J. Sep. Sci.

2013, 36, 1223–1230 1223

Marı́a Luz Gómez Pérez Research Article

Roberto Romero-González
José Luis Martı́nez Vidal
Antonia Garrido Frenich Analysis of veterinary drug residues in
Group “Analytical Chemistry of cheese by ultra-high-performance LC
Contaminants”, Department of
Chemistry and Physics, Research coupled to triple quadrupole MS/MS
Centre for Agricultural and Food
Biotechnology (BITAL), A simple, reliable, and fast multiresidue method has been developed for the determina-
University of Almerı́a, Almerı́a,
tion of 17 veterinary drugs belonging to several families (macrolides, sulfonamides, and
Spain
anthelmintics) in cheese at trace levels. Ultra-high-performance LC coupled to MS/MS has
been used for the analysis of these compounds in less than 9 min. Veterinary drug residues
Received October 30, 2012 have been extracted from cheese samples using a QuEChERS (quick, easy, cheap, effective,
Revised January 15, 2013
rugged, and safe)-based extraction procedure without applying any further clean-up step.
Accepted January 16, 2013
Matrix-matched calibration was used for quantification and recoveries were calculated at
three concentration levels (10, 50, and 100 ␮g/kg). The obtained values ranged from 70 to
110% for the selected compounds except for tylosin and josamycin at 100 ␮g/kg (111.7 and
112.7%, respectively). Intra- and interday precision were also evaluated and RSDs were lower
than 25% in all the cases. LOQs ranged from 0.3 ␮g/kg (for thiabendazole, oxfendazole,
mebendazole, josamycin, and fenbendazole) to 10.5 ␮g/kg (abamectin), whereas decision
limit and detection capability ranged from 2.3 (thiabendazole) to 11.3 (abamectin) and 4.2
(thiabendazole) to 14.3 ␮g/kg (abamectin), respectively. Finally, 13 samples were analyzed
and traces of thiabendazole were detected in two different cheeses.

Keywords: Cheese / MS/MS / QuEChERS / Ultra-high performance LC / Veterinary

drug residues
DOI 10.1002/jssc.201201003

1 Introduction ing these bacteria [8]. This provokes a delay in acidification,

deficiency in coagulation, soft texture, and bitter taste [5].
The large-scale animal production has favoured the use of vet- Furthermore, it must be highlighted that the pasteurization
erinary drugs (VDs) to improve the production and facilitate process does not completely eliminate VDs [9] and the intake
disease control [1]. However, VD residues or their metabo- of milk or dairy products (including cheese) containing these
lites can persist in animals and, therefore, come into the compounds, can cause adverse reactions already mentioned.
food chain. Thus, they have been detected in several types Several organizations such as European Union [10],
of animal tissues and animal products [2–4]. In this sense, US Food and Drug Administration (www.fda.gov/
the presence of VD residues in food can cause serious health medicaldevices/deviceregulationandguidance/databases/
problems as bacterial resistance and allergic reactions, as well ucm135680.htm) [11], and Codex Alimentarius (www.
as they can interfere during the manufacturing processes [5]. codexalimentarius.net/vetdrugs/data/index.html) [12] have
This last issue is important in the dairy sector, where the established maximum residue limits for VD in foodstuffs
presence of VDs is a serious problem because it affects the from animal origin, but there is a lack of legislation and
development of the manufacturing of cheeses and yogurts [6], information relating to the presence of these type of
causing economic losses [7]. Thus, during the elaboration of compounds in cheese, despite the important use of these
fermented dairy products like cheese or yogurt, lactic acid products in dairy cattle as well as the presence of VD residues
bacteria are used and they could be sensitive to a greater or in milk, which can be used for dairy products like cheese.
lesser extent to VDs present in milk, neutralizing or nullify- On the other hand, there is scarce literature related to the
determination of VD residues in cheese, using microbiolog-
ical [13] and chromatographic methodologies [14–17], apply-
Correspondence: Professor Antonia Garrido Frenich, Group “Ana-
lytical Chemistry of Contaminants”, Department of Chemistry and
ing fluorescence [14,17] or MS detection [15,16,18,19]. Micro-
Physics, Research Centre for Agricultural and Food Biotechnol- biological assays are easy to apply and they can provide suit-
ogy (BITAL), University of Almerı́a, Ctra. Sacramento s/n, Almerı́a, able results in a short time. However, this methodology does
Almerı́a 04120, Spain not allow the simultaneous determination of several classes
E-mail: agarrido@ual.es of VDs, and it only provides semiquantitive results. There-
Fax: +34 950015483
fore, chromatographic methods, mainly coupling LC-MS/
Abbreviations: Na2 EDTA, disodium ethylene diamine MS, have been applied in the last few years, considering that
tetraacetic acid; UHPLC, ultra-high-performance LC; VD, vet- it provides reliable determination of the target compounds at
erinary drug trace levels.


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
1224 M. L. G. Pérez et al. J. Sep. Sci. 2013, 36, 1223–1230

One of the main problems related to the determination of 2.2 Apparatus and software
these compounds in cheese is sample treatment, considering
the high content of protein and fat in this type of matrix, which Chromatographic analyses were performed in an ACQUITY
can often interfere in analytical procedures. Thus, current ex- UPLCTM system (Waters, Milford, MA, USA), using an
traction approaches are tedious and long, using solid-liquid Acquity UPLC BEH C18 column (100 × 2.1 mm), with
extraction with acetonitrile [17] and including clean-up steps 1.7-␮m particle size, from Waters. MS/MS detection was
based on SPE [14]. In the last few years, inexpensive and en- performed using an Acquity TQD tandem quadrupole mass
vironmentally friendly sample treatment procedures based spectrometer (Waters, Manchester, UK). The instrument
on matrix solid-phase dispersion, using hot water as extrac- was operated using an electrospray (ESI) source in positive
tant solvent have been proposed for the extraction of sulfon- mode. Data acquisition was performed using MassLynx 4.0
amides [15] or tetracyclines [16] from cheese. However, this software with QuanLynx software (Waters). Centrifugations
approach requires tedious extraction procedures, so in order were performed in a high-volume centrifuge from Centronic
to simplify sample treatment procedure, faster, and simpler (Barcelona, Spain). A Vortex mixer Heidolph (Schwabach,
procedures based on QuEChERS approach [20, 21] could be Germany), model Reax 2000 and an analytical AB204-S bal-
used, considering that it can be applied satisfactorily for the ance (Mettler Toledo, Greinfesee, Switzerland) were also
extraction of VDs from different matrices [22–24]. used. A Reax-2 rotary agitator from Heidolph was used for
Bearing in mind that current methods only determine sample extraction.
one compound [25] or a single class of VD residues in cheese,
such as sulfonamides [15, 19], tetracyclines [16, 18], or an-
thelmintics [17], the aim of this paper is the simultaneous 2.3 Ultra-high-performance LC–MS/MS analysis
determination of several classes of VDs in cheese (includ-
ing anthelmintics, sulfonamides, and macrolides), using a Chromatographic analyses were carried out using a gradient
simple and rapid extraction procedure based on QuEChERS elution with eluent A being methanol and eluent B consist-
approach combined with LC-MS/MS. ing on an aqueous solution of formic acid (0.01%, v/v). The
analysis started with 10% of eluent A, which was increased
linearly up to 100% in 5 min. This composition was held for
2 Materials and methods further 1.5 min before being returned to 10% of eluent A in
1 min, followed by a reequilibration time of 1 min, to give a to-
2.1 Reagents and chemicals tal run time of 8.5 min. The flow rate was set at 0.30 mL/min
and column temperature was maintained at 30⬚C. Aliquots of
Levamisole, sulfadimidine, sulfaquinoxaline, mebendazole, 5 ␮L of sample extract were injected into the chromatographic
fenbendazole, oxfendazole, thiabendazole, sulfachlorpyri- system.
dazine, sulfadimethoxine, emamectin, and abamectin were For MS/MS detection, the ionization source parameters
purchased from Riedel-de Haën (Seelze, Germany). Tylosin, in positive mode were: capillary voltage, 3.0 kV; extractor
tilmicosin, erythromycin, and josamycin were obtained from voltage, 4 V; source temperature, 120⬚C; desolvation temper-
Fluka (Steinheim, Germany). Albendazole and ivermectin ature, 350⬚C; cone gas flow, 80 L/h, and desolvation gas flow,
were purchased from Sigma-Aldrich (Madrid, Spain). 600 L/h (both gases were nitrogen). Collision-induced dissoci-
Stock standard solutions of individual compounds (with ation was performed using argon as collision gas at a pressure
concentrations between 200 and 300 mg/L) were prepared of 4 × 10−3 mbar in the collision cell. The optimum MS/MS
by dissolving the standard material in 100 mL of methanol conditions of the VDs were performed by column injection
(HPLC grade, Sigma). These solutions were kept at −20⬚C. of individual standards at 500 ␮g/L. Full-scan mass spectra
Then, a multicompound working solution (10 mg/L) was and product ion scan were acquired in order to obtain at least
prepared by combining suitable aliquots of each individual one precursor and two product ions for each compound for
standard stock solution and diluting them with appropriate both identification and quantification purposes, selecting the
amounts of methanol. This solution was kept at −20⬚C and most abundant product ion for quantification and the second
renewed every 2 weeks. one for confirmation.
Acetonitrile (HPLC-grade) was purchased from Sigma.
Acetic acid (purity >97%), formic acid (purity >98%), am-
monium formate, and magnesium sulfate anhydrous were 2.4 Sample preparation
obtained from Panreac (Barcelona, Spain). A 0.1 M Na2 EDTA
solution was prepared by dissolving disodium ethylene di- VDs were extracted from cheese samples using an extrac-
amine tetraacetic acid (Na2 EDTA) (Merck, Darmstadt, Ger- tion procedure based on QuEChERS methodology, previously
many) in water. Florisil silica material with 40-␮m particle used for the extraction of VD residues in milk [22]. Briefly,
size, 18% carbon load, and end capped and 12 mL SPE reser- the procedure was as follows: first, cheese was crushed and
voirs with two frits were provided by Varian (Harbor City, CA, homogenized, and then, 10 g of sample were weighed in a
USA). Ultrapure water was obtained from a Milli-Q Gradient polypropylene centrifuge tube (40 mL) and 10 mL of acetoni-
water system (Millipore, Bedford, MA, USA). trile containing 1% acetic acid were added. Then, 10 mL of


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
J. Sep. Sci. 2013, 36, 1223–1230 Liquid Chromatography 1225

Table 1. Retention time windows (RTWs) and MS/MS parameters of the selected veterinary drugs

COMPOUND RTW (min) Cone voltage (V) Quantification Confirmation

transition (m/z)a) transition (m/z)a)

Levamisole 2.45–2.55 36 205 > 123 (29) 205 > 117 (27)
Sulfachlorpyridazine 2.67–2.77 32 285 > 156 (15) 285 > 80 (50)
Sulfadimidine 2.90–2.99 35 279 > 124 (20) 279 > 92 (30)
Thiabendazole 3.09–3.13 30 202 > 175 (27) 202 > 131 (32)
Sulfadimethoxine 3.39–3.48 60 311 > 156 (20) 311 > 245 (18)
Sulfaquinoxaline 3.80–3.90 32 301 > 156 (18) 301 > 92 (30)
Tilmicosin 3.89–4.00 32 870 > 174 (40) 870 > 696 (40)
Oxfendazole 4.07–4.12 35 316 > 191 (22) 316 > 159 (35)
Tylosin 4.34–4.44 60 917 > 174 (40) 917 > 101 (45)
Erythromycin 4.45–4.54 32 717 > 158 (30) 717 > 116 (40)
Mebendazole 4.55–4.66 37 296 > 264 (25) 296 > 77 (46)
Josamycin 4.60–4.78 55 829 > 174 (35) 829 > 109 (35)
Albendazole 4.75–4.89 32 266 > 234 (20) 266 > 191 (35)
Fenbendazole 5.10–5.21 32 300 > 159 (35) 300 > 268 (20)
Emamectin 5.61–5.79 48 887 > 158 (39) 887 > 126 (39)
Abamectin 6.15–6.35 85 896 > 183 (55) 896 > 151 (60)
Ivermectin 6.27–6.50 75 898 > 754 (45) 898 > 183 (55)

a) Collision energies (eV) are given in brackets.

a solution of 0.1 M Na2 EDTA were added. The mixture was oxfendazole) have been included in this study, considering
stirred in a vortex for 1 min. After that, 4 g of magnesium that they may be found in cheese. The ESI probe in posi-
sulfate anhydrous and 1 g of sodium acetate were added and tive ion mode was selected as ionization technique due to its
the tube was stirred again for 1 min. Then, the mixture was sensitivity, ruggedness, easy handling and maintenance. In
centrifuged at 4500 × g for 5 min, and 2 mL of the supernatant all cases, [M + H]+ ions were found to be the most abun-
layer were taken and filtered through a Millex-GN nylon dant except for abamectin and ivermectin, which formed the
filter (0.20 ␮m, Millipore, Carrightwohill, Ireland). Finally, sodium adduct, [M + Na]+ (m/z 895.7 and 897.7, respectively).
1 mL of this filtered supernatant was diluted with 1 mL of Table 1 shows MS/MS transitions for quantification and con-
methanol/aqueous solution of formic acid 0.01% v/v (50:50, firmation, as well as cone voltages and collision energy values
v/v) and 5 ␮L were injected into the ultra-high-performance previously optimized for each of the selected compounds.
LC (UHPLC)-MS/MS system. Then, the extraction conditions were evaluated, using the
method developed previously for the extraction of VDs in
milk [22]. For that purpose, homogenized blank Edam cheese
2.5 Samples samples were spiked at 50 ␮g/kg with the target compounds
and three replicates were carried out with the proposed ex-
Thirteen samples of different types of cheese (Camem- traction method. In all the cases, matrix-matched calibration
bert, cream cheese, blue cheese, goat cheese, Brie, Emmen- using the specific experimental conditions for each exper-
taler, Mozzarella, Roquefort, cured lamb cheese, home cured iment was used to evaluate the recoveries of the selected
cheese, cheese with yogurt, integer cheese, and Edamer) were compounds.
purchased from local supermarkets in Almeria (Spain). All First, the amount of sample used during the extrac-
samples were analyzed following the procedure described tion process was evaluated, studying 3 and 10 g of sample.
above and those samples showing the absence of the target When 3 g of sample were used, 7 mL of water was added
compounds were used as blank samples in the preparation in order to favor the extraction of the compounds, consider-
of standards and recovery studies. ing the indications described previously for the application
of QuEChERS procedure when low water content samples
were analyzed [26]. Therefore, the ratio between amount of
3 Results and discussion sample and mL of sample extract is 0.3 and 1 g/mL, re-
spectively. The results are shown in Fig. 1A, and it can be
3.1 Optimization of the analytical method observed that lower recoveries were obtained when 3 g of
sample were used for all the compounds, except for sul-
The chromatographic method was based on a previously fachlorpyridazine, which shows similar results using both
reported UHPLC-MS/MS methodology [22]. New VDs amounts of sample. If 10 g of sample were used, recoveries
(sulfachlorpyridazine, thiabendazole, sulfadimethoxine, and ranged from 58 to 123%. Furthermore, it can be observed


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
1226 M. L. G. Pérez et al. J. Sep. Sci. 2013, 36, 1223–1230

Figure 1. Effect of (A) amount of sample,

(B) clean-up step with Florisil, and (C) ad-
dition of EDTA on the recovery of vet-
erinary drugs, when blank cheese sam-
ple was spiked at 50 ␮g/kg. Error bars in-
dicated the SD (n = 3). Veterinary drug
code: (1) levamisole; (2) sulfachlorpyri-
dazine; (3) sulfadimidine; (4) thiabendazole;
(5) sulfadimethoxine; (6) sulfaquinoxaline;
(7) tilmicosin; (8) oxfendazole; (9) tylosin;
(10) erythromycin; (11) mebendazole; (12)
josamycin; (13) albendazole; (14) fenbenda-
zole; (15) emamectin; (16) abamectin; and
(17) ivermectin.

that the variability of the obtained results is higher at lower affecting the recovery of the selected compounds. Therefore
amount of sample. This can be explained bearing in mind 10 g of sample were used for further experiments.
that although matrix effect could be more significant at Considering that QuEChERS usually includes a clean-
higher ratios (i.e. 1 g/mL), when lower ratios were used (i.e. up step, this has been evaluated using Florisil as adsorbent,
0.3 g/mL), the coextraction of interferences could be favored, bearing in mind the high fat content of cheese. For that,


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
J. Sep. Sci. 2013, 36, 1223–1230 Liquid Chromatography 1227

2 mL of the extract was passed through the Florisil cartridge, Table 2. Evaluation of matrix effect
comparing the obtained results with those obtained when
this clean-up step is avoided. Figure 1B shows the obtained Compound Slope (b) bmatrix /bsolvent
results and it can be noted that the addition of this clean-up
bsolvent bmatrix
step did not improve the recoveries of the VDs, except for
josamycin and abamectin. Therefore, a clean-up step was not
Levamisole 65.4 17.2 0.3
included in subsequent experiences. Sulfachlorpyridazine 47.5 36.9 0.8
Finally, the effect of the addition of EDTA was eval- Sulfadimidine 57.5 59.5 1.0
uated on the extraction of the target compounds, bearing Thiabendazole 170.6 76.9 0.5
in mind that it could improve the extraction of VDs that Sulfadimethoxine 120.3 201.9 1.7
could form chelate complexes with divalent cations and bind Sulfaquinoxaline 43.9 154.3 3.5
with proteins. For that purpose, 10 mL of Na2 EDTA solution Tilmicosin 3.6 1.9 0.5
(0.1 M) were added to the extraction solution and Fig. 1C Oxfendazole 153.4 213.3 1.4
shows the obtained results. It can be noted that if EDTA Tylosin 4.0 4.3 1.1
Erythromycin 95.8 48.2 0.5
was not added, the extraction of most of the selected com-
Mebendazole 814.2 776.8 0.9
pounds is dramatically reduced, and some of them, such as
Josamycin 19.2 9.5 0.5
thiabendazole, sulfadimethoxine, sulfaquinoxaline, oxfenda- Albendazole 454.6 743.6 1.6
zole, and tylosin, were not recovered from the sample. These Fenbendazole 1491.2 828.0 0.5
results are in accordance to previous studies in milk [22], al- Emamectin 1773.1 1103.2 0.6
though in other matrices, such as fish and meat, the results Abamectin 13.6 7.9 0.6
are different [27, 28], indicating that matrix could also affect Ivermectin 2.9 2.1 0.7
the extraction of VDs when EDTA was added to the extraction
solution. Therefore, EDTA was used for further experiments,
although more experiments will be carried out in order to
evaluate the proper role of EDTA in the extraction of VDs dazole, josamycin, fenbendazole, which ranged from 0.5 to
from complex matrices. 125 ␮g/kg. Three replicates for each calibration level were an-
alyzed. Peak area was selected as response and good linearity
within the tested interval was found with determination coef-
3.2 Validation of the proposed method ficients >0.99 in all the cases. Furthermore, the deviation of
the individual points from the calibration curve was always
Performance characteristics of the optimized method were es- <20%.
tablished by a validation procedure with spiked cheese sam- Trueness was studied through recovery studies, spik-
ples, studying matrix effect, linearity, trueness, intra- and ing blank samples at three fortification levels, 10, 50, and
interday precision, LOD and LOQ, decision limit (CC␣), de- 100 ␮g/kg. Recoveries ranged from 70 to 110% for all the
tection capability (CCß), and selectivity. Blank Edam cheese compounds assayed at the three levels evaluated (see Table
was used in all the experiments. 3) except for tylosin and josamycin at 100 ␮g/kg (111.8 and
When ESI is used as ionization technique, one of the 112.7%, respectively). These results indicate that good recov-
main problems is the decrease or increase of the signal due eries from cheese samples were obtained using the developed
to other components present in the matrix (matrix effect). To method.
evaluate the matrix effect, several concentrations (from 10 to The precision of the method was evaluated perform-
200 ␮g/kg) were analyzed in pure solvent and in blank cheese ing intra- and interday precision studies. Intraday preci-
samples. The slope ratios matrix/solvent for each compound sion was evaluated at the three concentration levels of the
were obtained, considering a tolerable signal suppression or recovery studies, performing five replicates for each level
enhancement effect if the slope ratio ranged from 0.8 to 1.2, (Table 3). It can be observed that repeatability, expressed as
whereas <0.8 or >1.2 implies a strong matrix effect. The ob- RSD was equal or <20%, except for several compounds as
tained results are shown in Table 2 and it can be observed sulfadimethoxine, tilmicosin, erythromycin, josamycin, and
that there is a strong matrix effect for most of the tested ivermectin at 10 ␮g/kg, and ivermectin at 50 ␮g/kg. Interday
compounds, except for sulfachlorpyridazine, sulfadimidine, precision was evaluated spiking one blank cheese sample at
tylosin, and mebendazole. For the other compounds, there 50 ␮g/kg, which was analyzed during five consecutive days.
is a decrease in the signal except for sulfadimethoxine, RSDs values were <20% for all the selected compounds ex-
sulfaquinoxaline, oxfendazole, and albendazole. Therefore cept for erythromycin (24.7%).
and to avoid matrix effect, matrix-matched calibration was LODs and LOQs were calculated analyzing blank sam-
used for a reliable quantification of VDs residues in ples spiked at 0.1, 0.5, 1, 2, 5, and 10 ␮g/kg, and they were
cheese. determined as the lowest concentration of the selected com-
Then, linearity was tested by spiking blank extract pounds that produce chromatographic peak at S/N of 3 and
cheese at five concentration levels ranged from 10 to 10, respectively (Table 4). In general good values were ob-
200 ␮g/kg, except for thiabendazole, oxfendazole, meben- tained and LODs ranged from 0.1 ␮g/kg (thiabendazole,


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
1228 M. L. G. Pérez et al. J. Sep. Sci. 2013, 36, 1223–1230

Table 3. Validation parameters of the developed method

Compound Recoverya) Interday precisionb)

10 ␮g/kg 50 ␮g/kg 100 ␮g/kg

Levamisole 108.4 (16.8) 76.8 (8.2) 99.2 (4.8) 8.5

Sulfachlorpyridazine 102.4 (16.2) 92.8 (10.6) 93.1 (5.5) 14.4
Sulfadimidine 106.3 (17.4) 90.3 (8.6) 86.4 (9.2) 13.5
Thiabendazole 107.3 (5.4) 76.7 (5.6) 79.3 (5.0) 14.1
Sulfadimethoxine 101.8 (20.9) 87.1 (10.7) 75.6 (5.3) 15.7
Sulfaquinoxaline 103.4 (7.7) 85.9 (7.4) 72.5 (3.0) 6.9
Tilmicosin 106.1 (20.1) 105.8 (13.6) 105.6 (9.8) 16.0
Oxfendazole 107.0 (7.3) 78.1 (8.7) 71.2 (5.3) 7.9
Tylosin 86.3 (17.2) 90.7 (6.9) 111.8 (7.1) 12.3
Erythromycin 75.3 (28.4) 76.8 (23.4) 100.2 (19.4) 24.7
Mebendazole 105.4 (6.5) 72.4 (7.0) 83.7 (4.8) 7.3
Josamycin 103.9 (21.4) 104.5 (14.2) 112.7 (11.6) 14.8
Albendazole 103.2 (6.6) 78.0 (6.3) 84.5 (4.5) 9.2
Fenbendazole 102.0 (8.0) 77.0 (3.6) 84.3 (4.3) 5.3
Emamectin 101.8 (10.4) 86.8 (4.6) 96.4 (4.3) 6.1
Abamectin 91.2 (18.7) 97.9 (9.6) 84.5 (11.3) 13.1
Ivermectin 101.8 (24.5) 109.8 (21.5) 86.7 (17.5) 23.0

a) Repeatability values, expressed as RSD, are given in brackets (n = 5).

b) RSD values obtained at 50 ␮g/kg. Samples were analyzed in five consecutive days.

Table 4. LODs, LOQs, decision limits (CC␣), and detection capa- abendazole) to 11.3 ␮g/kg (abamectin), whereas CCß ranged
bilities (CCß) of the selected compounds from 4.2 ␮g/kg (thiabendazole) to 14.3 ␮g/kg (abamectin).
Finally, selectivity was evaluated extracting and analyzing
LOD LOQ CC␣ CCß blank cheese samples. The absence of any signal at the same
(␮g/kg) (␮g/kg) (␮g/kg) (␮g/kg)
elution time as the target compounds indicates that there
were no matrix interferences that may give a positive signal.
Levamisole 0.7 2.4 5.6 9.8
Sulfachlorpyridazine 1.7 5.5 5.8 10.2
Identification of the target compounds was carried out by
Sulfadimidine 0.4 1.4 4.9 6.4 searching in the appropriate retention time windows, which
Thiabendazole 0.1 0.3 2.3 4.2 were obtained by mean retention time ± three SD of the
Sulfadimethoxine 0.5 1.7 3.7 8.7 retention time of blank samples spiked.
Sulfaquinoxaline 0.2 0.7 3.4 5.7
Tilmicosin 1.1 3.6 4.4 7.2
Oxfendazole 0.1 0.3 2.9 4.8
Tylosin 1.8 4.8 5.9 7.2 3.3 Sample analysis
Erythromycin 0.8 2.6 5.4 8.1
Mebendazole 0.1 0.3 3.0 5.4 The developed method was applied to different types of cheese
Josamycin 0.1 0.3 3.4 6.5
samples. Thirteen samples collected from several stores lo-
Albendazole 0.3 0.8 3.9 5.8
cated in Almeria were analyzed. For quantification purposes,
Fenbendazole 0.1 0.3 3.2 5.4
Emamectin 0.2 0.6 3.0 4.9 calibration curves prepared for each cheese evaluated in this
Abamectin 3.1 10.5 11.3 14.3 study were used. In order to ensure the reliability of the
Ivermectin 1.6 5.5 5.9 9.8 results, an internal quality control was used. This implies
the analysis of a reagent blank, obtained by performing the
whole procedure without sample, which removed false pos-
itives due to contamination of the reagents or instruments;
oxfendazole, mebendazole, josamycin, and fenbendazole) to a blank extract to eliminate the false positives caused by a
3.1 ␮g/kg (abamectin). On the other hand, LOQs ranged contamination in the extraction procedure or by the presence
from 0.3 ␮g/kg (thiabendazole, oxfendazole, mebendazole, of an interference; a matrix-matched calibration using blank
josamycin, and fenbendazole) and 10.5 ␮g/kg (abamectin). cheese, to check the sensitivity and linearity; and spiked blank
Furthermore, CC␣ and CCß were calculated based on a lin- sample at 50 ␮g/kg to evaluate recovery yield.
ear regression model according to the procedure described by Only thiabendazole was detected in two cheeses (blue
Verdon et al. [29]. The obtained results are shown in Table 4 cheese and goat cheese). Thus, thiabendazole was found in
and it can be observed that CC␣ ranged from 2.3 ␮g/kg (thi- blue cheese at 0.7 ␮g/kg, while goat cheese showed traces of


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
J. Sep. Sci. 2013, 36, 1223–1230 Liquid Chromatography 1229

acknowledges her grant (FPI) from the MINECO (AGL2010-

21370). R.R.G. is also grateful for personal funding through
the Ramón y Cajal Program (MINECO-European Social Fund,
ESF).

The authors have declared no conflict of interest.

5 References

[1] Stolker, A. A. M., Brinkman, U. A. T., J. Chromatogr. A

2005, 1067, 15–53.
[2] Kinsella, B., O’Mahony, J., Malone, E., Moloney, M.,
Cantwell, H., Furey, A., Danaher, M., J. Chromatogr. A
2009, 1216, 7977–8015.
[3] Bogialli, S., Di Corcia, A., Anal. Bioanal. Chem. 2009, 395,
947–966.
[4] Chafer-Pericas, C., Maquieira, A., Puchades, R., Trends
Anal. Chem. 2010, 29, 1038–1049.
[5] Berruga, M. I., Novés, B., Molina, M. P., Román,
M., Molina, A., Int. J. Dairy Technol. 2008, 61,
372–378.
[6] Bogialli, S., Di Corcia, A., Laganá, A., Mastrontani, V.,
Sergi, M., Rapid Commun. Mass Spectrom. 2007, 21,
Figure 2. UHPLC-MS/MS chromatograms of (A) a positive sample
237–246.
of thiabendazole at 0.7 ␮g/kg and (B) a blank sample spiked with [7] Font, H., Adrian, J., Galve, R., Estévez, M. C., Castellari,
thiabendazole at 5 ␮g/kg. M., Gratacós-Cubarsı́, M., Sánchez-Baeza, F., Marco, M.
P., J. Agric. Food Chem. 2008, 56, 736–743.
[8] Yamaki, M., Berruga, M. I., Althaus, R. L., Molina, M. P.,
the compound (<LOQ). Figure 2 shows the chromatogram Molina, A., Food Addit. Contam., Part A 2006, 23, 660–
of the positive sample of thiabendazole in blue cheese. 667.
[9] de Oliveira, R. C., Paschoal, J. A. R., Reyes, F. G. R., Food
Addit. Contam., Part B 2010, 3, 156–162
4 Concluding remarks [10] Commission Regulation (EU) No. 37/2010 of 22 Decem-
ber 2009, on pharmacologically active substances and
their classification regarding maximum residue limits in
A fast and simple method based on QuEChERS extraction
foodstuffs of animal origin, Off. J. Eur. Commun. 2010,
procedure and UHPLC-MS/MS was developed for the si- L15, 1–72.
multaneous determination of different classes of VDs (an-
[11] The Code of Federal Regulations, Title 21, Food and
thelmintics, sulfonamides, and macrolides) in cheese in less Drugs, PART 56, Revised April 2011, US FDA, Silver
than 9 min. The extraction procedure is based on a simple Spring, MD 2011.
extraction with acetonitrile, adding EDTA to facilitate the ex- [12] Codex Alimentarius, Veterinary Drug Residues in Food,
traction of the various classes of VDs. Furthermore, the use 34th Session of the Codex Alimentarius Commission,
of UHPLC coupled to MS/MS allows a rapid determination FAO, Rome 2011.
of the selected compounds. Good quantitative results for the [13] Usleber, E., Dade, M., Schneider, E., Dietrich, R., Bauer,
assayed compounds and good validation parameters in terms J., Märtlbauer, E., J. Agric. Food Chem. 2008, 56, 6857–
of linearity, trueness, precision, LOQ, CC␣, and CCß were 6862.
obtained. Despite strong matrix effect was observed, it was [14] Anastasio, A., Esposito, M., Amorena, M., Catellani, P.,
successfully compensated using matrix-matched calibration. Serpe, L., Cortesi, M. L., J. Agric. Food Chem. 2002, 50,
Considering the advantages of the proposed method, this 5241–5245.
could be applied for regular monitoring of these compounds [15] Berardi, G., Bogialli, S., Curini, R., Di Corcia, A., Laganá,
in cheese by routine laboratories, although further studies A., J. Agric. Food Chem. 2006, 54, 4537–4543.
should be carried out in order to include other classes of VD [16] Bogialli, S., Coradazzi, C., Di Corzia, A., Laganá, A.,
such as tetracyclines and quinolones. J. AOAC Int. 2007, 90, 864–871.
[17] Imperiale, F. A., Busetti, M. R., Suárez, V. H., Lanusse, C.
We gratefully acknowledge Spanish Ministry of Economy E., J. Agric. Food Chem. 2004, 52, 6205–621.
and Competitiveness-FEDER (MINECO-FEDER, AGL-2007- [18] Al-Mazeedi, H. M., Abbas, A. B., Alomirah, H. F.,
63569, and AGL2010-21370) for financial support. M.L.G.P. Al-Jouhar, W. Y., Al-Mufty, S. A., Ezzelregal, M. M.,


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com
1230 M. L. G. Pérez et al. J. Sep. Sci. 2013, 36, 1223–1230

Al-Owaish, R. A., Food Addit. Contam., Part A 2010, 27, Garcı́a-Campaña, A. M., J. Chromatogr. A 2011, 1218,
291–301. 4966–4971.
[19] Clark, S. B., Turnipseed, S. B., Madson, M. R., Hurlbut, [25] Imperiale, F., Ortiz, P., Cabrera, M., Farias, C., Sallovitz,
J. A., Kuck, L. R., Sofos, J. N., J. AOAC Int. 2005, 88, J. M., Iezzi, S., Pérez, J., Alvárez, L., Lanusse, C., Food
736–743. Addit. Contam., Part A 2011, 28, 438–445.
[20] Anastassiades, M., Lehotay, S. J., Stajnbaher, D., Schenk, [26] Mastovska, K., Dorweiller, K. J., Lehotay, S. J.,
F. J., J. AOAC Int. 2003, 86, 412–431. Wegscheid, J. S., Szpyka, K. A., J. Agric. Food Chem.
[21] Lehotay, S. J., Mastovska, K., Yun, S. J., J. AOAC Int. 2010, 58, 5959–5972.
2005, 88, 630–638. [27] Pereira-Lopes, R., Cazorla-Reyes, R., Romero-González,
[22] Aguilera-Luiz, M. M., Martı́nez-Vidal, J. L., Romero- R., Martı́nez-Vidal, J. L., Garrido-Frenich, A., J. Chro-
González, R., Garrido-Frenich, A., J. Chromatogr. A 2008, matogr. A 2012, 895–896, 39–47.
1205, 10–16. [28] Pereira-Lopes, R., Cazorla-Reyes, R., Romero-González,
[23] Kinsella, B., Whelan, M., Cantwell, H., McCormack, M., R., Garrido-Frenich, A., Martı́nez-Vidal, J. L., Talanta
Furey, A., Lehotay, S. J., Danaher, M., Talanta 2010, 83, 2012, 89, 201–208.
14–24. [29] Verdon, E., Hurtaud-Pessel, D., Sanders, P., Accredit.
[24] Lombargo-Agüi, M. Gámiz-Gracia, L., Cruces-Blanco, C., Qual. Assur. 2006, 11, 58–62.


C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com