Me

VL

a

ARRAA

KSAFPHLs

1

pamcuaps[oem[lc

0d

Journal of Chromatography A, 1209 (2008) 162–173

Contents lists available at ScienceDirect

Journal of Chromatography A

journa l homepage: www.e lsev ier .com/ locate /chroma

ulti-class determination of antimicrobials in meat by pressurized liquidxtraction and liquid chromatography–tandem mass spectrometry

anesa Carretero, Cristina Blasco, Yolanda Picó ∗

aboratori de Bromatologia i Toxicologia, Facultat de Farmàcia, Universitat de València, Av. Vicent Andrés Estellés s/n, 46100 Burjassot, València, Spain

r t i c l e i n f o

rticle history:eceived 23 June 2008eceived in revised form 31 August 2008ccepted 4 September 2008vailable online 7 September 2008

eywords:creeningntimicrobialsoodressurized liquid extraction

a b s t r a c t

A multi-residue method using pressurized liquid extraction (PLE) and liquid chromatography–tandemmass spectrometry (LC–MS/MS) has been developed for determining trace levels of 31 antimicrobials,including �-lactams, lincosamides, macrolides, quinolones, sulfonamides, tetracyclines, nitroimidazolesand trimethoprim. The extraction method required pre-homogeneization of the meat with EDTA-washedsand and subsequent one-static-cycle extraction for 10 min with 40 ml of water at 1500 psi and 70 ◦C. Theeffect of operation temperature, pressure, flush volume, and static cycles on PLE performance was stud-ied. Average recoveries ranged from 75 to 99% with relative standard deviations

atog

rb

aafimawewaiipctbowrao

tanhsaaaamlliff

2

2

ftqtnsprtfihTdTrtosp

icorcrtfo

(DfUVoawsa

2

kSdDwtawa

tscmpmb

bToaspla

tptrev2a

V. Carretero et al. / J. Chrom

esidues and other contaminants [5]. However, these methods haveeen scarcely reported [18–22].

Recently, Bogialli et al. [23–25] developed a simple, rapidnd sensitive confirmatory method for determining sulfonamidentimicrobials applicable to different matrices, such as cattle andsh muscle tissues, bovine liver and kidney, milk and eggs. Theethod is based on the pressurized liquid extraction (PLE) usinglaboratory hand-made system and heated water as extractant,hich offers as distinct advantage the elimination of toxic and

xpensive solvents. Since then, several methods based on PLE usingater have been tested for the determination of ampicillin and

moxycillin in bovine tissues and milk [26], sulfonamide residuesn raw meat, infant foods and cheese [27–29], aminoglycosidesn bovine milk [30], tetracycline antibiotics in bovine, swine, andoultry muscle tissues and in cheese [31,32], macrolides and lin-omycin in bovine milk and yoghurt [33] and quinolones in bovineissues and milk [34,35]. These methods have differential aspectsecause they have always been developed and optimized for onlyne class of antimicrobials with structural similarities. However,hen looking at the common procedural scheme, the analyst might

easonably suppose that it would be useful and presently feasible tochieve consensus conditions able to acceptably recover all classesf antimicrobials.

This study affords, for the first time, the development of a selec-ive, sensitive and consistent multi-class method for determiningntimicrobial residues in meat based on PLE and LC–MS/MS. A sig-ificant improvement has been attained by replacing laboratoryand-made PLE systems by Dionex commercial ASE, which greatlyimplifies and accelerates sample extraction, as well as offers somedvantages in terms of reliability and efficiency. The investigatedntimicrobials were selected from those that are currently avail-ble in the European market for the treatment and prevention ofnimal microbial infections. Their structures are shown in Fig. 1. Theethod was validated by evaluating a set of parameters, including

inearity, decision limit (CC�), detection capability (CC�), detectionimit (LOD), quantification limit (LOQ), recovery, precision, selectiv-ty, stability and ruggedness. Confirmation and quantitation resultsor 25 incurred bovine and porcine meat samples, which were takenrom local markets and slaughterhouses, are also reported.

. Experimental

.1. Reagents and materials

Thirty-one antimicrobials (see Fig. 1 and Table 1) were obtainedrom Sigma–Aldrich Logistik (Schnelldorf, Germany) of >95% cer-ified purity. Dapsone was used as internal standard (I.S.) foruantification purposes because it is not allowed in the livestockreatment [7]. Ronidazole, one of the target analytes, is listed ason-authorized product too. Although, theoretically, there is theame chance for finding dapsone or ronidazole incurred samples,ractically, the rapid alert system for food and feed (RASFF) annualeport pointed out that nitrofurans and its metabolites representhe biggest proportion of notifications [36]. Dapsone was not noti-ed and presents other advantages to be used as I.S., such asigher absolute recoveries and an intermediate retention time.he stock standard solutions of each compound were prepared byissolving 10.0 mg in methanol to obtain a final volume of 10 ml.hese solutions were stored in stained glass-stopper bottles under

efrigeration at 4 ◦C. Standard working mixtures, at various concen-rations, were daily prepared by appropriate dilution of aliquotsf the stock solutions in methanol–water (25:75, v/v). A workingtandard mixture containing 100 �g ml−1 of each compound wasrepared in methanol for using it as spiking solution. This work-

2

pi

r. A 1209 (2008) 162–173 163

ng standard mixture was used to spike control tissue samples atoncentrations ranging from 10 �g kg−1 (if the sensitivity allows it;therwise, the analyte was spiked at its LOQ) to twice the maximumesidue limit (MRL) and to prepare matrix-matched standards atoncentrations ranging from 10 to 1000 �g kg−1, with the previousestriction, for calibration curves. Individual and mixed stock solu-ions were replaced once a month. Matrix-matched standards andortified samples (or incurred) were spiked with the same amountf I.S., on the day of the analysis at a concentration of 50 �g kg−1.

Formic acid was from Panreac (Barcelona, Spain) and methanolliquid chromatography grade) from Merck (Darmstadt, Germany).eionized ultra pure water (

164 V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173

Fig. 1. Chemical structures of the selected antimicrobial agents.

V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173 165

Table 1SRM conditions used in the LC–MS/MS analysis

Compound Time windowa (min) tR (min) CVb SRM 1c (m/z) SRM 2c (m/z) CEd Dwel time (s) Product ion ratio

Ronidazole 1: 2.4–6.0 4.88 20 201 → 140 201 → 109 20 0.2 0.08Sulfadiazine 5.03 20 251 → 108 251 → 156 20 0.2 0.87Lincomycin 2: 5.0–7.5 5.95 15 407 → 126 407 → 359 25 0.2 0.10Sulfathiazole 6.46 20 256 → 108 256 → 156 20 0.2 0.86Sulfapyridine 6.73 25 250 → 108 250 → 156 20 0.2 0.80Trimethoprim 3: 6.5–8.5 7.23 20 291 → 123 291 → 230 25 0.2 0.40Pipemidic acid 7.30 45 304 → 189 304 → 217 25 0.2 0.37Marbofloxacin 7.63 35 363 → 72 363 → 320 15 0.2 0.68Ofloxacin 4: 7.5–9.5 8.35 40 362 → 261 362 → 318 25 0.8 0.39Norfloxacin 8.74 35 320 → 205 320 → 233 40 0.8 0.62Sulfadimidine 8.79 30 279 → 124 279 → 156 20 0.8 0.81Ciprofloxacin 8.98 45 332 → 245 332 → 288 35 0.8 0.20Tetracycline 5: 7.8–8.5 8.54 30 445 → 410 445 → 427 15 0.8 0.60Cefalexine 8.79 25 380 → 106 380 → 198 20 0.8 0.65Dapsone (I.S.) 6: 8.0–10.0 8.95 20 249 → 107 249 → 153 20 0.8 0.02Oxytetracycline 7: 8.0–10.0 8.98 30 461 → 426 461 → 444 15 0.8 0.61Sulfamethoxypyridazine 9.45 35 281 → 108 281 → 156 20 0.2 0.80Enrofloxacin 8: 8.5–10.5 9.19 45 360 → 245 360 → 316 25 0.2 0.50Danofloxacin 9.32 50 358 → 340 358 → 283 30 0.2 0.52Chlortetracycline 9: 10.0–11.5 11.05 30 479 → 444 479 → 462 20 0.2 0.04Sulfisoxazole 10: 10.0–12.0 11.07 25 268 → 108 268 → 156 20 0.2 0.95Sulfabenzamide 11: 10.5–12.5 11.13 25 277 → 108 277 → 156 20 0.2 0.88Tilmicosin 12: 11.0–14.0 12.53 15 870 → 174 870 → 696 45 0.2 0.29Sulfadimethoxine 13: 12.0–14.0 13.08 35 311 → 108 311 → 156 25 0.2 0.48Sulfaquinoxaline 14: 12.5–14.0 13.49 30 301 → 108 301 → 156 20 0.2 0.73Erythromycin 15: 14.0–16.0 14.95 15 734 → 158 734 → 576 25 0.2 0.74Tylosin 16: 14.0–16.0 15.11 20 917 → 174 917 → 772 35 0.2 0.10Flumequine 17: 14.5–17.0 15.98 25 262 → 202 262 → 244 35 0.2 0.20Josamycin 18: 15.0–18.0 16.23 30 829 → 109 829 → 174 35 0.2 0.75Cloxacilline 19: 17.0–20.0 17.56 20 468 → 160 468 → 178 25 0.2 0.62Dicloxacilline 18.95 25 502 → 160 502 → 212 25 0.2 0.70Sulfasalazine 20: 18.0–22.0 19.29 30 399 → 119 399 → 156 35 0.2 0.39

a Each time window correspond to a different total ion chromatogram (TIC).b CV, cone voltage (V).

firmat

twi

(nwpAsp((cc

tsfiotvao3

pf

2

lrdsptfa

ba2i2t

cpdtt

c The first SRM transition was used for quantification and the second one for cond CE, collision energy (eV).e [M+H+CH3OH]+ is detected.

rometer (model Micromass Quattro Micro API, Waters), equippedith a dual Z-spray electrospray/atmospheric pressure chemical

onization source (Waters, Milford, MA, USA).Separation was achieved on an XTerra MS C18 LC column

100 mm × 2.1 mm I.D., 3.5 �m) from Waters. Mobile phase compo-ent A was methanol with 10 mM formic acid, while component Bas water with 10 mM formic acid too. At 0.2 ml min−1, the mobilehase gradient profile was as follows (t in min): t 0, A = 10%; t 20,= 90%; t 21, A = 10%, kept for 10 min. The injection volume was

et to 20 �l. MS/MS detection in the Quattro Micro-LC system waserformed with the ESI source operating in the positive ionizationPI) mode, under the time-scheduled selected reaction monitoringSRM) conditions shown in Table 1. Two SRMs were monitored perompound: the first one and most abundant was used for quantifi-ation, and the second one for confirmation.

The selection of the ionization mode and the optimization ofhe various parameters influencing the MS signal, including thepecific cone and collision energies for each analyte, were per-ormed by on-column injection (20 �l) of standard solutions of thendividual compounds and mixtures of all of them. The resultingptimized values were as follows: capillary voltage 3.2 kV; sourceemperature, 125 ◦C; desolvation temperature, 400 ◦C; extractoroltage 3 V and radio frequency (RF) lens 0.4 V. Nitrogen was useds both the nebulizing and the desolvation gas at 630 l h−1. For

peration in the MS/MS mode, argon was used as collision gas at.2 × 10−3 mbar.

Instrument control and data acquisition and evaluation wereerformed with the MassLynx 4.0 software package purchasedrom Micromass.

c

twF

ion.

.4. Validation procedure

An analyte was considered as positively identified when the fol-owing criteria were met: (i) the ratio of the relative (to the I.S.)etention time of the analyte to that of the same analyte in stan-ard solution was within ±2.5% tolerance; (ii) the presence of aignal at each of the two SRMs for the analyte (the use of two SRMser compound counts for four identification points, as defined byhe EU Commission Decision 2002/657/EC [9] and it is a warrantyor the selectivity of the method); (iii) the peak area ratio of SRM2gainst SRM1 was within the tolerance fixed by the EU criteria [9].

Calibration curves, CC�, CC�, recoveries, precision (repeata-ility and within-laboratory reproducibility), selectivity, stabilitynd ruggedness were assessed according to the EU regulation002/657/EC [9]. In addition to the quality parameters included

n that Decision, LODs and LOQs were determined. Decision002/675/EC is only a guideline and several authors have suggestedhe need of a more rigorous validation procedure [10,38].

The linearity of the analytical methods was proved building thealibration curves for each compound using extracts of blank sam-les (n = 7) spiked from 10 �g kg−1 (or from the LOQ if the sensitivityoes not allows it) to 1000 �g kg−1. Each level was prepared inriplicate. The linear regression analysis was carried out by plot-ing the peak area ratio of the analyte and I.S. versus the analyte

oncentration.

The LOD was calculated at a signal-to-noise (S/N) ratio of 3, whilehe LOQ value was calculated by using a S/N of 10. LODs and LOQsere obtained by the transition with highest S/N ratio in SRM mode.

or the LOQ, the confirmatory transition should be, at least, visible

1 atog

ibMtdla

tFwtt1bdsts

aalbwccbid

3

3

wutascctp(r�ot

eecc((mqfattfe

ep

cdoucc[sttt

3

eptrI

trnaro9at

aiaceate

hNfitttmr

tsce

i2

66 V. Carretero et al. / J. Chrom

n the chromatogram. CC� values were determined by analysing 20lank samples fortified with antimicrobials at their correspondingRLs. CC� was calculated as the decision limit CC� plus 1.64 times

he corresponding standard deviation (ˇ = 5%), supposing that stan-ard deviation at the CC� level equals standard deviation at the MRL

evel. For those compounds, without an established MRLs, the CC�nd CC� were calculated at the LOQ.

The recovery experiments were carried out at three concentra-ion levels (LOQ, MRLs and 2 MRL) in independent quintuplicates.or recovery studies, 1 g portion of chopped pig or cattle meatas accurately weighed in a porcelain capsule and spiked with

he antimicrobial standard solution, taken care to uniformly spreadhem on the sample. The spiked sample was left for, at least,0 min at room temperature to ensure the appropriate distri-ution in the matrix. This time span was established checkingifferent intervals (from analyzing the sample immediately afterpiking to doing it after 30 min) as the minimum time requiredo reach the appropriate distribution of the analytes in theamples.

Precision (within- and between-day) was calculated from thenalysis of 16 blank bovine or porcine samples fortified with allnalytes at each of the three specified fortification levels. Within-aboratory precision was obtained by following the same protocolut performing the analyzes in three different days. The specificityas assessed by analysing blank tissue samples. The stability was

hecked preserving standard and spiked samples under differentonditions. The ruggedness of the method was mainly establishedy the previous optimization procedure, which gave an idea on thenfluence of different factors, and also by the analysis of meat fromifferent animal species.

. Results and discussion

.1. LC–MS/MS optimization

Optimization of the various MS/MS experimental conditionsas performed by injection of standard solutions of the individ-al compounds and of mixtures of all of them. Identification ofhe parent ion and selection of the optimum cone voltage for eachnalyte were performed in the full scan mode by recording masspectra from m/z 50 to 1000 in the PI mode, at different values ofone voltage (from 15 to 45 V at intervals of 5 V). At the selectedone voltages (see Table 1), compounds undergo soft fragmen-ation showing one predominant ion, which corresponds to therotonated molecule [M+H]+, except for the group of �-lactamscloxacillin, cefalexin and dicloxacillin) whose precursor ion cor-esponds to [M+H+MeOH]+. The formation of this precursor ion for-lactams has been previously reported [17] and it can probablyriginated by either, methanolysis in the methanol solution or tohe formation of the methanol adduct.

The MS/MS conditions were adjusted under various collisionnergies, from which the two more intense product ions forach compound were selected in the final method. This studyovers seven classes of antimicrobials—3 �-lactams (cloxacillin,efalexin, dicloxacillin), 1 lincosamide (lincomycin), 4 macrolideserithromycin, josamycin, tilmicosin, tylosin), 1 nitroimidazoleronidazol), 8 quinolones (enrofloxacin, norfloxacin, ofloxacin,

arbofloxacin, ciprofloxacin, danofloxacin, pipemidic acid, flume-uine), 10 sulfonamides (sulfabenzamide, sulfadimethoxine, sul-amethoxypyridazine, sulfapyridine, sulfaquinoxaline, sulfadi-

zine, sulfisoxazole, sulfathiazole, sulfadimidine, sulfasalazine), 3etracyclines (tetracycline, chlortetracycline, oxytetracycline) andrimethoprim. The fragmentation of the precursor ion follows dif-erent patterns according to each class. Fragmentation pattern ofach class has already been compiled in several reviews [1,5] and

mpT

b

r. A 1209 (2008) 162–173

xperimental studies [13,28,34,35]. The structure of most of theroduct ions selected in this study is outlined in those references.

The three tetracyclines included in the study, tetracycline,hlortetracycline and oxytetracycline epimerized under acidic con-itions. This epimerization could leads to an earlier retention timef the epimer compared to the parent drug (about 1 min earlier,nder the chromatographic conditions described). However, in thisase, tetracyclines were not epimerized through the extraction pro-edure, probably, because the extraction conditions were neutral13,34,35]. Ronidazole is rapidly metabolized, by oxidation of theide-chain linked to C2, in products that have a similar toxic poten-ial as the parent compound. However, in this case, we only includedhe parent compound as a preliminary screening of the ability ofhe method to determine this class of substances.

.2. PLE optimization

In the optimization of the PLE procedure, all important param-ters affecting extraction efficiency were evaluated: temperature,ressure, treatment of sand, static time, cell size, number of extrac-ion cycles and flush volume. For this optimization, the calculatedecoveries were the absolute ones including that obtained for the.S.

Four different heating temperatures (50, 70, 90 and 100 ◦C) wereested to establish the extraction temperature able to give goodecovery of the analytes and the lowest amount of matrix compo-ents, because these last ones could contaminate the ion sourcend/or interfere with the rest of the analysis. Fig. 2a shows theecovery values for all the antimicrobials tested. The improvementf the recoveries was directly proportional to the temperature until0 ◦C, then it decreases. The best results were obtained between 70nd 90 ◦C, with recoveries ranging from 50 to 101%. Thus, an extrac-ion temperature of 70 ◦C was used for subsequent experiments.

Four different pressures, 1000, 1500, 2000 and 2500 psi, weressayed. Recoveries did not improve when values other than thenitial one (1500 psi) were used. This parameter is not considereds a critical one because the purpose of pressurizing the extractionell is to prevent the solvent boiling at the extraction temperature,nsuring that the solvent remains in contact with the sample. Prob-bly, pressure has little or no effect on the analyte recovery becausehe temperature (70 ◦C) is below the boiling point of water. Thextraction pressure was set at 1500 psi for further experiments.

Three different sand treatments (sand treated with EDTA, withydrazine or with both) were tested (Fig. 2b). The action of thea2EDTA is to deactivate metal impurities present on the sand sur-

ace by formation of stable complexes, and that of the hydrazines to displace the antimicrobial fraction, if any, reversibly boundo certain sites of the tissue components [31–34]. Treatment ofhe sand with 0.2 mol l−1 Na2EDTA provided improved recovery ofetracycline, quinolone, and �-lactams. On the contrary, the treat-

ent with hydrazine or with mixtures hydrazine–EDTA diminishedecoveries.

Six static times (2, 5, 10, 15 and 20 min) were checked, showinghat recoveries increased with the static time until 10 min. Longertatic times (data not shown) did not influence the extraction effi-iency. The extraction time was set to 10 min to assure a rapidxtraction as well as high and reproducible recoveries.

Four cell sizes (5, 11, 22 and 33 ml) were used. Recoveryncreased in parallel to the cell size until 22 ml (Fig. 2c). The cells of2 and 33 ml provided similar results. The cell size determines how

uch solvent is used, because the cell is flushed in an established

ercentage, larger cells are flushed with more volume of solvent.he cells of 22 ml were chosen.

The recoveries obtained are very similar from one to three cycles,eing a bit lower when three cycles were used. Possibly, the higher

V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173 167

peratu

tb

tpmfi6wasi

3

Ttocst

wDisatcefcsocqsTc

Fig. 2. PLE recoveries using different (a) tem

he number of cycles, the greater the degradation of the antimicro-ials. One cycle was selected as optimum.

The flush percentage refers to the amount of solvent flushedhrough the cell following the static heating step, expressed as aercentage of the cell volume. Increasing the flush volume allowsore solvent to pass through the sample but it also increases the

nal volume of the extract. For optimizing the solvent flush step, 30,0, 100 and 150% of the extraction cell, volume (22 ml) of solventas pumped into the cell. No difference in recovery was seen for

ny compound, so the percentage of flush was set at 60% for theubsequent extractions. The other amounts of solvent would havencreased the final volume of the extract, decreasing the sensitivity.

.3. Method validation

The specificity was assessed by analyzing blank tissue samples.he absence of background peaks with a S/N > 3, at the retention

imes of the target compounds, showed that the method is freef endogeneous interferences. Fig. 3 shows the extracted total ionhromatograms (TIC) resulting from the analysis of a pork tissueample spiked with the target analytes, at concentrations equalo half of their respective MRLs. Twenty separate data acquisition

rcfdn

res, (b) treatment of sand and (c) cells size.

indows were used, some of them including several compounds.uring the course of these experiments (∼3 months), slight shifts

n analyte elution time, which required the adjustment of an acqui-ition window, were observed. When handling complex matricesnd using the ESI source, partial suppression of analyte ioniza-ion in the electrosprayed solution by competitive ionization ofo-eluting matrix components is often observed [25,32,33]. Thextent of this effect is related to both concentrations and affinitiesor the proton (or cations) of the co-extracted and co-eluted matrixomponents. In this study, significant variations of the analyte ionignal intensities were observed if injected from any tissue extractr from a standard solution (data not shown). Signal suppressionan be as high as 40%. Analytes in bovine or pork tissue extract wereuantified by calibration curves obtained from extract spiked withtandard solutions. The calibration parameters are summarized inable 2. As can be noted, appropriate linearities were observed:orrelation coefficients were r = 0.999 for all the analytes, while

esiduals are below 20% in the low level and below 10% in highoncentration range. The slope of the equations was very similaror meat extracts of both animal species, indicating that this factoroes not influence the results, and only a few intercept values wereegative.

168 V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173

Fig. 3. SRM LC–MS/MS chromatogram resulting from the analysis of a bovine sample spiked with the 31 target analytes at 50 �g kg−1 and the I.S. at 50 �g kg−1.

V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173 169

Table 2Calibration parameters obtained from extract of non-incurred cattled and pig samples spiked from 10 �g kg−1 – excepting tetracycline from 30 �g kg−1, cloxacillin, cefalexinand pipemidic acid from 50 �g kg−1, tilmicosin, flumequine and ronidazole 25 �g kg−1 and enrofloxacin 15 �g kg−1 – to 1000 �g kg−1

Compound Cattle matrix Pig matrix

Slope Intercept r2 Residualsa Residualsb Slope Intercept r2 Residualsa Residualsb

Tetracycline 0.92 0.02 0.996 5.2 0.4–10 0.96 0.02 0.999 12.57 0.5–9.3Chlortetracycline 2.93 0.06 0.990 6.2 1.3–9.7 2.87 0.01 0.992 6.34 0.9–8.4Oxytetracycline 2.49 −0.18 0.999 8.0 0.6–8.9 2.57 0.01 0.994 7.90 3.0–6.9Josamycin 5.79 0.03 0.996 9.4 0.2–5.8 6.66 0.09 0.992 7.90 1.3–7.2Lincomycin 3.15 0.02 0.998 2.6 0.1–4.5 3.21 −0.18 0.993 9.08 0.3–4.2Cloxacillin 0.71 0.01 0.999 4.2 0.3–7.8 0.73 0.02 0.997 6.34 0.6–5.9Cefalexin 0.22 0.01 0.999 11.7 0.5–3.2 0.24 −0.35 0.997 7.32 2.1–6.9Dicloxacillin 2.36 0.04 0.997 10.3 0.2–3.3 2.46 0.00 0.998 6.54 1.3–7.2Erythromycin 4.29 −0.06 0.999 6.9 1.6–12.1 4.46 0.01 0.998 7.71 0.9–6.8Tylosin 0.95 0.01 0.997 2.1 1.8–9.1 0.99 0.03 0.996 10.10 1.2–7.9Tilmicosin 7.27 −0.23 0.997 12.1 3.1–12.0 7.12 0.05 0.998 8.10 1.3–8.0Enrofloxacin 0.74 0.00 0.992 15.9 2.7–10.8 0.72 0.10 0.993 5.76 0.6–12.1Norfloxacin 1.13 0.01 0.994 16.4 2.0–5.5 1.17 0.00 0.991 6.73 2.0–11.2Ofloxacin 1.85 0.03 0.996 2.6 1.2–6.9 2.13 0.01 0.998 9.08 3.4–10.1Marbofloxacin 1.46 0.04 0.995 7.8 0.8–7.4 1.66 0.08 0.997 5.56 2.8–8.9Ciprofloxacin 4.12 0.10 0.992 3.9 2.3–8.8 4.70 0.03 0.994 7.71 3.2–8.2Danofloxacin 0.59 0.00 0.997 8.7 0.3–3.4 0.67 0.32 0.995 6.32 3.4–7.4Pipemidic acid 0.20 0.01 0.999 11.9 1.8–7.9 0.22 0.04 0.998 7.71 2.8–10.3Flumequine 3.02 0.08 0.998 2.0 2.1–5.2 3.14 0.04 0.999 7.90 2.1–6.9Sulfabenzamide 1.35 0.03 0.994 2.3 0.8–4.3 1.40 0.04 0.994 9.49 2.8–5.1Sulfadimethoxine 2.76 0.31 0.998 9.9 0.3–2.4 3.07 0.01 0.994 5.76 2.4–8.3Sulfamethoxypyridazine 1.71 0.04 0.998 7.4 0.5–8.7 1.77 0.10 0.996 6.73 2.7–7.2Sulfapyridine 6.63 0.04 0.999 5.2 0.6–9.7 7.43 0.03 0.997 8.30 1.7–6.9Sulfaquinoxaline 3.66 0.04 0.999 2.3 0.3–6.4 4.13 0.02 0.997 10.49 1.4–7.2Sulfadiazine 1.21 0.01 0.998 6.8 0.4–5.0 1.38 0.01 0.993 5.56 2.4–6.8Sulfadimidine 3.80 0.10 0.996 1.3 2.4–6.2 3.96 0.10 0.995 10.54 0.9–10.3Sulfasalazine 2.69 0.03 0.999 4.6 1.6–8.4 2.80 0.20 0.998 7.12 2.5–9.2Sulfathiazole 1.63 0.02 0.998 3.4 2.7–10.2 1.70 0.02 0.997 6.73 2.6–7.9Sulfisoxazole 0.39 0.01 0.998 11.3 2.3–6.9 0.45 0.01 0.999 13.57 1.4–8.2Trimethoprim 1.80 0.10 0.999 2.6 1.6–8.3 2.23 0.01 0.995 6.34 0.2–7.3Ronidazole 0.13 0.00 0.992 8.4 0.9–9.2 0.16 0.09 0.993 7.90 2.0–9.2

a Deviation of the measurement from its value predicted by the regression line at the lower concentration level.b Deviation of the measurement from its value predicted by the regression line at the remaining level.

Table 3LODs, LOQs, MRLs, CC�s and CC�s in bovine muscle of the analytes studied

Analyte LOD (�g kg−1) LOQ (�g kg−1) MRLs (�g kg−1) CC� (�g kg−1) CC� (�g kg−1)

Tetracycline 10 30 100 106 108Chlortetracycline 3 10 100 107 115Oxytetracycline 3 10 100 101 107Josamycin 3 10 200 203 209Lincomycin 3 10 100 104 117Cloxacillin 15 50 300 308 327Cefalexin 15 50 200 197 202Dicloxacillin 3 5 300 303 309Erythromycin 3 5 200 208 216Tylosin 3 5 100 101 107Tilmicosin 5 25 50 52 58Enrofloxacin 3 10 100a 101 103Norfloxacin 5 15 – 17 19Ofloxacin 3 10 – 12 14Marbofloxacin 3 10 150 156 162Ciprofloxacin 3 10 100a 103 108Danofloxacin 3 10 200 206 213Pipemidic acid 15 50 – 53 59Flumequine 5 25 200 206 214Sulfabenzamide 3 10 100 101 105Sulfadimethoxine 3 10 100 102 106Sulfamethoxypyridazine 3 10 100 103 109Sulfapyridine 3 10 100 102 108Sulfaquinoxaline 3 10 100 105 113Sulfadiazine 3 10 100 101 105Sulfadimidin 3 10 100 103 108Sulfasalazine 3 10 100 101 106Sulfathiazole 3 10 100 106 112Sulfisoxazole 3 10 100 109 115Trimethoprim 3 10 50 53 60Ronidazole 5 25 – 29 32

a MRL of 100 �g kg−1 established by the European Union for the sum of enrofloxacin and its metabolite ciprofloxacin.

170 V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173

Table 4Accuracy and precision for the analytes at three spiking levels (n = 5)

Analyte (%) Recovery (X ± RSD, %) Repeatabilitya (%) Reproducibilityb

LOQ (�g kg−1) MRL (�g kg−1) 2 MRLs (�g kg−1)

Tetracycline 70 ± 14 73 ± 12 78 ± 7 6.1 9.2Chlortetracycline 86 ± 12 86 ± 11 87 ± 9 7.1 10.0Oxytetracycline 71 ± 11 74 ± 10 79 ± 8 7.3 10.0Josamycin 79 ± 13 82 ± 11 85 ± 6 7.1 10.7Lincomycin 86 ± 15 88 ± 12 92 ± 10 9.9 11.3Cloxacillin 77 ± 17 84 ± 13 85 ± 5 9.2 11.8Cefalexin 74 ± 16 77 ± 11 82 ± 7 9.4 10.4Dicloxacillin 88 ± 15 91 ± 11 98 ± 8 8.2 13.0Erythromycin 85 ± 12 86 ± 10 87 ± 9 7.4 9.5Tylosin 89 ± 14 91 ± 12 94 ± 9 8.0 10.2Tilmicosin 79 ± 13 82 ± 10 85 ± 8 7.6 11.5Enrofloxacin 86 ± 10 88 ± 9 92 ± 8 7.4 9.2Norfloxacin 79 ± 12 89 ± 10 89 ± 7 7.8 9.6Ofloxacin 93 ± 11 92 ± 8 96 ± 6 7.8 9.3Marbofloxacin 77 ± 15 77 ± 13 79 ± 10 9.3 11.1Ciprofloxacin 85 ± 12 86 ± 11 88 ± 10 9.0 10.9Danofloxacin 71 ± 11 81 ± 8 81 ± 5 9.2 11.2Pipemidic acid 75 ± 13 79 ± 9 84 ± 5 8.8 10.3Flumequine 94 ± 12 93 ± 10 92 ± 7 8.6 10.3Sulfabenzamide 87 ± 16 88 ± 13 94 ± 10 8.2 10.0Sulfadimethoxine 86 ± 15 89 ± 12 92 ± 9 8.3 10.7Sulfamethoxypyridazine 81 ± 17 83 ± 12 91 ± 7 8.5 10.5Sulfapyridine 87 ± 14 90 ± 9 97 ± 5 9.1 10.2Sulfaquinoxaline 85 ± 18 89 ± 12 95 ± 6 9.3 11.5Sulfadiazine 98 ± 12 96 ± 10 95 ± 7 6.9 9.3Sulfadimidin 85 ± 11 87 ± 10 91 ± 8 7.2 10.4Sulfasalazine 89 ± 10 89 ± 9 91 ± 8 7.4 10.7Sulfathiazole 84 ± 13 85 ± 11 87 ± 10 8.6 11.9Sulfisoxazole 89 ± 15 90 ± 12 92 ± 9 9.4 12.1Trimethoprim 72 ± 12 75 ± 11 85 ± 10 9.8 12.7R

tebat5tcsvapa

osSpws

s1ap2mbmc

Dtbw

eapcica

3

(cts2imii

onidazole 77 ± 18 79 ± 13a Repeatability obtained at the LOQ level, n = 16.b Reproducibility obtained at the LOQ level, n = 16 in three different days.

Table 3 shows the LODs, LOQs, the MRLs, CC� and CC�. Data rela-ive to bovine and porcine samples did not differ considerably fromach other, only data resulting from the analysis of antimicrobials inovine tissue are shown. The LODs and LOQs were estimated fromSRM mass chromatograms resulting from the analysis of one of

he two matrices (bovine muscle) spiked with the analytes at the0 �g kg−1 level. Once estimated, LODs were checked by injectinghree different extracts of samples spiked at the estimated con-entration and LOQs by calculating recovery and precision. Thetatistical risk of making a wrong decision is expressed by the CC�alues with an error of 5% (probability of false non-compliance ≤5%)nd the CC� values with an error ˇ = 5% (probability of falsely com-liant samples ≤5%). These parameters were established for eachntimicrobial at its MRL when possible.

Precision (repeatability and within-laboratory reproducibility)f the method was determined using, independently, spiked blankamples at three different levels (16 spiked samples for each level).amples were analyzed on three different days, using daily pre-ared standard solutions. Recoveries varied between 75 and 99%ith RSDs not higher than 18%. The results, summarized in Table 4

how the good accuracy of the method.Stability of standards and extracts was tested with an aqueous

tandard of 400 �g kg−1, and extracts of samples spiked at 25 and00 �g kg−1. After preparation, they were stored at 4 and at −20 ◦Cnd measured immediately, 24 and 48 h later. The results obtainedoint out that both, the standard and the extracts, are stable at least

days in the aforementioned storage conditions. Ruggedness of theethod was clearly ascertained during the optimization procedure,

y establishing the consequences of the deliberate introduction ofinor reasonable variations and by the similar results obtained

hecking different animal species.

slqc

98 ± 8 9.3 12.3

The method is an adequate screening one that is, according to theecision 2002/657/EC [9], devised for a high throughput and used

o sift large number of samples for potential non-compliant resultsecause the presence and/or absence of the studied compoundsere accomplished in

V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173 171

Fig. 4. SRM LC–MS(MS) chromatogram obtained from two of the samples analyzed (A) bovine meat sample No. 45 and (B) porcine sample No. 37. Peak identification andconcentration of antimicrobials as in Table 5.

172 V. Carretero et al. / J. Chromatogr. A 1209 (2008) 162–173

Table 5Confirmatory and quantitative LC–MS/MS analysis of incurred bovine and porcine muscle tissue

Sample Analyte detected Incurred samples Matrix-matched standard Level (�g kg−1)

Product ions Ion ratio tR (min) Ion ratio tR (min)

Bov1 Enrofloxacin 316/245 0.496 ± 0.010 9.19 0.508 ± 0.007 9.21 59 ± 7Ciprofloxacin 245/288 0.200 ± 0.002 8.98 0.202 ± 0.003 8.99 24 ± 3

Bov19 Lincomycin 359/126 0.111 ± 0.004 5.93 0.108 ± 0.003 5.96 63 ± 8Bov 22 Enrofloxacin 316/245 0.496 ± 0.010 9.18 0.496 ± 0.010 9.19 39 ± 4

Ciprofloxacin 245/288 0.200 ± 0.002 8.97 0.200 ± 0.002 8.98 46 ± 6Bov 25 Sulfamethoxypyridazine 156/108 0.819 ± 0.004 9.50 0.822 ± 0.001 9.46 87 ± 7Bov 32 Chlortetracycline 462/444 0.049 ± 0.001 11.02 0.051 ± 0.004 11.05 43 ± 9Bov 45 Enrofloxacin 316/245 0.516 ± 0.002 9.21 0.502 ± 0.004 9.17 25 ± 3

Ciprofloxacin 245/288 0.199 ± 0.003 9.00 0.200 ± 0.001 8.96 10 ± 2Bov 51 Tilmicosin 696/174 0.332 ± 0.004 12.50 0.317 ± 0.006 12.54 41 ± 4Bov 56 Chlortetracycline 462/444 0.053 ± 0.005 10.99 0.052 ± 0.004 11.02 39 ± 7Bov 59 Sulfadiazine 156/108 0.878 ± 0.007 5.09 0.893 ± 0.005 5.05 83 ± 8Bov 68 Enrofloxacin 316/245 0.505 ± 0.005 9.19 0.509 ± 0.002 9.19 23 ± 4

Ciprofloxacin 245/288 0.203 ± 0.004 8.98 0.205 ± 0.003 8.98 37 ± 6Por 5 Oxytetracycline 444/426 0.600 ± 0.003 9.02 0.602 ± 0.003 9.02 54 ± 5Por 10 Oxytetracycline 444/426 0.612 ± 0.002 9.03 0.604 ± 0.006 9.01 42 ± 5Por 19 Cefalexin 198/106 0.687 ± 0.005 8.83 0.698 ± 0.004 8.85 112 ± 12Por 23 Enrofloxacin 316/245 0.507 ± 0.007 9.19 0.499 ± 0.007 9.19 53 ± 9

Ciprofloxacin 245/288 0.199 ± 0.003 8.98 0.200 ± 0.002 8.98 15 ± 3Por 24 Chlortetracycline 462/444 0.048 ± 0.002 11.03 0.059 ± 0.001 11.05 72 ± 6Por 37 Cloxacillin 178/160 0.640 ± 0.006 18.53 0.643 ± 0.002 18.50 176 ± 12Por 44 Enrofloxacin 316/245 0.510 ± 0.009 9.19 0.506 ± 0.008 9.19 26 ± 8

05P 01P 03P 05

((nriifd

iasl

4

cclhdCpttpMht

ccrttt

ttf

A

tg0tr

R

[[

[

Ciprofloxacin 245/288 0.209 ± 0.0or 49 Sulfamethoxypyridazine 156/108 0.806 ± 0.0or 62 Tylosin 772/174 0.101 ± 0.0or 70 Oxytetracycline 444/426 0.596 ± 0.0

sulfamethoxypiridazine and sulfadiazine), two tetracyclineschlortetracycline) and only one (1.35%) lyncomicine. However,one of them surpassed the MRLs established. The most commonesidues that appeared in the porcine samples were tetracyclinesn four samples (5.2%), quinolones in three (3.9%) and �-lactamicsn two (2.6%). Additionally, one macrolide (tylosin) and one sul-onamide (sulfamethoxypyridazine) were found, each one in aifferent sample (equivalent to 1.3%).

Fig. 4 shows a typical chromatogram of a bovine sample contain-ng quinolones (enrofloxacin and its metabolite ciprofloxacin), and

porcine one containing cloxacillin. Both chromatograms clearlyhow the unequivocal identification of the antimicrobial and theack of interfering peaks at the retention time of the other ones.

. Conclusions

A multi-residue procedure for the confirmation and quantifi-ation of 31 antimicrobial residues that belong to seven differenthemical classes: �-lactams, quinolones, sulfonamides, macrolides,incosamides, nitroimidazoles, tetracyclines and trimethoprimas been developed. This method involved PLE and LC–MS/MSetermination. The method validation in meat, according to theommission Decision No. 2002/657/CE, showed that it is sim-le, rapid, rugged, sensitive and specific, with recoveries higherhan 65%. Confirmation and quantitation were achieved at concen-rations below the tolerance levels. Additionally, LODs and LOQsointed out that residue concentration 100 times lower than theRLs can be determined. When tested on food samples intended for

uman consumption, the method was able to identify and quantifyhe antibiotics present in the samples.

The present study shows that despite the incidence of antimi-robial residues in meat, the contamination level could not be

onsidered as a serious public health problem according to EUegulations. To prevent exposure to antimicrobials, it is necessaryo reduce and control their use in farms and livestock installa-ions. Nevertheless, residue monitoring programs are appropriateo ensure minimal residue levels. These results clearly demonstrate

[

[[[[

8.98 0.207 ± 0.003 8.98 42 ± 99.47 0.804 ± 0.002 9.46 64 ± 5

15.13 0.106 ± 0.003 15.10 35 ± 29.01 0.602 ± 0.003 8.99 79 ± 4

hat LC–MS/MS is a useful tool to carry out confirmatory and quan-itative analysis of antimicrobials in meat in the food safety andood science fields.

cknowledgements

This work has been supported by the Conselleria de Sanitat ofhe Generalitat Valenciana through the project of research in pro-rammes of Health, Prevention and Prediction of Illness (Project No.17/2007) and by the Conselleria dı̌Empresa, Universitat i Ciència ofhe Generalitat Valenciana throught the R + D project for emergingesearch groups (GV/2007/264).

eferences

[1] A. Gentili, D. Perret, S. Marchese, Trac-Trends Anal. Chem. 24 (2005) 704.[2] C. Blasco, C.M. Torres, Y. Pico, Trac-Trends Anal. Chem. 26 (2007) 895.[3] A. Di Corcia, M. Nazzari, J. Chromatogr. A 974 (2002) 53.[4] V. Andreu, C. Blasco, Y. Picó, Trac-Trends Anal. Chem. 26 (2007) 534.[5] A.A.M. Stolker, U.A.T. Brinkman, J. Chromatogr. A 1067 (2005) 15.[6] FAO/WHO, Food Standards, Codex Alimentarius, Veterinary Drug Residues in

Food, Maximum residue limits, FAO, Rome, 2008, (available at http://www.codexalimentarius.net/mrls/vetdrugs/jsp/vetd q-e.jsp).

[7] Commission Regulation 508/1999/EC amending Annexes I to IV to Coun-cil Regulation (EEC) No 2377/90 laying down Community procedure for theestablishment of maximum residue limits of veterinary medicinal products infoodstuffs of animal origin, European Community, Brussels, 1999, p. 16 (avail-able at http://europa.eu.int).

[8] General Priciples for Evaluatin the Safety of Compounds in Food ProducingAnimals, US Food and Drugs Administration (FDA), Rockville, MD, 1986.

[9] Commission Decision 2002/657/EC implementing Council Directive 96/23/ECconcerning the performance of analytical methods and interpretation of results,European Union, Brussels, 2002, pp. 66–98 (available at http://europe.eu.int).

10] J.P. Antignac, B. Le Bizec, F. Monteau, F. Andre, Anal. Chim. Acta 483 (2003) 325.11] M. Clemente, M.P. Hermo, D. Barron, J. Barbosa, J. Chromatogr. A 1135 (2006)

170.12] S. Fraselle, V. Derop, J.M. Degroodt, J. Van Loco, Anal. Chim. Acta 586 (2007)

383.

13] M.P. Heymo, E. Nemutlu, S. Kir, D. Barron, J. Barbosa, Anal. Chim. Acta 613 (2008)

98.14] P. Mottier, I. Hure, E. Gremaud, P.A. Guy, J. Agric. Food Chem. 54 (2006) 2018.15] U. Santos, F. Ramos, Curr. Pharm. Anal. 2 (2006) 53.16] E. Verdon, P. Couedor, P. Sanders, Anal. Chim. Acta 586 (2007) 336.17] M.S. Díaz-Cruz, M.J. López de Alda, D. Barceló, J. Chromatogr. A 1130 (2006) 72.

http://www.codexalimentarius.net/mrls/vetdrugs/jsp/vetd_q-e.jsphttp://europa.eu.int/http://europe.eu.int/

atog

[[

[

[[

[

[

[

[

[

[[

[

[

[

[

[

[

[

V. Carretero et al. / J. Chrom

18] A. Juan-Garcia, G. Font, Y. Pico, Electrophoresis 28 (2007) 4180.19] R. Yamada, M. Kozono, T. Ohmori, F. Morimatsu, M. Kitayama, Biosci. Biotechnol.

Biochem. 70 (2006) 54.20] A.A.M. Stolker, P. Rutgers, E. Oosterink, J.J.P. Lasaroms, R.J.B. Peters, J.A. van Rhijn,

M.W.F. Nielen, Anal. Bioanal. Chem. 391 (2008) 2309.21] K. Granelli, C. Branzell, Anal. Chim. Acta 586 (2007) 289.22] T. Goto, Y. Ito, S. Yamada, H. Matsumoto, H. Oka, J. Chromatogr. A 1100 (2005)

193.23] S. Bogialli, R. Curini, A. Di Corcia, M. Nazzari, R. Samperi, Anal. Chem. 75 (2003)

1798.24] S. Bogialli, R. Curini, A. Di Corcia, M. Nazzari, M. Sergi, Rapid Commun. Mass

Spectrom. 17 (2003) 1146.25] S. Bogialli, R. Curini, A. Di Corcia, M. Nazzari, M.L. Polci, J. Agric. Food Chem. 51

(2003) 4225.26] S. Bogialli, V. Capitolino, R. Curini, A. Di Corcia, M. Nazzari, M. Sergi, J. Agric.

Food Chem. 52 (2004) 3286.27] A. Gentili, D. Perret, S. Marchese, M. Sergi, C. Olmi, R. Curini, J. Agric. Food Chem.

52 (2004) 4614.28] G. Font, A. Juan-Garcia, Y. Pico, J. Chromatogr. A 1159 (2007) 233.29] G. Berardi, S. Bogialli, R. Curini, A. Di Corcia, A. Lagana, J. Agric. Food Chem. 54

(2006) 4537.

[

[

r. A 1209 (2008) 162–173 173

30] S. Bogialli, R. Curini, A. Di Corcia, A. Lagana, M. Mele, M. Nazzari, J. Chromatogr.A 1067 (2005) 93.

31] S. Bogialli, R. Curini, A. Di Corcia, A. Lagana, G. Rizzuti, J. Agric. Food Chem. 54(2006) 1564.

32] S. Bogialli, C. Coradazzi, A. Di Corcia, A. Lagana, M. Sergi, J. AOAC Int. 90 (2007)864.

33] S. Bogialli, A. Di Corcia, A. Lagana, V. Mastrantoni, M. Sergi, Rapid Commun.Mass Spectrom. 21 (2007) 237.

34] S. Bogialli, G. D’Ascenzo, A. Di Corcia, G. Innocenti, A. Lagana, T. Pacchiarotta,Rapid Commun. Mass Spectrom. 21 (2007) 2833.

35] S. Bogialli, G. D’Ascenzo, A. Di Corcia, A. Lagana, S. Nicolardi, Food Chem. 108(2008) 354.

36] RASFF annual report 2007, Directorate General for Health & Consumer,European Commission, Brussels, 2008 (available at http://ec.europa.eu/food/food/rapidalert/index en.htm).

37] Council directive 96/23/CE on measures to monitor certain substances andresidues thereof in life animal and animal products and repealing the Direc-tives 85/358/EEC and 86/469/EEC and Decisions 89/187/EEC and 91/664/EEC,European Community, Brussels, 1996, p. 10 (available at http://europa.eu.int).

38] H.T. Ronning, K. Einarsen, T.N. Asp, J. Chromatogr. A 1118 (2006) 226.

http://ec.europa.eu/food/food/rapidalert/index_en.htmhttp://europa.eu.int/

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具

http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/cloud/http://www.xuebalib.com/http://www.xuebalib.com/vip.htmlhttp://www.xuebalib.com/db.phphttp://www.xuebalib.com/zixun/2014-08-15/44.htmlhttp://www.xuebalib.com/

Multi-class determination of antimicrobials in meat by pressurized liquid extraction and liquid chromatography-tandem mass spectrometryIntroductionExperimentalReagents and materialsSamples and sample extractionLC-MS/MS instrumentation and conditionsValidation procedure

Results and discussionLC-MS/MS optimizationPLE optimizationMethod validationApplication to samples from markets and slaughterhouses

ConclusionsAcknowledgementsReferences