectors from the Health Authority as part ofhe Spanish National Residue Control Plan at livestock farms andlaughterhouses. As soon as the samples arrived at the laboratory,hey were stored at −20 ◦C until analysis.

.4.2. Enzymatic hydrolysis step5 ml of sodium acetate 0.2 M (pH 5.2) were added to 5 ml of urine

ample after the addition of 100 �l of internal standard solution. Inuality control samples 20 �l of the working mixed standard solu-ion was also added. The hydrolysis was performed by adding 20 �lf helix pomatia �-glucuronidase/arylsulfatase, vortex mixed andncubated at 37 ◦C for 3 h. Samples were subsequently centrifugedt 4000 rpm during 15 min.

.4.3. Sample extraction and clean up procedureFor the evaluation of the different extraction and clean up proce-

ures, blank urines were spiked with the working standard solutionn order to obtain the target screening concentration (CC� level) forach compound. These concentrations are shown in Table 1. Thehree sample preparation procedures were performed as follows:

.4.3.1. Dilution procedure. The method was based on a previous

ork described by Kaufman et al. [10] where the hydrolysed urineas diluted 1:10 (v/v) with 5% acetonitrile and injected directly into

he chromatographic system after homogenization. In this studyamples pre-treatment was based on directly diluted 1:1 (v/v) the

screening strategy.

hydrolysed urine with purified water before the injected into thechromatographic system after homogenization.

2.4.3.2. Solid-phase extraction (SPE). Two types of solid-phaseextraction cartridges were evaluated: Bond Elut Certify and OasisHLB. The procedures used were based on previous published works[28,29]. For the Bond Elut cartridges, the hydrolysed urine wasapplied to the cartridges following the published method [28]. Theextracts were evaporated under a stream of nitrogen (40 ◦C) andfinally reconstituted with 200 �l of acetonitrile:water (20/80, v/v)containing 0.1% of acetic acid. Analyses were performed by injecting10 �l of the final extract in the UHPLC–HRMS system.

The Oasis HLB was used following the generic recommendedprotocol for a wide range of compounds published by Waters [29].In this way, cartridges were preconditioned with 5 ml of methanoland 5 ml of water. Then, the hydrolysed sample was loaded and thecartridge was washed with 5 ml of 5% methanol in water solution.The elution was carried out with 5 ml of methanol.

After the SPE clean up procedure the extracts were evapo-rated under a stream of nitrogen (40 ◦C), reconstituted with 200 �lof acetonitrile:water (20/80, v/v) containing 0.1% of acetic acid,placed in an ultrafree® MC centrifugal filter units of 0.2 �m andfinally ultracentrifuged at 11000 rpm and 4 ◦C for 3 min. The anal-yses were performed by injecting 10 �l of the final extract in theUHPLC–HRMS system.

2.4.3.3. QuEChERS methodology. QuEChERS methodologies based

on AOAC Official 2007.01 and European EN 15662 procedureswere tested, both including buffered extraction with acetonitrileand dispersive-SPE. AOAC procedure contains greater amountsof salts, primary secondary amine, C18 sorbent and MgSO4 than

58 N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65

Table 1Mass list, used internal standard, elemental composition, retention time and accurate mass data at the screening target level of the compounds analysed.

Compound name Internalstandard

Elementalcomposition

Diagnostic ion Theoreticalm/z ion (Da)

RT(min)

Recommendedconcentration(�g/l)a

Screeningtarget level orCC� (�g/l)

Average masserrorb (ppm,n = 20)

RALsBeta-zearalanol (I) C18H26O5 [M−H]− 321.1707 5.70 2 1 1.964Beta-zearalenol (I) C18H24O5 [M−H]− 319.1551 5.77 1 2.102Alpha-zearalanol (II) C18H26O5 [M−H]− 321.1707 6.09 2 1 2.226Alpha-zearalenol (II) C18H24O5 [M−H]− 319.1551 6.19 1 2.375Zearalanone (I) C18H24O5 [M−H]− 319.1551 6.83 2 1 2.634Zearalenone (MYC) (I) C18H22O5 [M−H]− 317.1394 6.88 1 2.198

Steroids4-Beta-hydroxystanozolol (VI) C21H32N2O2 [M+H]+ 345.2537 5.52 1 1.5623′-Hydroxystanozolol (VI) C21H32N2O2 [M+H]+ 345.2537 5.62 1 1.4724-Alpha-hydroxystanozolol (VI) C21H32N2O2 [M+H]+ 345.2537 5.78 1 1.280Beta-trenbolone (III) C18H22O2 [M+H]+ 271.1693 5.87 2 1 1.370Beta-boldenone (IV) C19H26O2 [M+H]+ 287.2006 5.89 1 1 1.252Beta-nortestosterone (V) C18H26O2 [M+H]+ 275.2006 6.08 1 1 1.04216-Beta-hydroxystanozolol (VI) C21H32N2O2 [M+H]+ 345.2537 6.11 2 1 1.490Boldione (VII) C19H24O2 [M+H]+ 285.1849 6.13 1 1.155Alpha-boldenone (IV) C19H26O2 [M+H]+ 287.2006 6.20 1 1 1.447Alpha-trenbolone (III) C18H22O2 [M+H]+ 271.1693 6.23 2 1 1.452Alpha-nortestosterone (V) C18H26O2 [M+H]+ 275.2006 6.34 1 1 1.266Beta-testosterone (III) C19H28O2 [M+H]+ 289.2162 6.41 1 1.354Androstenedione (IV) C19H26O2 [M+H]+ 287.2006 6.70 1 1.040Alpha-testosterone (III) C19H28O2 [M+H]+ 289.2162 6.72 1 1.309Methyltestosterone (VII) C20H30O2 [M+H]+ 303.2319 6.81 1 1.949Stanozolol (VI) C21H32N2O [M+H]+ 329.2587 8.11 2 1 1.324

StilbenesDienestrol (VIII) C18H18O2 [M−H]− 265.1234 7.14 2 1 0.681Hexestrol (VIII) C18H22O2 [M−H]− 269.1547 7.24 2 1 0.798Diethylstilbestrol (VIII) C18H20O2 [M−H]− 267.1391 8.48 1 1 1.561

�-AgonistsMetaproterenol (XVII) C11H17NO3 [M+H]+ 212.1281 2.85 10 10 1.537Zilpaterol (IX) C14H19N3O2 [M+H]+ 262.1550 3.00 1 1 1.325Terbutaline (IX) C12H19NO3 [M+H]+ 226.1438 3.03 3 3 1.632Salbutamol (XI) C13H21NO3 [M+H]+ 240.1594 3.03 1 1 1.536Cimaterol (XI) C12H17N3O [M+H]+ 220.1444 3.11 0.5 0.5 1.001Cimbuterol (X) C13H19N3O [M+H]+ 234.1601 3.52 0.5 0.5 1.590Fenoterol (VI) C17H21NO4 [M+H]+ 304.1543 3.61 1 1 1.349Ritodrine (IX) C17H21NO3 [M+H]+ 288.1594 3.72 0.5 0.5 1.325Clencyclohexerol (IX) C14H20Cl2N2O2 [M+H]+ 319.0975 3.97 1 1 1.959Hydroxymethylclenbuterol (IX) C12H18Cl2N2O2 [M+H]+ 293.0818 4.22 0.2 0.2 1.235Ractopamine (IX) C18H23NO3 [M+H]+ 302.1751 4.22 1 1 1.499Chlorprenaline (VII) C11H16ClNO [M+H]+ 214.0993 4.28 1 2.399Clenproperol (X) C11H16Cl2N2O [M+H]+ 263.0712 4.31 0.5 0.5 1.277Formoterol (IX) C19H24N2O4 [M+H]+ 345.1809 4.47 2 2 2.683Tulobuterol (IX) C12H18ClNO [M+H]+ 228.1150 4.51 0.2 0.2 1.712Clenbuterol (X) C12H18Cl2N2O [M+H]+ 277.0869 4.58 0.2 0.2 1.119Chlorbrombuterol (X) C12H18BrClN2O [M+H]+ 321.0364 4.70 0.2 0.2 0.994Brombuterol (X) C12H18Br2N2O [M+H]+ 364.9859 4.81 0.2 0.651Isoxsuprine (X) C18H23NO3 [M+H]+ 302.1751 4.93 0.5 0.5 1.365Clenpenterol (X) C13H20Cl2N2O [M+H]+ 291.1025 4.99 0.5 0.2 1.978Mabuterol (XVI) C13H18ClF3N2O [M+H]+ 311.1133 5.09 0.2 0.2 0.453Clenisopenterol (VII) C13H20Cl2N2O [M+H]+ 291.1025 5.48 1 1 2.013Mapenterol (IX) C14H20ClF3N2O [M+H]+ 325.1289 5.62 0.2 0.2 1.399Clenhexerol (IX) C14H22Cl2N2O [M+H]+ 305.1182 6.32 1 1.631Salmeterol (IV) C25H37NO4 [M+H]+ 416.2795 7.27 1 0.5 0.949

SedativesAtenolol (XI) C14H22N2O3 [M+H]+ 267.1703 3.04 1 1.572Xylazine (X) C12H16N2S [M+H]+ 221.1107 4.56 0.5 2.181Azaperol (X) C19H24FN3O [M+H]+ 330.1976 4.63 0.5 1.489Azaperone (XI) C19H22FN3O [M+H]+ 328.1820 5.04 1 1.731Carazolol (IX) C18H22N2O2 [M+H]+ 299.1754 5.18 0.5 1.633Haloperidol metabolite II (IX) C21H25ClFNO2 [M+H]+ 378.1631 5.89 0.5 1.897Acepromazine (X) C19H22N2OS [M+H]+ 327.1526 6.22 0.5 1.197Haloperidol (X) C21H23ClFNO2 [M+H]+ 376.1474 6.34 0.5 1.770Propionylpromazine (IX) C20H24N2OS [M+H]+ 341.1682 7.03 0.5 1.540Chlorpromazine (XII) C17H19ClN2S [M+H]+ 319.1030 7.58 3 1.497

NitroimidazolesHydroxymetronidazole (XIII) C6H9N3O4 [M+H]+ 188.0666 2.34 2 1.322HMMNI (XII) C5H7N3O3 [M+H]+ 158.0560 2.63 2 2.031Metronidazole (XIII) C6H9N3O3 [M+H]+ 172.0717 2.79 1 1.638Ronidazole (XIII) C6H8N4O4 [M+H]+ 201.0618 2.91 2 1.768Dimetridazole (XIV) C5H7N3O2 [M+H]+ 142.0611 3.20 1 2.293Hydroxyipronidazole (XV) C7H11N3O3 [M+H]+ 186.0873 3.70 1 2.199Ipronidazole (XV) C7H11N3O2 [M+H]+ 170.0924 4.51 1 2.143

N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65 59

Table 1 (Continued)

Compound name Internalstandard

Elementalcomposition

Diagnostic ion Theoreticalm/z ion (Da)

RT(min)

Recommendedconcentration(�g/l)a

Screeningtarget level orCC� (�g/l)

Average masserrorb (ppm,n = 20)

CorticosteroidsPrednisolone (I) C21H28O5 [M+CH3COO]− 419.2075 4.96 1 2.303Methylprednisolone (I) C22H30O5 [M+CH3COO]− 433.2232 5.26 0.2 2.276Dexamethasone (XVI) C22H29FO5 [M+CH3COO]− 451.2138 5.36 2 0.4 1.759Betamethasone (XVI) C22H29FO5 [M+CH3COO]− 451.2138 5.36 2 0.4 1.759Flumethasone (II) C22H28F2O5 [M+CH3COO]− 469.2043 5.40 0.2 1.918

NSAIDsNaproxen (XIV) C14H14O3 [M−H]− 229.0870 6.16 10 1.167Oxyphenylbutazone (XVII) C19H20N2O3 [M−H]− 323.1401 6.21 5 2.554Phenylbutazone (XVII) C19H20N2O2 [M−H]− 307.1452 8.25 5 2.852Mefenamic acid (XVII) C15H15NO2 [M−H]− 240.1030 9.12 5 1.197

AmphenicolsFlorfenicol amine (XII) C10H14FNO3S [M+H]+ 248.0751 2.75 0.2 0.875Thiamphenicol (XVIII) C12H15Cl2NO5S [M−H]− 353.9975 3.53 0.2 1.966Florfenicol (XVIII) C12H14Cl2FNO4S [M−H]− 355.9932 4.30 0.2 0.857Chloramphenicol (XVIII) C11H12Cl2N2O5 [M−H]− 321.0051 4.57 0.3 0.2 2.324

Thyreostats2-Thiouracil (XV) C4H4N2OS [M+H]− 126.9972 1.17 10 10 2.644Tapazol (XVII) C4H6N2S [M+H]+ 115.0324 2.59 10 10 2.1466-Propyl-2-thiouracil (XII) C7H10N2OS [M−H]− 169.0441 3.25 10 10 2.0712-Mercaptobenzimidazole (XV) C7H6N2S [M+H]+ 151.0324 3.45 10 1.5866-Phenyl-2-thiouracilL (XVII) C10H8N2OS [M−H]− 203.0285 3.68 20 2.548

OthersDapsone (IX) C12H12N2O2S [M+H]+ 249.0692 4.10 1 1.214Monoacetyldapsone (VIII) C14H14N2O3S [M−H]− 289.0652 4.34 5 0.565

Internal standardsBeta-zearalanol-D4 (I) C18D4H22O5 [M−H]− 325.1959 5.70 5 2.632Alpha-zearalanol-D4 (II) C18D4H22O5 [M−H]− 325.1959 6.07 5 2.452Beta-boldenone-D3 (III) C19D3H23O2 [M+H]+ 290.2194 5.89 5 1.281Beta-trembolone-D3 (IV) C18D3H19O2 [M+H]+ 274.1881 5.83 5 1.465Beta-nortestosterone-D3 (V) C18D3H23O2 [M+H]+ 278.2194 6.08 5 1.42416-Beta-hydroxystanozolol-D3 (VI) C21D3H29N2O2 [M+H]+ 348.2725 6.11 5 1.401Methyltestosterone-D3 (VII) C20D3H27O2 [M+H]+ 306.2507 6.79 5 1.362Diethylstilbestrol-D6 (VIII) C18D6H14O2 [M−H]− 273.1767 8.43 5 1.807Ractopamine-D5 (IX) C18D5H18NO3 [M+H]+ 307.2065 4.19 5 1.641Clenbuterol-D9 (X) C12D9H9Cl2N2O [M+H]+ 286.1434 4.54 5 1.583Atenolol-D7 (XI) C14D7H15N2O3 [M+H]+ 274.2143 3.04 5 1.315Chlorpromazine-D6 (XII) C17D6H13ClN2S [M+H]+ 325.1407 7.58 5 0.609Ronidazole-D3 (XIII) C6D3H5N4O4 [M+H]+ 204.0807 2.93 5 1.588Dimetridazole-D3 (XIV) C5D3H4N3O2 [M+H]+ 145.0799 3.18 5 1.354Ipronidazole-D3 (XV) C7D3H8N3O2 [M+H]+ 173.1112 4.49 5 0.340Dexamethasone-D4 (XVI) C22D4H25FO5 [M+H]+ 455.2389 5.35 5 1.963Phenylbutazone-D10 (XVII) C19D10H10N2O2 [M−H]− 317.2080 8.15 5 1.754Chloramphenicol-D5 (XVIII) C11D5H7Cl2N2O5 [M−H]− 326.0364 4.57 5 2.457

s obta

tetsi

2

apieppe

lw

b

a CRL guidance paper of 7th December 2007.b Average mass error (absolute values) calculated from the individual mass error

he EN procedure. Three combinations were tested: (Q1) ENxtraction salt packet + AOAC fatty dispersive-SPE; (Q2) EN extrac-ion salt packet + EN fatty dispersive-SPE and (Q3) EN extractionalt packet. The three QuEChERS based procedures are givenn Fig. 2.

.5. Matrix effect and recovery studies

The evaluation of matrix effect and recovery was conductedccording to the strategy developed by Matuszevski [30]. For thisurpose, three different sets (A, B and C) were prepared by spik-

ng urine samples at CC� level and processed using the optimizedxtraction of the method described (Q1). Matrix effect (ME%) androcedure recovery (RE%) of the methods were evaluated by com-aring the absolute peak areas of the three sets. The procedure forach set is described as follows:

Set A: consisted of six replicates of a standard solution at CC�

evel, including internal standards, in mobile phase. Ten microliters

ere injected into the UHPLC–HRMS system.Set B: consisted of six replicates of blank urine extracts obtained

y the optimized sample preparation spiked, after extraction and

ined at the screening target level in the validation.

clean-up, with 200 �l of the standards solution used on set A. Thespiked extracts were ultra centrifuged and placed into vials and10 �l injected into the UHPLC–HRMS system.

Set C: consisted of six replicates of blank urine spiked at threelevels before the extraction procedure. The samples were processedby the optimized sample preparation and finally 10 �l injected intothe UHPLC–HRMS system.

From the results obtained ME and RE values were calculated asfollows:

ME (%) = B

A× 100

RE (%) = C

B× 100

where A is the mean peak area obtained from the set A. B and C

are the mean peak areas obtained from the set B and the set C,respectively. Also, corrected ME and corrected RE were calculatedby using, area ratios (area analyte/area internal standard) insteadof area analyte.

60 N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65

three

2

iGamtm(ti

Diwtiismcfgm

rst±

Fig. 2. Extraction protocol for the

.6. Validation process

The validation procedure was carried out based on the spec-fications of the Commission Decision 2002/657/EC [31], the CRLuidance paper of 7th December 2007 on the state of the art ofnalytical methods for national residue control plans [32] and,ore specifically, on the recommendations of the Guideline for

he validation of screening methods for residues of veterinaryedicines proposed by the Community Reference Laboratories

CRLs) in 2010 [33]. The definitions of the terms used in the valida-ion framework established by the European Union are summarizedn Supplementary data.

As established for screening methods, Selectivity/Specificity andetection Capability (CC�) were evaluated. Regarding Selectiv-

ty/Specificity, 20 bovine urine samples from different animalsere analysed to check for any interferences at the expected reten-

ion times of the compounds and to estimate the effect of thenterferences. To calculate the Detection Capability (CC�) or screen-ng target level and identify the cut-off level, the 20 blank urineamples were spiked at a concentration lower than the recom-ended [32] and the false compliant result rate was studied to

heck if a value <5% was obtained. Threshold value (T) and cut-offactor (Fm) were also determined for each analyte following theuideline approach in order to check if Fm > T and Fm > B (B is theean response of the blank samples).Confirmation of analytes was accomplished on the basis of the

elative retention time (RRT) calculated for corresponding internaltandard and the accurate mass. For analyte identification massolerance and retention time deviation should be within 5 ppm and2.5%, respectively.

QuEChERS based methods tested.

2.7. Quality control and cut-off level

In each analytical batch a quality control sample (blank fortifiedat CC� level) was used to establish the cut-off level and conse-quently to check if the response of the samples is at or above thislevel. The cut-off level was calculated as the ratio of the response ofthe analyte to the response of the corresponding internal standard.

3. Results and discussion

3.1. UHPLC separation and detection by Orbitrap/HRMS

A total of 87 target substances were included in theUHPLC–HRMS method to perform a daily routine screening ofbanned/unauthorized substances in bovine urine. In order to pro-duce a fast and efficient chromatographic method, an UHPLCsystem was used. A column packed with sub-2 �m was selected,because this sort of columns lead to significant improvementsin terms of efficiency and shorter chromatographic runs whencompared with conventional columns packed with 3 �m or 5 �mparticles.

The compounds analysed covered a wide spectrum of polarities,including very polar substances such as nitroimidazoles, thyreo-statics or some �-agonists to non-polar ones such as stilbens orsome NSAIDs. For the chromatographic analysis of all the com-pounds tested a column with high retention capacity of the most

polar molecules was selected (Hypersil GOLD aQ, polar endcappedC18 phase), providing a total run time of 13.5 min (including 3 minof equilibration time). A C18 column (Hypersil Gold) was also testedbut poor retention of the most polar compounds was obtained.

N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65 61

nction

Fftllpws

maoatmtaae

cstFotg

ya(

maacaibeIm[

(5ss

eu

Fig. 3. Identification of the analytes (%) as a fu

urthermore, the Hypersil GOLD aQ has shown better peak shapeor some of the analytes. Retention times reported in Table 1 provedo be very stable, with relative standard deviation of the mean atess than 0.5% for all compounds. The peak widths at the base-ine ranged from 12 to 20 s which permits more than twelve dataoints through the elution profile of all substances if the analyzerorks with a resolution of 50,000 FWHM (2 Hz) and with polarity

witching at a fast scan rate.Moreover, different compositions were investigated for the

obile phase including acetonitrile or methanol as organic phasend water with formic acid, acetic acid, and ammonium acetater ammonium formiate as aqueous phase. The presence of formiccid or ammonium formiate in the mobile phase inhibited ioniza-ion of stilbens both with methanol and acetonitrile. With the use of

ethanol, most of the epimers were not resolved enough. The mix-ure of ammonium acetate 10 mM – acetonitrile and acetic acid 0.1%cetonitrile provided good results in all cases. Finally, 0.1% aceticcid aqueous solution with acetonitrile was selected because it isasier to prepare. The gradient is described in Section 2.

With the chromatographic conditions selected, the epimericompounds betametasone and dexametasone were not completelyeparated, so an alternative gradient was optimized to enableheir adequate separation in the case of the suspected sample.ig. S1, supplementary data shows an example of a chromatogrambtained for the analysis of these substances where it can be notedhat the two isomers are separated by the alternative chromato-raphic gradient employed at a flow rate of 400 �l/min.

For the detection of the analytes in the Orbitrap ExactiveTM anal-ser the elemental composition and the monoisotopic mass of eachdduct, both in positive ([M+H]+) and negative ionization mode[M−H]−, [M+CH3COO]−) were previously calculated (Table 1).

The determination of accurate mass (AM) is used to confirm ele-ental formula of compounds and to detect the presence of the

nalytes in a screening method. When no isobaric interferencesre present, resolving power does not affect the accuracy and pre-ision of the AM measurements. However, the analysis of residuesnd contaminants in food safety is one of the fields where themportance of the high resolving power has been demonstratedecause the probability of having overlapping isobaric interfer-nces is extremely high, avoiding a correct mass assignment [34].n these cases, the resolving power is a key parameter for a correct

ass assignment and consequently avoiding false negative results27,35].

The Orbitrap was operated at a resolution of 50,000 FWHM2 Hz). The chosen resolution was selected in combination with a

ppm mass extraction window, in order to define an appropriatecreening strategy that provides good sensitivity and appropriate

electivity. This strategy is detailed in Fig. 1.

To minimize the number of false positives a narrow mass-xtraction window and low tolerance for mass accuracy needs to besed. However, for screening methods it is essential to avoid false

of the different extraction procedures tested.

negative results, which could be achieved when very good massaccuracy (<5 ppm) is ensured in the real samples. As can be seen inTable 1, the measurements performed at R = 50,000 showed goodmass accuracy (<5 ppm) for blank urines fortified at CC� level.

As an example of the suitability of these settings to avoid falsenegative, Fig. S2, supplementary data, presents the full-scan massspectra and the extracted ion chromatograms with 5 ppm extrac-tion window of 2-thiouracil (a) and formoterol (b) from a CC�level spiked sample acquired at two resolutions 10,000 FWHMand 50,000 FWHM. The mass deviation at a resolution of 10,000is higher than the 5 ppm extraction window due to the presenceof unidentified interferences, giving a false negative result. How-ever, these analytes can be detected with a mass deviation less than5 ppm at 50,000 FWHM leading to a confident identification. It hasbeen reported that the selectivity of the data acquired at 50,000FWHM in HRMS full scan mode exceeds the use of two transitionsin low resolution mass spectrometry (LC–MS/MS methods) [21].

3.2. Extraction and clean-up procedure

One of the aims of this work was to develop amultiresidue–multiclass extraction method for the simulta-neous determination of several classes of veterinary drugs. For thispurpose, the performance of three extraction procedures (dilution,SPE, QuEChERS) was assessed, based on the number of extractedand identified analytes.

Fig. 3 shows the percentage of identified compounds when thedifferent extraction/clean-up procedures were used at the CC�level. In the case of the diluted method, only 32% of analytes wereidentified with this approach. RALs, NSAIDs, tranquillizers, thyre-ostats and amphenicols were correctly identified but analytes suchas 16-�-OH-stanozolol, hexestrol, ritodrine, prednisolone, metil-prednisolone and tapazol have not been detected and identified.Although dilution is a simple and attractive sample preparation,this approach seems to be limited to the detection of banned sub-stances at low concentrations.

On the other hand, SPE procedures gave satisfactory resultsusing Bond Elut Certify (C8 sorbent + strong cation exchanger)for steroids, RALs, stilbens, tranquillizers, corticosteroids andNSAIDs but not for �-agonists, nitroimidazoles, amphenicols,thyreostatics and dapsone (47% of identification). OASIS HLBcartridges (hydrophilic–lipophilic balance sorbent) provided sat-isfactory results for most of the groups except nitroimidazoles,NSAIDs, thyreostats and dapsone (70% of identification). Theseresults can be explained due to the lower retention capacity of theC8 material compared with the HLB sorbent which is able to retainpolar and non-polar compounds.

Finally, the results of the combination of QuEChERS methodolo-gies show that the Q1 procedure is the optimum method because allthe targeted compounds are correctly detected and identified. Thegreater amount of primary secondary amine and C18 sorbents to

62 N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65

F ion (cos

rod

Hgo(w

3

mbb

m[

3

p[atie8c

pwtMswslwmwpt

ig. 4. Matrix effect and recovery before correction (ME% and RE%) and after correcttudied (see Fig. S3 for all compounds).

etain matrix components in this approach (Q1) led to a decreasef ion suppression phenomenon and an improvement in analyteetection.

Conventional QuEChERS implies a dispersive-SPE clean-up step.owever some authors have avoided it because clean extracts andood responses were obtained, achieving a fast methodology. Inur work fewer compounds were identified by using Q3 extractionwithout the dispersive-SPE step) (92%), as no matrix componentsere removed.

.3. Matrix effect and absolute recovery

It is well known that in liquid chromatography coupled withass spectrometry the matrix effect is an important issue that must

e evaluated and discussed in the context of method developmentefore studying its performance characteristics.

In LC–MS/MS the ion suppression and the signal enhancementay be caused by different factors as reported by different authors

36,37].

.3.1. Matrix effectThe results obtained in the matrix effect study have been inter-

reted taking into account the study conducted by Ferrer et al.38]. The matrix effect was classified into three different categoriesttending to the calculated values. There was no matrix effect whenhe ME factor was between 80% and 120%, because the repeatabil-ty of the results would be close to this range. A medium matrixffect was considered when the values ranged between 40% and0% or 120% and 150%. A percentage below 40% or above 150% waslassified as a high matrix effect.

Some of the results achieved including internal standards areresented in Fig. 4 and Fig. S3, supplementary data. In this figurese can appreciate that the matrix effect varies remarkably from 6

o 113%. For the 21% of the analytes no matrix effect was found.ost of these analytes were steroids, RAL’s, some nitroimidazoles,

almeterol, naproxen and flumethasone. A medium matrix effectas obtained for the 45% of compounds. This category includes

tilbens, most of the beta-agonists, nitroimidazoles and tranquil-izers, and other compounds. In contrast a high signal suppression

as observed both for the most polar and smaller analytes like

etaproterenol, 6-phenyl-2-thiouracil and 6-propyl-2-thiouracil,ith ME values of 6, 7 and 14% respectively. This high ion sup-ression can be explained by different causes. On the one hand,he smaller molecules can undergo this matrix effect due to the

rrected ME% and corrected RE%) for some of the compounds and internal standards

post-interface signal suppression [20,34,35]. On the other hand,the most polar analytes were more susceptible to undergo ion sup-pression since they present shorter retention times, and so theycan elute nearly to the elution front along with many other matrixcomponents.

3.3.2. RecoveryThe RE factor was also investigated. The variability of the recov-

eries calculated for all compounds is high such as in the ME factorstudy. In Fig. 4 and Fig. S3, supplementary data, it can be observedthat the 55% of the analytes give a RE value around or above 60%.However the method gives low recoveries for some compoundslike oxyphenylbutazone, tapazol or 6-phenyl-2-thiouracil with val-ues between 4 and 9%. These results may originate from severalcauses. High polar compounds probably suffer poor extraction fromthe aqueous to the acetonitrile layer and/or the analytes could beretained in the stationary phase. These hypotheses were not theaim of this study and will be investigated in further work.

3.3.3. Overcoming matrix effect and recoveryCommonly, the use of isotopically labelled internal standard

(I.S.), matrix matched standards calibration and/or dilution of thefinal extracts ensure the correction of these issues.

Due to the limitation on sensitivity, the dilution of final extractshas been discarded. On the other hand, we need to calculate if thedetected substances provide a response above the cut-off level, forthis quantitation is very useful to employ internal standards. In thisstudy we used 18 internal standards. The choice of the appropriateI.S. for each compound was a difficult task and was performed tak-ing into account the ME and RE values obtained, for both analytes ofinterest and I.S. According to this, the selected I.S. (Table 1) shouldpresent ME and RE values similar to the compound for which itattempts to correct.

The corrected ME% and RE% values (Fig. 4 and Fig. S3) show thatthe 93% of the analytes do not present a matrix effect as the cor-rected values are within the range of 80–120% established beforefor this category. The other analytes present medium matrix effectswith values ranging from 43% (florfenicol) and 75% (flumethasone),except the 6-propyl-2-thiouracil with a 33% of matrix effect. In thecase of corrected RE%, only naproxen, 2-mercaptobenzimidazole,

and 6-phenyl-2-thiouracil have recoveries below 60%.

From these results we can conclude that the methodcould be improved by using isotopically labelled internal stan-dard for the most problematic analytes. For the analysis of

N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65 63

F w of 5u

6tn2wiby

3

atsto

saw

tsesmsfis

ig. 5. UHPLC–HRMS extracted ion chromatograms (narrow mass tolerance windorine sample spiked at the screening target level.

-propyl-2-thiouracil and florfenicol with high and medium ME%,he corrected recoveries are 86 and 90% respectively, so it wouldot be necessary to use the isotopically labelled I.S. In contrast for-mercaptobenzimidazole, 6-phenyl-2-thiourcil and cimbuterolith medium matrix effect, their corresponding corrected recover-

es are unfavourable with values below 60%. In these cases it woulde advisable to use the isotopically labelled I.S. in quantitative anal-sis.

.4. Validation of results

A complete validation was carried out for the analytes with screening target concentration level equal or below the 50% ofhe regulatory limit (Table 1). However, for the substances with acreening target concentration ranging between 50% and 100% ofhe regulatory limit, a rough validation was performed (20 insteadf 40–60 samples [33]).

After the analysis of 20 samples, no signal were detected at theame retention time of the studied analytes, which indicates thatt 50,000 FWHM the method presented an appropriate specificity,ithout interferences from the matrix components.

In order to establish the appropriate spiked levels to calculatehe CC� for each substance, these criteria were followed: (a) forubstances with recommended concentration (RC), different lev-ls of fortification < RC (e.g., ½ RC or lower) were checked; (b) forubstances without RC, the regulated levels established for other

atrices were taken into account. In all cases, the final CC� were

elected after checking that a mass assignment <5 ppm was ful-lled. As an example, Fig. 5 shows a chromatogram of 10 substancespiked at CC� level.

ppm) of different compounds from diverse groups of veterinary drugs of a bovine

The results of B, T and Fm obtained from the validation of theCC� are presented in Table SI. The criteria Fm > B established in theCommission Decision 2002/657/EC was achieved for all analytes.Likewise, Fm was lower than T for all the substances of the study,which means that the rate of false positive is below 5%.

The CC� summarized in Table 1 were similar or lower thanthose obtained by Leporati et al. [11]. In this work 37 illicit sub-stances from different families (anabolic steroids, corticosteroids,promazines and �-agonists) in urine were detected by a LC–MS/MSmethod with detection limits ranged between 1 and 3 ng/ml.

3.5. Application to real samples: non-compliant results

The proposed analytical method was applied to 58 real urinesamples from Valencia Region (Spain) to monitor the content ontargeted analytes. The samples were collected at slaughterhousesand from live animals in farms. During the period of monitor-ing, carazolol, ractopamine, chloramphenicol and 2-thiouracil werefound in different samples which were considered as suspect, astheir area ratio (area analyte/area internal standard) was higherthan the one obtained for the quality control at CC� level.

The suspected sample of carazolol was analysed by making useof the developed methodology and HRMS. The Orbitrap instrumentoperated simultaneously in both full scan and “all ions fragmen-tation” (AIF) mode using HCD at 25 eV in positive mode. Fig. 6shows the accurate mass extracted chromatograms of the par-ent ion (without fragmentation) and the characteristic fragments

obtained (with positive HCD) employing a 5 ppm mass window. Thetheoretical and experimental isotopic pattern (12C/13C) of the cara-zolol was also shown and compared, which probes the presence ofthis analyte in the sample.

64 N. León et al. / J. Chromatogr. A 1258 (2012) 55– 65

F ion wi l patt

4

wbbalats

aobCe

isttl

mamb

A

p

ig. 6. Positive sample to carazolol. (a) Extracted ion chromatograms of the nativenjected with positive HCD fragmentation at 25 eV. (c) Theoretical and experimenta

. Conclusions

A comprehensive screening method based on UHPLC–HRMSas successfully developed and validated for a wide-range of

anned and non-regulated veterinary drugs in bovine urine. Touild-up a satisfactory screening strategy with suitable sensitivitynd selectivity, the mass spectrometer parameters were optimizedeading to the conclusion that a resolving power of 50,000 FWHM,n AGC of 106 ions, and an extraction windows of 5 ppm providehe most sensitive, selective and accurate measurements for theseubstances in urine.

In order to select an appropriate sample preparation method for wide-range screening, three generic sample preparation method-logies were studied. The QuEChERS method was selected as theest as it enables the detection and identification of drugs withC�s lower or equal than the recommended values established forach substance.

The method was validated based on the EU criteria for screen-ng methods (2002/657/EC) and the Guideline for the validation ofcreening methods proposed by the Community Reference Labora-ories. In all cases, the CC� established levels were equal or lowerhan the recommended concentrations established by EU referenceaboratories.

From our point of view, the use of high-resolution liquid chro-atography combined with high-resolution mass spectrometry is

powerful and reliable tool for the identification of substances inulti-class multi-residue analysis, and could be used on a routine

asis in the food safety official laboratories.

cknowledgements

This work has been developed under the financial sup-ort of Conselleria de Sanitat (Generalitat Valenciana), Spain

ithout HCD fragmentation. (b) Native ion and characteristic fragments of carazololern 12 C/13 C of carazolol.

(Ref: 018/2011). The authors would like to thank Thermo FischerScientific for their assistance and technical support.

Appendix A. Supplementary data

Supplementary data associated with this article can befound, in the online version, at http://dx.doi.org/10.1016/j.chroma.2012.08.031.

References

[1] Council Directive 96/22/EC of 29 April 1996 concerning the prohibition on theuse in stock farming of certain substances having a hormonal or thyreostaticaction and beta-agonists, and repealing Directives 81/602/EEC and 88/146/EECand 88/299/EEC, Official Journal of the European Union, L 22.

[2] Council Directive 96/23/EC of 29 April 1996 on measures to monitor cer-tain substances and residues thereof in live animals and animal products andrepealing Directives 85/358/EEC and 86/469/EEC and Decisions 89/187/EEC and91/664/EEC, Official Journal of the European Union, L 125.

[3] Council Regulation 470/2009 of 6 may 2009 laying down Community pro-cedures for the establishment of residue limits of pharmacologically activesubstances in foodstuffs of animal origin, repealing Council Regulation (EEC)No 2377/90 and amending Directive 2001/82/EC of the European Parliamentand the Council and Regulation (EC) No 726/2004 of the European Parliamentand the Council, Brussels, Belgium, Official Journal of the European Union, L152 (2010).

[4] Council Regulation 37/2010/EC of 22 December of 2009 on pharmacologicallyactive substances and their classification regarding maximum residues limits infoodstuffs of animal origin, Brussels, Belgium, Official Journal of the EuropeanUnion, L 15 (2010).

[5] R.J.B. Peters, A.A.M. Stolker, J.G.J. Mol, A. Lommen, E. Lyris, Y. Angelis, A. Von-aparti, M. Stamou, C. Georgakopoulos, M.W.F. Nielen, Trends Anal. Chem. 29(2010) 1250.

[6] Commission staff working document on the implementation of national

residue monitoring plans in the member states in 2009, Brussels, Belgium.

[7] A.A.M. Stolker, U.A.Th. Brinkman, J. Chromatogr. A 1067 (2005) 15.[8] A.A.M. Stolker, T. Zuidema, M.W.F. Nielen, Trends Anal. Chem. 26 (2007) 967.[9] D. Courttheyn, B. Le Bizec, G. Brambilla, H.F. De Brabander, E. Cobbaert, M. Van

de Wiele, J. Vercammen, K. De Wasch, Anal. Chim. Acta 473 (2002) 71.

atogr.

[

[

[

[

[

[

[

[[[

[

[

[

[

[

[[

[

[

[

[[

[

[

[

[

[

N. León et al. / J. Chrom

10] A. Kaufmann, P. Butcher, K. Maden, M. Widmer, Anal. Chim. Acta 586 (2007)13.

11] M. Leporati, P. Capra, P. Brizio, V. Ciccotelli, M.C. Abete, M. Vicentini, J. Sep. Sci.35 (2012) 400.

12] B. Kinsella, J. O’Mahony, E. Malone, M. Moloney, H. Cantwell, A. Furey, M. Dana-her, J. Chromatogr. A 1216 (2009) 7977.

13] S.J. Lehotay, K.A. Son, H. Kwon, U. Koesukwiwat, W. Fu, K. Mastovska, E. Hoh,N. Leepipatpiboon, J. Chromatogr. A 1217 (2010) 2548.

14] S.W. Lee, J.-H. Choi, S.-K. Cho, H.-A. Yu, A.M.A. El-Aty, J.-H. Shim, J. Chromatogr.A 1218 (2011) 4366.

15] M.M. Aguilera, J.L. Martínez, R. Romero-González, A. Garrido, J. Chromatogr. A1025 (2008) 10.

16] M. Fiori, C. Civitareale, S. Mirante, E. Magarò, G. Brambilla, Anal. Chim. Acta 529(2005) 207.

17] E.N.M. Ho, D.K.K. Leung, T.S.M. Wan, N.H. Yu, J. Chromatogr. A 1120 (2006) 38.18] G. Kaklamanos, G. Theodoridis, T. Dabalis, J. Chromatogr. B 877 (2009) 8059.19] A. Tölgyesi, V. Sharma, S. Fekete, J. Fekete, A. Simon, S. Farkas, J. Pharmaceut.

Biomed. Anal. 64–65 (2012) 40.20] A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Talanta 85 (2011)

991.21] A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Anal. Chim. Acta 673

(2010) 60.22] A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Anal. Chim. Acta 700

(2011) 86.23] R. Romero-Gonález, M.M. Aguilera-Ruiz, P. Plaza-Bolanos, A. Garrido, J.L.

Martínez, J. Chromatogr. A 1218 (2011) 9353.24] Q. Hu, R. Noll, H. Li, A. Makarov, M. Hardman, G. Cooks, J. Mass Spectrom. 40

(2005) 430.

[

[

A 1258 (2012) 55– 65 65

25] R. Perry, R. Cooks, R. Noll, Mass Spectrom. Rev. 27 (2008) 661.26] A. Kaufmann, P. Butcher, K. Maden, M. Widmer, J. Chromatogr. A 1194 (2008)

66.27] A. Kaufmann, P. Butcher, K. Maden, S. Walker, M. Widmer, Analyst 136 (2011)

1898.28] G.N.W. Leung, E.W. Chung, E.N.M. Ho, W.H. Kwok, D.K.K. Leung, F.P.W. Tang,

T.S.M. Wan, N.H. Yu, J. Chromatogr. B 825 (2005) 47.29] Waters’ Recommended protocol for Solid Phase Extraction (SPE) obtained from

http://www.waters.com/webassts/cms/library/docs//lcSP.pdf.30] B.K. Matuszevski, J. Chromatogr. B 830 (2006) 293.31] 17 August 2002, Commission Decision 2002/657/EC of 12 August 2002 con-

cerning the performance of analytical methods and the interpretation ofresults, Brussels, Belgium, Official Journal of the European Communities, L 221(2002).

32] CRL Guidance paper of 7th December 2007, CRLs view on state of the art ana-lytical methods for national residue control plans, p. 1.

33] CRL Guidelines of 20th January 2010 for the validation of screening methodsfor residues of veterinary medicines, p. 1.

34] M. Kellman, H. Muenster, P. Zomer, H. Mol, J. Am. Soc. Mass Spectrom. 20 (2009)1464.

35] A. Makarov, E. Denisov, O.J. Lange, J. Am. Soc. Mass Spectrom. 17 (2006)977.

36] A.V. Eeckhaut, K. Lanckmans, S. Sarre, I. Smolders, Y. Michotte, J. Chromatogr.

B 877 (2009) 2198.

37] F. Gosetti, E. Mazzucco, D. Zampieri, M.C. Gennaro, J. Chromatogr. A 1217 (2010)3929.

38] C. Ferrer, A. Lozano, A. Agüera, A. Jimenez, A.R. Férnandez-Alba, J. Chromatogr.A 1218 (2011) 7634.