Ultrasound-assisted dispersive extraction for the high pressure liquid chromatographic determination of tetracyclines residues in milk with diode array detection

Food Chemistry 150 (2014) 328–334

Contents lists available at ScienceDirect

Food Chemistry journal homepage: www.elsevier.com/locate/foodchem

Analytical Methods

Ultrasound-assisted dispersive extraction for the high pressure liquid chromatographic determination of tetracyclines residues in milk with diode array detection Eftichia Karageorgou, Marina Armeni, Ioulia Moschou, Victoria Samanidou ⇑ Laboratory of Analytical Chemistry, Department of Chemistry, Aristotle University of Thessaloniki, Thessaloniki 541 24, Greece

a r t i c l e

i n f o

Article history: Received 28 June 2013 Received in revised form 7 October 2013 Accepted 1 November 2013 Available online 13 November 2013 Keywords: Tetracyclines Milk Ultrasound Dispersion Chromatography

a b s t r a c t Ultrasound assisted matrix solid phase dispersive extraction was applied for the selective isolation and clean-up of tetracyclines (oxytetracycline, tetracycline, epi-chlorotetracycline, chlorotetracycline and doxycycline) from milk. Target analytes were determined by an accurate and sensitive chromatographic analytical method, which was validated to meet the European Legislation criteria. The separation was performed on a LiChroCART-LiChrospherÒ 100 RP-18 (5 lm, 250  4 mm) analytical column, operated at ambient temperature, followed by diode array detection. Validation included investigation of linearity, selectivity, stability, limits of detection and quantitation, decision limit, detection capability, trueness, precision and ruggedness according to the Youden’s approach. Limits of quantitation of examined tetracyclines were from 14.5 to 56.6 lg/kg significantly lower than respective Maximum Residue Limits, whereas recoveries ranged from 82.0% to 108%. The applicability of the method was evaluated using milk samples purchased from local market. Accuracy of the method was additionally proved by analysis of bovine milk certified reference material (BCRÒ-492). Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Antimicrobials of tetracyclines’s group (TCs) were first discovered in 1945. They are broad-spectrum antibiotic agents extensively used to control bacterial infections in humans and animals. They exhibit activity against infections caused by both Gram-positive and Gram-negative bacteria, chlamydia, mycoplasmas, rickettsiae and protozoan parasites. TCs are given to animals destined for human consumption to promote growth. However the presence of residues in milk or edible animal tissues may potentially cause allergic reactions or may lead to toxic and dangerous effects on human health (Samanidou, Nikolaidou, & Papadoyannis, 2007a). In order to protect human health, the European Union (EU) has enacted maximum residue limits (MRLs) for the presence of TCs in foodstuff of animal origin. The use of veterinary drugs in the EU is regulated by Commission Regulation (EU) No. 37/2010 (2009), which describes the procedure for the establishment of MRLs for veterinary medicinal products in food, whereas Decision 2002/ 657/EC (European Commission Decision 2002/657/EC, 2002), defines the performance criteria and the interpretation of results for analytical methods in the official control of residues in products of animal origin (Samanidou, Nikolaidou, & Papadoyannis, 2007b). ⇑ Corresponding author. Tel.: +30 2310997698; fax: +30 2310997719. E-mail address: [email protected] (V. Samanidou). 0308-8146/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.foodchem.2013.11.008

Analysis of tetracyclines in milk is performed mainly by HPLC with various detection techniques (Cinquina, Longo, Anastasi, Gianneti, & Cozzani, 2003; Frenich, Anguilera-Luiz, Vidal, & Romero-Gonzalez, 2010; Fritz & Zuo, 2007; Furusawa, 2003; Kaale, Chambuso, & Kitwala, 2008; Kishida, 2011; Robert et al., 2013; Samanidou et al., 2007a; Shariati, Yamini, & Esrafili, 2009; Tsai et al., 2010; Young & Tran, 2011; Zhao, Zhang, & Gan, 2004), although Flow Injection Analysis and immunochemical techniques have been also proposed (Gao, Zhao, Wang, & Wang, 2013; Rodriguez, Espinosa, Aguilar-Arteaga, Ibarra, & Miranda, 2010). Sample pre-treatment techniques used for the isolation of tetracyclines from milk include Solid Phase Extraction (SPE) (Fritz & Zuo, 2007; Furusawa, 2003) with prior deproteinization using TCA, EDTA, and EDTA-McIlvaine buffer solution (Cinquina et al., 2003; Samanidou et al., 2007b; Young & Tran, 2011), liquid–liquid extraction (Kaale et al., 2008; Zhao et al., 2004) and magnesium hydroxide co-precipitation method (Tsai et al., 2010) Recently, for the isolation of tetracyclines from milk a carrier mediated hollow fiber liquid phase microextraction was used (Shariati et al., 2009), and for the same purpose a magnetic solid phase extraction (MSPE) was performed, which involves the extraction and clean-up by silica based magnetic support dispersion on non-pretreated milk samples, followed by the magnetic isolation and desorption of the analytes by acidified methanol (Rodriguez et al., 2010). Also isolation of the target compounds

E. Karageorgou et al. / Food Chemistry 150 (2014) 328–334

was achieved by using centrifugal ultrafiltration device with Ultrafree-MC/PL [low binding regenerated cellulose membrane, nominal molecular weight limit (NMWL) = 5000, capacity <0.5 mL, without prior sample preparation (Kishida). These pre-treatment techniques demand either specific equipment not available in every analytical laboratory, or high consumption of solvents and long time of sample preparation. Nowadays, for the isolation of target compounds of complex matrices like milk, the trend is to use simple, fast, of low cost and almost solvent-free sample preparation techniques. Such a technique is Matrix Solid Phase Dispersion (MSPD). By a thorough search in the literature, MSPD was used only once for the isolation of tetracyclines residues from egg with no success according to the authors (Frenich et al., 2010). Recently matrix solid-phase dispersion extraction was applied for the determination of tetracycline residues in milk by capillary electrophoresis (Mu, Liu, Xu, Tian, & Luan, 2012). The uniqueness of MSPD as sample preparation technique is that it’s especially suitable for the extraction of solid, semi-solid and/or highly viscous food and biological matrices, consisted in obtaining isolation of target analytes by dispersing sample matrix onto a solid support. In this way any difficulty encountered by employing the classical SPE approach is eliminated. MSPD is a flexible technique regarding selectivity because various SPE sorbents can be used, allowing also combination of them. Sonication can assist the process of analytes extraction and sample clean-up as it provides an efficient contact between the solid and the extractant, which typically result in higher recovery rates of the target analytes (Karageorgou & Samanidou, 2010). The authors have previously introduced the combined use of ultrasound power in matrix solid phase dispersion (MSPD) with new sorbent materials like QuEChERS for the preparation of samples with satisfactory results for the multi-residue determination of other antimicrobials namely cephalosporins, penicillins, amphenicols and quinolones in milk (Karageorgou & Samanidou, 2010; Karageorgou, Myridakis, Stephanou, & Samanidou, 2013; Karageorgou & Samanidou, 2011; Karageorgou, Samanidou, & Papadoyannis, 2012). In this study ultrasound assisted matrix solid phase dispersion is proposed for the determination of five tetracyclines by HPLCDAD. Target analytes are oxytetracycline (OTC), tetracycline (TC), epi-chlorotetracycline (epi-CTC), chlorotetracycline (CTC) and doxycycline (DC). To the best of our knowledge, this is the first attempt that ultrasound assisted matrix solid phase dispersion is applied successfully on the isolation of tetracyclines residues from milk. Validation of the proposed method was performed according Commission Decision 2002/657/EC, determining linearity, selectivity, stability, decision limit, detection capability, trueness, precision and ruggedness according to the Youden’s test approach. Limits of detection and quantitation were also calculated although not required by Commission Decision 2002/657/EC for the integrity of the study. The bias of the method was further proved by the analysis of a certified reference material.

2. Materials and methods 2.1. Instrumentation A quaternary low pressure gradient HPLC–DAD system was used for the chromatographic analysis, purchased from Shimadzu (Kyoto, Japan). The solvent lines were mixed in an FCV-10ALVP mixer. An LC-10ADVP pump equipped with a Shimadzu SCL10ALVP System Controller, permitting fully automated operation, was used to deliver the mobile phase to the analytical column.

329

Sample injection was performed via a Rheodyne 7725i injection valve (Rheodyne, Cotati California, USA) equipped with a 20 lL loop. Detection was achieved by an SPD-M10AVP photodiode array detector, supplied with data acquisition software Lab Solutions-LC solutions by Shimadzu. Degassing of the mobile phase was achieved by helium sparging in the solvent reservoirs by a DGU10B degassing unit. A glass vacuum filtration apparatus obtained from Alltech Associates was employed for the filtration of ammonium acetate, using Whatman cellulose nitrate 0.2 lm membrane filters (Whatman Laboratory Division, Maidstone, England). A Glasscol small vortexer (Terre Haute, IN, USA) and an ultrasonic bath Transonic460/H (35 kHz, 170 W, Elma, Germany) were employed for the pretreatment of milk samples. All evaporations were performed with a ReactiVap 9-port evaporator model 18780 by Pierce (Rockford, IL, USA). Q-Max RR syringe filters (0.22 lm nylon membrane) used were purchased from Frisenette ApS (Knebel, Denmark). A LiChroCART-LiChrospherÒ 100 RP-18 (5 lm, 250  4 mm) analytical column by Merck (Darmstadt, Germany) was used for the chromatographic separation. Four SPE products were investigated towards their efficiency for the isolation of tetracyclines from milk: Nexus Abselut (60 mg/ 3 mL) by Agilent Technologies Inc., (Santa Clara, CA, USA), OasisHLB (200 mg/6 mL) by Waters (Milford, Massachusetts, USA), Merck-Lichrolut RP-18 (200 mg/3 mL) by Merck and Bond Elut Plexa (60 mg/3 mL) by Agilent Technologies Inc., 2 mL dispersive SPE tubes QuEChERS containing 150 mg magnesium sulfate, 50 mg PSA (primary, secondary amines) and 50 mg C18EC were purchased by Agilent Technologies Inc., and were used in MSPD mode. Certified reference material (BCRÒ-492) used, (consisted of 10 mL of a dry residue of lyophilised skimmed milk) was obtained from a bovine animal treated with oxytetracycline. This material has been certified by BCR (Community Bureau of Reference, EU, Geel, Belgium).

2.2. Chemicals and reagents Tetracycline (purity) >98% (TC), oxytetracycline hydrochloride >95% (OTC) and Chlorotetracycline hydrochloride >97% (CTC) were purchased from Fluka Chemie (Buchs, Belgium). Doxycycline hyclate >98% (DC) and caffeine (IS) were purchased from Sigma (Steinheim, Germany). epi-Chlorotetracycline hydrochloride >97% was purchased from Acros Organics (New Jersey, USA). HPLC grade methanol and acetonitrile were obtained from Fisher Scientific (Steinheim, UK). Oxalic acid was supplied by Merck. High purity water, obtained by a Milli-Q purification system (Millipore, Bedford, MA, USA), was used throughout study.

2.3. Chromatography Target analytes were separated by gradient elution using A: 0.01 M Oxalic acid and B: ACN. The initial volume ratio 85:15 (v/ v) changed to 70:30 (v/v) within 12 min and finally remained stable for 5 min. Flow rate was set at 1.2 mL/min providing an inlet pressure of 160–180 bar. Column performance was evaluated by calculation of number of theoretical plates N, tailing factor Tf, relative retention time RRT, retention factor k, resolution factor Rs, and the precision of the retention time and peak area. Peak identification was performed by spectral information provided by the diode array detector. Monitoring and quantitation was performed at 355 nm for tetracyclines and at 275 nm for the IS.

330

E. Karageorgou et al. / Food Chemistry 150 (2014) 328–334

2.4. Preparation of standards Stock standard solutions of each antibiotic were prepared in methanol (100 ng/lL) and found to be stable for four months, when stored at 4 °C in the dark. Working standards were prepared by further dilution in methanol every 5 days. All working standards contained caffeine as internal standard at a concentration of 1 ng/ lL. Aliquots of 20 lL were injected onto the column and quantitative analysis was based on peak area measurements as ratios versus peak area of internal standard. 2.5. Milk samples Milk samples were supplied from local market. Different varieties of milk were analysed: (a) skimmed milk 0% fat (fresh and high-temperature, short-time pasteurized milk), (b) semiskimmed milk 1.5% fat (fresh and high-temperature, short-time pasteurized milk), (c) full-fat milk 3.5% fat, (d) semi-skimmed milk evaporated dehydrated milk 4% fat and (e) full-fat evaporated dehydrated milk 7.5% fat. 2.6. Sample preparation The reversible adsorption capabilities of different commercially available sorbents for the isolation of antibiotics were initially evaluated under various protocols using aqueous standard solutions at 5 ng/lL concentration level. Therefore Bond Elut Plexa, Oasis HLB, Lichrolut RP-18 and Abselut Nexus preconditioned with 2 mL methanol and 2 mL water were evaluated under different elution systems, using methanol, acetone, acetonitrile, isopropanol, mixture of methanol–acetonitrile (1:1 v/v), mixture of 0.01 M oxalic acid–acetonitrile (1:1, v/v) and serial elution with 0.01 M oxalic acid followed by methanol or acetonitrile. No washing step was applied in order to minimize analytes loss. The sorbent that was found to provide higher recover rates was subsequently used for SPE optimization with standard solutions and then applied to milk in the mode dispersive extraction, which was enhanced ultrasonically. Samples of milk, which were analysed and found not to contain detectable residues of the analytes, were used as negative controls (blank samples) and were spiked at the MRL concentration level for tetracyclines. Various protocols were tested in order to achieve the best recoveries. The sorbent material was preconditioned by flushing 2 mL of methanol and 2 mL of water. Then it was emptied into a glass beaker, where 500 lg of spiked milk and 500 lL H2O were added. The mixture was blended with the dispersant material in a mortar with a pestle, while homogenisation was enhanced by sonication for 10 min in an ultrasound bath. Water in ultrasonic bath was periodically renewed to ensure that temperature was not significantly increased. Afterwards the sample was transferred into an empty cartridge reservoir and compressed. Quantitative transfer was assured from glass beaker and mortar with the minimum required volume of water. At this step the sorbent was dried by vacuum. Finally target analytes were eluted with 1 mL 0.01 M aqueous oxalic acid, followed by 1 mL acetonitrile. The samples were subsequently evaporated to dryness under a nitrogen stream in a water bath at 35 °C and the residues were dissolved in 200 lL of methanol. Aliquots of 20 lL were injected into the HPLC system.

between-day precision), decision limit (CCa), detection capability (CCb) and ruggedness. Limits of detection and quantitation were also calculated although not required by Commission Decision 2002/657/EC for the integrity of the assay. The linearity response of tetracyclines was initially studied using standard solutions, using working standards injected in triplicate, covering the entire working interval of 0.5–20 ng/lL (data not shown). In milk samples, linearity response was examined by fortifying matrix samples with a series of mixed standards of the examined compounds. Calibration curves were constructed by triplicate injections. Linear regression analysis was performed using the ratio of the standard area to internal standard area against the analytes concentration. The calculations for the limits of detection (LODs) were based on the standard deviation of y-intercepts of regression analysis (r) and the slope (S), using the following equation LOD = 3.3 r/S. Limits of quantitation (LOQs) were calculated by the equation LOQ = 10 r/S (Guidance for Industry Q2 (R1), 2005). Trueness was expressed in terms of recovery and bias as relative standard deviation using pooled negative milk samples fortified at three concentration levels (0.5  MRL, MRL and 1.5  MRL). Within-day precision was studied by five replicate measurements at these concentration levels, while between-day precision was conducted during routine operation of the system over the period of five consecutive days by triplicate analysis. Recovery was calculated as the percentage of the found mass of the analyte on the spiked sample toward the mass that was initially spiked. The decision limit, CCa, was calculated as the mean measured concentration at the MRL of each compound plus 1.64 times the SD of within-day precision at this concentration. The detection capability, CCb, was calculated as CCa plus 1.64 times the SD of within-day repeatability at CCa. Statistical analysis was performed at the 95% confidential level and the number of replicate analyses was 20. Selectivity was tested by analysing three blank milk samples of each category of milk mentioned in Section 2.2 in order to verify the absence of interfering peaks at the retention time of each antibiotic taken into account. Stability of target analytes in milk was examined by analyzing aliquots of spiked milk samples stored at 4 °C and 20 °C. Stability of extracted samples was also examined under the same conditions. The ruggedness of the method was assessed according to the Youden’s approach (Youden & Steiner, 1975). The basic idea is that several variations are introduced at once, instead of studying one alteration at a time. The experimental design described in Decision 2002/657/EC was applied and seven small but deliberate changes of the operating parameters (variables) were introduced. Their influence on the method results was assessed (European Commission Decision, 2002). The changes involved: sonication time, sorbent material, evaporation temperature, milk volume, concentration of wash solvent (oxalic acid), volume of elution solvent, as well as the nature of elution solvent. Negative pooled milk sample was spiked at 100 lg/kg and recovery of target analytes was estimated. Standard deviation of the differences Di (SDi): was calculated according to the equation:

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X 2 2 ðDi =7Þ

2.7. Method validation according to European Decision 2002/657/EC

SDi ¼

The developed method was validated according to the criteria set by European Commission Decision (2002). The following parameters were taken into account: selectivity, linearity response, trueness (bias and recovery), precision (repeatability and

When SDi is significantly larger than the standard deviation of the method carried out under intermediate precision conditions it is a foregone conclusion that all factors together have an effect on the result, even if every single factor does not show a significant

331

E. Karageorgou et al. / Food Chemistry 150 (2014) 328–334 Table 1 Ruggedness tested by applying seven small but deliberate changes in the operating parameters. Parameter

Units

Nominal levels

High level capital case values

Low level lower case values

A = sonication time B = MSPD sorbent material C = evaporation temperature D = milk amount E = oxalic acid concentration at wash step F = ACN elution volume G = Elution solvent Observed results

min

10 Plexa 25 0.5 0.01 1.0 ACN

10 Plexa 80 1.0 0.02 1.0 MeOH

8 OASIS 25 0.5 0.01 0.5 ACN

°C g m/L mL

influence and that the method is not sufficiently robust against the chosen modifications. The investigated factors and their levels of variation are reported in Table 1. 3. Results and discussion 3.1. Chromatography The gradient elution program yielded optimum separation of the five target antibiotics and the internal standard within 16 min. 3.2. Sample preparation As shown in Table 2, Bond Elut Plexa yielded the highest recovery rates among the tested sorbents and 0.01 M oxalic acid followed by 1 mL ACN proved to be the best eluent. Thus, MSPD optimization with spiked milk was performed with the chosen sorbent and eluent and good recoveries were achieved in the mode of ultrasound assisted dispersive extraction. Preconditioning of Plexa sorbent was found to improve its performance, whereas sonication provided an efficient contact between the solid and the extractant, thus resulting in higher recovery rates. Absolute recovery after ultrasound assisted MSPD was 110.7% for OTC, 58.7% for TC, 79.1% for epi-CTC, 79.6% for CTC and finally 79.3% for DC. The recovery of IS was 88% and was added to the milk sample together

Experiment number 1

2

3

4

5

6

7

8

A B C D E F G s

A B c D e f g t

A b C d E f g u

A b c d e F G v

a B C d e F g w

a B c d E f G x

a b C D e f G y

a b c D E F g z

with all others compounds. All examined analytes were well resolved from complex milk matrix. Typical chromatograms of blank and spiked milk sample after MSPD are illustrated in Fig. 1.

3.3. Method validation The developed method was validated according to the EU requirements for the analysis of veterinary drug residues in animal products and European Commission, Health and Consumer Protection Directorate—General, Guidelines for the implementation of decision 2002/657/EC, SANCO/2004/2726-rev, December 4, according to 2002/657/EC guidelines, using cow’s milk. Samples of milk, which were analysed and found not to contain detectable residues of the analytes, were used as negative controls (blank samples). The results of validation were calculated for individual substances. The results of figures of merit that were investigated are described in the following paragraphs.

3.3.1. Linearity and sensitivity Calibration curves constructed using fortified milk samples after MSPD, were obtained by least-squares linear regression analysis of the peak area ratio of analyte to the internal standard versus analyte injected amount. All calibration and sensitivity data are presented in Table 3. Regression equations revealed good

Table 2 Optimization of sample preparation. Sorbent

Eluent

SPE optimization with standard solutions Bond Elut Plexa 2 mL MeOH Oasis HLB Lichrolut RP-18 Abselut Nexus Bond Elut Plexa 2 mL acetone 2 mL ACN 2 mL isopropanol 2 mL MeOH-ACN (1:1 v/v) 2 mL 0.01 M oxalic acid -ACN (1:1, v/v) 1 mL 0.01 M oxalic acid followed by 1 mL MeOH 1 mL 0.01 M oxalic acid followed by 1 mL ACN Sorbent

Recoveries (%) OTC

TC

epi-CTC

CTC

DC

73.2 59.8 27.2 49.8 26.0 82.4 73.3 22.1 96.4 74.7 109

55.9 44.6 20.7 34.2 21.2 73.1 59.2 18.1 70.3 27.0 89.8

121.9 42.2 35.1 48.4 47.4 40.6 64.4 30.6 128 116 129

55.8 32.0 24.6 32.0 7.0 51.2 33.4 18.8 104 95.5 110

132 70.7 58.5 77.8 34.4 94.6 81.0 36.1 138 160 119

Optimized conditions

MSPD optimization with spiked milk Bond Elut Plexa Wash step(2 mL H2O) Without wash Combination with 125 mg of QuEChERS sorbent Elution with oxalic acid (pH = 4.8) Elution with 1 mL 0.4 M McIlvaine’s buffer solution followed by 1 mL 0.01 M oxalic acid Elution with 1 mL 0.01 M oxalic acid followed by 1 mL ACN (without sonication) Elution with 1 mL 0.01 M oxalic acid followed by 1 mL ACN (ultrasonically assisted)

Recoveries (%) 62.1 43.9 34.3 73.9 82.8 83.4 111

31.8 10.4 10.3 18.3 9.4 47.3 58.7

78.7 57.4 33.0 41.0 49.7 68.5 79.1

55.3 45.1 16.0 30.0 16.4 61.0 79.6

79.5 95.8 28.4 35.7 28.8 70.8 79.3

332

E. Karageorgou et al. / Food Chemistry 150 (2014) 328–334

Fig. 1. Chromatogram of blank milk sample after MSPD at (A1) 275 nm, and (A2) 355 nm. Chromatogram of fortified milk samples at 100 lg/kg at (B1) 275 nm, and (B2) 355 nm. Peaks: IS: 4.701 min, OTC: 7.300 min, TC: 9.063 min, epi-CTC: 11.570 min, CTC: 14.023 min, DC: 15.600 min.

Table 3 Linearity (triplicate injection), sensitivity and detectability data in milk matrix. Compound

Data analysis

r

LOD (lg/kg)

LOQ (lg/kg)

MRL (lg/kg)

CCa (lg/kg)

CCb (lg/kg)

OTC TC epi-CTC CTC DC

y = (0.00225397 ± 7.82136E05)x + (0.082173667 ± 0.008448039) y = (0.00042 ± 1.1E05)x + (0.018445 ± 0.001186) y = (0.000485 ± 6.49E06)x + (0.037236 ± 0.000701) y = (0.000836 ± 2.86E05)x + (0.013378 ± 0.003094) y = (0.000484 ± 2.53E05)x + (0.019476 ± 0.002738)

0.9994 0.9997 0.9999 0.9994 0.9986

12.4 9.3 4.8 12.2 18.7

37.5 28.2 14.5 36.9 56.6

100 100 100 100 100

113 99.1 104 99.1 101

115 102 109 118 105

x = lg/kg. y = peak area ratio.

333

E. Karageorgou et al. / Food Chemistry 150 (2014) 328–334 Table 4 Linearity (triplicate injection), sensitivity and detectability data in milk matrix. Compound

Data analysis

r

LOD (lg/kg)

LOQ (lg/kg)

MRL (lg/kg)

CCa (lg/kg)

CCb (lg/kg)

OTC TC epi-CTC CTC DC

y = (0.00225397 ± 7.82136E05)x + (0.082173667 ± 0.008448039) y = (0.00042 ± 1.1E05)x + (0.018445 ± 0.001186) y = (0.000485 ± 6.49E06)x + (0.037236 ± 0.000701) y = (0.000836 ± 2.86E05)x + (0.013378 ± 0.003094) y = (0.000484 ± 2.53E05)x + (0.019476 ± 0.002738)

0.9994 0.9997 0.9999 0.9994 0.9986

12.4 9.3 4.8 12.2 18.7

37.5 28.2 14.5 36.9 56.6

100 100 100 100 60

113 99.1 104 99.1 62

115 102 109 118 70.8

x = lg/kg. y = peak area ratio.

correlation coefficients ranging between 0.9986 and 0.9999 over the examined range. 3.3.2. Limit of quantitation, decision limit and detection capability All observed LOQ values were from 14.5 to 56.6 lg/kg, lower than respective MRL values. In compliance with the 2002/657/EC decision, the validation procedure includes the determination of CCa (limit of decision) and CCb (capability of detection). CCa values revealed after spiking 20 blank milk samples at MRL level and ranged from 62 to 113 lg/ kg, whereas CCb values by analyzing 20 blank spiked samples at corresponding CCa level for each analyte, and ranged from 70.8 to 118 lg/kg. Results are presented in Table 3 together with linearity and sensitivity data. 3.3.3. Selectivity The application of the whole procedure to fifteen blank milk samples in order to verify the method selectivity demonstrated that no interferences were detected. One major endogenous peak (approx. 5.254 min) and a number of minor importance ones (as for example a peak at 11.648 min, close to the epi-CTC peak) appear in the chromatogram of all milk samples, but as they elute at different retention times from the examined antibiotics and IS, they do not cause any interference. 3.3.4. Recovery and precision The precision of the method based on within-day repeatability was assessed by replicate (n = 5) measurements of three spiked milk samples at 50, 100, and 150 lg/kg. Relative recovery rates from the spiked samples were determined at three different concentrations by comparing the peak area ratios for extracted tetracyclines and the values derived from the respective calibration curve. The between-day precision of the method was established using milk samples at the same concentration levels as above. A triplicate determination of each concentration was conducted during routine operation of the system over a period of 5 days. RSD values for all the examined antimicrobials agents were lower than 12.7%, while mean apparent recovery rates were for OTC: 90.2–106%, for TC: 89.5–108%, for epi-CTC: 82.0–104%, for CTC: 98.8–104% and finally for DC: 91.7–103%. The results of the within-day repeatability and the between-day precision of the method applied to fortified matrix samples are given in detail in Table 4.

3.3.5. Analysis of certified reference material BCRÒ-492 The certified reference material BCRÒ-492 was used in order to check the trueness of the developed analytical method. The sample consists of a residue of lyophilised skimmed milk obtained from a bovine animal treated with oxytetracycline (OTC) at a concentration of 307 ± 14 lg/L. After preparation of the sample by the MSPD protocol described in Section 2.6, the measured concentration was 282 ± 12 lg/L for OTC leading to absolute recovery of 92.0%. 3.3.6. Stability Tetracyclines in milk samples were found to be stable for one week, when stored at 4 °C or 2 weeks at 20 °C. Extracted samples were found stable for at least 3 days at 4 °C and for 1 week, when stored at 20 °C, according to the degradation criterion of 15%. (Guidelines for the Validation of Analytical Methods used in Residue Depletion Studies, 2009). 3.3.7. Youden ruggedness Method ruggedness was estimated for minor changes by means of the ‘‘seven parameters/eight experiments’’ design, also known as the Youden and Steiner test. Seven different factors were chosen in the entire analytical process, because of their possible critical influence. These included sonication time, sorbent material, evaporation temperature, milk volume, concentration of oxalic acid, volume of elution solvent, and nature of elution solvent. Blank cow’s milk samples spiked at 100 lg kg1 were used for all the experiments. The Youden approach is a fractional factorial design. The basic idea is the introduction of several variations at once instead of is studying one alteration at a time. A set of A–G denote the nominal values for seven different factors that could influence the results, if their nominal values are changed slightly. Their alternative values are denoted by the corresponding lower case letters a–g. Eight determinations have to be made, which use a combination of the chosen factors (A–G). Observed results expressed as analyte’s amount from each Youden experiment is presented by letters ‘‘s– z’’ as shown in Table 1. Calculations for ruggedness testing of minor changes according to the Youden approach were conducted according to European Commission Decision 2002/657/EC. The results arising from the eight relevant robustness experiments are reported in Table 5. Sonication time, sorbent material and ACN elution volume, have a negative impact on OTC and a positive on epi-CTC. Evaporation temperature influences in a negative way all compounds and especially OTC. This effect was expected as

Table 5 Youden’s test results. Compound

SD (method)

SD

Da

Db

Dc

Dd

De

Df

Dg

OTC TC epi-CTC CTC DC

2.79 12.7 23.9 12.8 16.7

125 78.4 76.8 87.8 75.3

31.5 5.45 8.33 10.1 3.54

18.6 25.1 175 114 120

297 147 71.1 179 139

49.6 62.4 4.76 14.5 2.80

71.9 19.2 6.10 24.1 14.4

35.1 9.20 27.7 12.5 4.18

10.7 105 18.4 22.3 9.26

334

E. Karageorgou et al. / Food Chemistry 150 (2014) 328–334

tetracyclines are temperature sensitive. The nature of organic solvent at elution step influences positively all tetracyclines and the most TC in favor of acetonitrile. The amount of milk added has a positive impact on OTC and a negative on TC. Finally, CTC is negatively affected by oxalic acid concentration during elution step, while the same factor has a positive impact on OTC. 3.3.8. Application to real samples Three milk samples of different brand of each category described in Section 2.5, were analyzed, fifteen in total, according to the developed method and no tetracyclines residues were detected. 4. Conclusions In the present study, an ultrasound-assisted MSPD method was successfully applied to the multi-residue analysis of five tetracyclines residues in milk by HPLC-diode array detection. Validation was performed according to the EU Decision 2002/ 657/EC for the validation of an analytical method for residues in animal products. The results of validation process demonstrate that the method is suitable for application in European Union statutory veterinary drug residue surveillance programs. To the best of our knowledge, dispersive extraction assisted by sonication was applied successfully for the first time as a sample preparation technique for the determination of tetracyclines in milk. In this method, MSPD revealed high extraction yields and the endogenous compounds eluted were well resolved from target analytes, with a simple procedure that requires no special devices. The applicability of the method and its suitability to the routine analysis was proved by the analysis of fifteen milk samples, purchased from local market. Furthermore, the reliability of the proposed method was proved by its successful application to the analysis of the certified reference material BCRÒ-492. Acknowledgements This research has been co-financed by the European Union (European Social Fund – ESF) and Greek national funds through the Operational Program ‘‘Education and Lifelong Learning’’ of the National Strategic Reference Framework (NSRF) – Research Funding Program: Heracleitus II. Investing in knowledge society through the European Social Fund. References Cinquina, A. L., Longo, F., Anastasi, G., Gianneti, L., & Cozzani, R. (2003). Validation of a high-performance liquid chromatography method for the determination of oxytetracycline, tetracycline, chlortetracycline and doxycycline in bovine milk and muscle. Journal of Chromatography A, 987, 227–233. Commission Regulation (EU) No. 37/2010 of 22 December 2009, L.15/1-72. European commission decision 2002/657/EC, (2002). Official Journal of European, Communities, 221, 8–36. Frenich, A., Anguilera-Luiz, M., Vidal, J., & Romero-Gonzalez, R. (2010). Comparison of several extraction techniques for multiclass analysis of veterinary drugs in

eggs using ultra-high pressure liquid chromatography–tandem mass spectrometry. Analytica Chimica Acta, 661, 150–160. Fritz, J. W., & Zuo, Y. (2007). Simultaneous determination of tetracycline, oxytetracycline, and 4-epitetracycline in milk by high-performance liquid chromatography. Food Chemistry, 105, 1297–1301. Furusawa, N. (2003). Isolation of tetracyclines in milk using a solid-phase extracting column and water eluent. Talanta, 59, 155–159. Gao, F., Zhao, H., Wang, P., & Wang, J. (2013). Production of monoclonal antibody against doxycycline for immunoassay of seven tetracyclines in bovine muscle and milk. Journal of Environmental Science and Health, Part B: Pesticides, Food Contaminants, and Agricultural Wastes, 48, 92–100. Guidance for Industry. (2005). Q2 (R1) Validation of analytical procedure: Text and methodology. ICH. Guidelines for the Validation of Analytical Methods used in Residue Depletion Studies. (2009) Brussels: VICH Kaale, E., Chambuso, M., & Kitwala, L. (2008). Analysis of residual oxytetracycline in fresh milk using polymer reversed-phase column. Food Chemistry, 107, 1289–1293. Karageorgou, E., Myridakis, A., Stephanou, E., & Samanidou, V. (2013). Multi-residue LC-MS/MS analysis of cephalosporins and quinolones in milk following ultrasound assisted matrix solid phase dispersive extraction combined with QuEChERS methodology. Journal of Separation Science, 36, 2020–2027. Karageorgou, E., & Samanidou, V. (2010). Application of ultrasound-assisted matrix solid-phase dispersion extraction to the HPLC confirmatory determination of cephalosporin residues in milk. Journal of Separation Science, 33, 2862–2871. Karageorgou, E., & Samanidou, V. (2011). Development and validation according to European Union Decision 2002/657/EC of an HPLC–DAD method for milk multiresidue analysis of penicillins and amphenicols based on dispersive extraction by QuEChERS in MSPD format. Journal of Separation Science, 34, 1893–1901. Karageorgou, E., Samanidou, V., & Papadoyannis, I. (2012). Ultrasound-assisted matrix solid phase dispersive extraction for the simultaneous analysis of blactams (four penicillins and eight cephalosporins) in milk by high performance liquid chromatography with photodiode array detection. Journal of Separation Science, 35, 2599–2607. Kishida, K. (2011). Simplified extraction of tetracycline antibiotics from milk using a centrifugal ultrafiltration device. Food Chemistry, 126, 687–690. Mu, G., Liu, H., Xu, L., Tian, L., & Luan, F. (2012). Matrix solid-phase dispersion extraction and capillary electrophoresis determination of tetracycline residues in milk. Food Analytical Methods, 5, 148–153. Robert, C., Gillard, N., Brasseu, P., Pierret, G., Ralet, N., Dubois, M., et al. (2013). Rapid multi-residue and multi-class qualitative screening for veterinary drugs in foods of animal origin by UHPLC–MS/MS. Journal of Food Additives & Contaminants: Part A, 30, 443–457. Rodriguez, J., Espinosa, J., Aguilar-Arteaga, K., Ibarra, I., & Miranda, J. M. (2010). Determination of tetracyclines in milk samples by magnetic solid phase extraction flow injection analysis. Journal of Microchim Acta, 171, 407–413. Samanidou, V., Nikolaidou, K., & Papadoyannis, I. (2007a). Development and validation of an HPLC confirmatory method for the determination of seven tetracycline antibiotics residues in milk according to the European Union Decision 2002/657/EC. Journal of Separation Science, 30, 2430–2439. Samanidou, V., Nikolaidou, K., & Papadoyannis, I. (2007b). Advances in chromatographic analysis of tetracyclines in foodstuffs of animal origin – A review. Journal of Separation & Purification Reviews, 36, 1–69. Shariati, S., Yamini, Y., & Esrafili, A. (2009). Carrier mediated hollow fiber liquid phase microextraction combined with HPLC–UV for preconcentration and determination of some tetracycline antibiotics. Journal of Chromatography B, 877, 393–400. Tsai, W., Huang, T., Chen, H., Huang, J., Hsue, M., Chuang, H., et al. (2010). Determination of tetracyclines in surface water and milk by the magnesium hydroxide coprecipitation method. Journal of Chromatography A, 1217, 415–418. Youden, W. J., & Steiner, E. H. (1975). Statistical manual of AOAC. AOAC-I: Association of Official Analytical Chemists (pp. 33–36). Young, M., & Tran, K. (2011). Optimized Extraction and Cleanup Protocols for LCMS/MS Multi-Residue Determination of Veterinary Drugs in Edible Muscle Tissues. Waters application note. OASIS. Zhao, F., Zhang, X., & Gan, Y. (2004). Determination of tetracyclines in ovine milk by high-performance liquid chromatography with a coulometric electrode array system. Journal of Chromatography A, 1055, 109–114.