Analysis of Veterinary Drug and Pesticide Residues Using the Ethyl Acetate Multiclass/Multiresidue Method in Milk by Liquid Chromatography-Tandem Mass Spectrometry

A rapid and simple multiclass, ethyl acetate (EtOAc) multiresidue method based on liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) detection was developed for the determination and quantification of 26 veterinary drugs and 187 total pesticide residues in milk. Sample preparation was a simple procedure based on liquid–liquid extraction with ethyl acetate containing 0.1% acetic acid, followed by centrifugation and evaporation of the supernatant. The residue was dissolved in ethyl acetate with 0.1% acetic acid and centrifuged prior to LC-MS/MS analysis. Chromatographic separation of analytes was performed on an Inertsil X-Terra C18 column with acetic acid in methanol and water gradient. The repeatability and reproducibility were in the range of 2 to 13% and 6 to 16%, respectively. The average recoveries ranged from 75 to 120% with the RSD (n = 18). The developed method was validated according to the criteria set in Commission Decision 2002/657/EC and SANTE/11945/2015. The validated methodology represents a fast and cheap alternative for the simultaneous analysis of veterinary drug and pesticide residues which can be easily extended to other compounds and matrices.


Introduction
Veterinary drugs are widely used in medical and veterinary practices to treat and prevent disease as well as improve feed efficiency and increase animal growth rates [1]. Pesticides are also widely used to enhance food production by protecting food crops from potentially harmful and destructive pests [2]. However, the resulting occurrence of contaminants and/or residues in the human diet represents an issue of high concern.
According to the European Union, the maximum residue limit (MRL) in dairy milk is 100 g/kg for tetracycline and sulfenamide, 50 g/kg for macrolides and quinolones, and 10 g/kg for pesticides. Sensitive analytic methods have been developed to monitor and detect the MRL values in the dairy milk [3]. There are ultra-high pressure liquid chromatography mass spectrometry (UHPLC-MS/MS) methods reported to detect multiple residues of -lactams [4,5], as well as pesticides and mycotoxins [6], and some antihelminthic drugs and phenylbutazone [7].
Milk is a complex food that is high in fat and protein, and such ingredients may cause interactions in the analytical processes. Therefore, sample preparation is required, particularly in extraction and cleanup. Formerly, sample preparation methods were based on a few compounds or a single class of such drugs. Applying common extraction procedures and developing chromatographic conditions are difficult in multiclass and multiresidue analyses. Solid phase extraction methods have been applied, after the phases of protein precipitation and centrifugation, in order to observe the fluoroquinolones [8], veterinary drugs [9], mycotoxins, and pesticides in milk [6,10]. However, these methods are generally found to be time-consuming and require large volumes of organic solvents.

Journal of Analytical Methods in Chemistry
Multiresidue veterinary drugs that were developed for milk tests depend on various extraction and cleanup principles. One of the most accepted approaches is to dilute a sample of milk with a solvent like acetonitrile and then to centrifuge and evaporate the obtained supernatant organic extract [11,12]. Some multiclass analytical method applications by LC-MS/MS or LC-TOF/MS, related to homogenized or raw milk, that have the ability to specify undesirable chemicals, such as tetracycline, quinolone, sulfonamide, peptide, hormone, nonsteroidal anti-inflammatory anthelmintic drugs, mycotoxin, and pesticides, can be found in the literature [7]. Yet most of these methods are unable to offer satisfactory recovery of a large range of compounds of different polarities [13,14].
Most methods for the analysis of veterinary residues have some disadvantages, including high solvent consumption, tedious SPE cleanup steps that require extended time for analysis, and high costs. Therefore, these types of methods are not applied for routine analyses. The Quick Easy Cheap Effective Rugged Safe (QuEChERS) methodology, which was originally developed for pesticide analysis, has recently been proposed for the analysis of veterinary drugs using different matrices [15][16][17][18]. However, QuEChERS was found to be inconvenient for the recovery of polar veterinary drugs, including penicillin, tetracycline, and quinolone [13,18,19]. Therefore, there is still a great need for simple and rapid multiresidue analytical methods for simultaneously determining veterinary drug and pesticide residues in milk.
In this study, we prepared milk samples by using a procedure based on a simple liquid-to-liquid extraction. This method utilized a simple and quick sample preparation procedure using a single extraction step. Through this method, milk samples were analyzed for the determination of both veterinary drugs and pesticide residues by utilizing liquid chromatography-tandem mass spectrometry (LC-MS/MS). As a result, the reduced use of chemicals and steps in the sample preparation phase, together with the avoidance of a sample cleanup step, simplified the sample pretreatment and reduced the overall total cost. Finally, in addition to reducing analyses costs, the method provided a higher recovery of compounds of various polarities and improved the simplicity of detection efforts.

Samples.
All pasteurized whole milk samples were purchased from local markets. Also, raw milk was used for interference and specificity/selectivity as a blank.
Standard Solutions. Individual stock solutions of the veterinary drugs and pesticides were prepared in acetonitrile at a concentration of 1000 mg/kg. A mixed intermediate standard solution was prepared by diluting the stock standard solutions of the veterinary drugs and pesticides in acetonitrile at a concentration of 10 mg/kg. Stock and intermediate standard solutions were stored at 4 ∘ C in amber flasks and were found stable for at least 6 months.

Ethyl Acetate Extraction without Salting Procedure.
Milk samples, upon arrival at our laboratory, were kept at refrigerator temperature (10 ± 4 ∘ C) until analysis. For the preparation an aliquot of approximately 5 mL milk sample was pipetted in a 50 mL polypropylene centrifuge tube. Then, 200 mcL acetic acid was added to 10 mL of ethyl acetate. After vortex for 3 minutes, the mixture was centrifuged at 5000 rpm for 10 minutes. The upper phase was taken in 15 mL centrifuge tube and was dried under a gentle stream of nitrogen, and the residue was reconstituted with 1000 mcL of mobile phase A/mobile phase B (80/20). The sample was vortexed vigorously for 10 minutes. The extract was filtered through a 0.45 m filter prior to LC-MS/MS analysis.

Acetonitrile Extraction without Salting Procedure.
Approximately 5 mL milk sample was pipetted in a 50 mL polypropylene centrifuge tube. Then, 10 mL of acetonitrile and 200 mcL acetic acid were added to milk. After mixing by a vortex stirrer for 3 minutes, the mixture was centrifuged at 5000 rpm for 10 minutes. The upper phase was taken in 15 mL centrifuge tube and was dried under a gentle stream of nitrogen, and the residue was reconstituted with 1000 mcL of mobile phase A/mobile phase B (80/20). The sample was vortexed vigorously for 10 minutes. The extract was filtered through a 0.45 m filter prior to LC-MS/MS analysis.

QuEChERS Extraction
Procedure. Approximately 5 mL milk sample was pipetted in a 50 mL polypropylene centrifuge tube. Then, 2 g of magnesium sulfate and 1 g of sodium acetate were added to milk samples [15]. Then, 10 mL of acetonitrile and 100 mcL acetic acid were added to milk samples. After vortex for 3 minutes, the mixture was centrifuged at 5000 rpm for 10 minutes. The upper phase was taken in 15 mL centrifuge tube and was dried under a gentle stream of nitrogen, and the residue was reconstituted with 1000 mcL of mobile phase A/mobile phase B (80/20). The extract was transferred to a 2 mL Eppendorf microtube containing 50 mg PSA and 200 mg magnesium sulfate. Then, the tube was centrifuged at 4000 rpm during 5 minutes. The extract was filtered through a 0.45 m filter prior to LC-MS/MS analysis.  binary pump (Shimadzu UFLC LC-20AD model), Shimadzu automatic injector (Autosampler SIL-20A HT model), and a column oven (CTO-20AC). Analytical columns, Symmetry5 C18 2.1 × 150 mm id, 5 m particle size (Waters, Milford, MA), and Waters XTerra C18 150 mm × 2.1 mm id, 5 m particle size (Waters, Milford, MA), were tested. Chromatographic separation of veterinary drugs and pesticides was carried out on a Waters Symmetry C18 column. The method used a gradient mobile phase containing 0.1% acetic acid water and mobile phase B containing methanol. The column temperature was maintained at 40 ∘ C with a flow rate of 0.3 mL/min. The gradient profile was scheduled as follows: initial proportion (98% A and 2% B) for 0.3 minutes, linear increase to 80% (B) until 7 minutes, and hold of 80% (B) for 3 minutes. The injection volume was 50 L. The chromatographic system was coupled to electrospray ionization (ESI) source followed by an Applied Biosystems MDS SCIEX 4500 Q TRAP mass spectrometer. The MS/MS detector conditions were as follows: curtain gas 20 mL/min, exit potential 10 V, ion source gas 1 and ion source gas 2 set at 50 mL/min, ion spray voltage 5500 V, and turbo spray temperature set at 550 ∘ C. MS data were acquired in the positive ion ESI mode using two alternating MS/MS scan events. Two transitions were monitored for each analyte. The selected molecular ion and optimized collision voltages of product ions used for quantification, confirmation, and ion ratio were summarized in Table 1. Applied Biosystems SCIEX Analyst software version 1.6 was employed for data acquisition and processing. Quantification was by comparison with a six-point calibration (0.0, 0.01, 0.025, 0.05, 0.1, and 0.2 mg/kg) in matrix-matched calibration.

Validation Study.
The analytical method developed for determination of veterinary drug and pesticide residues in milk was validated according to EU Decision 2002/657/EC [16] and SANTE/11945/2015 [17]. The following parameters were evaluated in the validation procedure: selectivity, sensitivity, linearity, precision (intraday and interday reproducibility), accuracy and CC and CC , LOD, and LOQ.

Optimization of the Extraction Procedures.
Ethyl acetate extraction without salt procedure was chosen to be performed in this study because of its advantages. There was no need to use salt and it could give lower detection limit in terms of volatile characteristic of ethyl acetate.
Recovery values showed no difference among three different extraction procedures (acetonitrile extraction, QuECh-ERS extraction, and ethyl acetate extraction without salting procedure) ( Figure 1).
The recovery values expressed as recovery % are all within the reference range of 70-120%. Comparing three procedures, EtOAc without salt provided recoveries between 100% and 120% for a higher number of veterinary drugs and pesticides (26 veterinary drugs and 134 pesticides; total of 160 compounds) than QuEChERS (82 compounds) and ACN (100 compounds), as it can be observed in Figure 2. In terms of extraction recoveries, EtOAc was found to be a suitable extraction procedure for all 26 veterinary drugs and most of the pesticides analyzed in this study. Only one analyte (propham) showed > 120 for EtOAc.
Accuracy was evaluated in terms of relative standard deviation (RSD) by spiking blank samples with the corresponding volume of the multicompound working standard solution. RSD was evaluated at 50 g/kg by spiking six blank samples at each level for three procedures that provided similar RSD values. These values were within 1 < RSD < 10 for 75% of each analyte in the three procedures. These results indicated that the EtOAc without salt method was precise, accurate, and reliable for the analysis of the veterinary drug and pesticide compounds in the milk samples as an alternative method.

Validation Study
3.3.1. Selectivity. The selectivity of the method was assessed by duplicate analysis of 10 blank milk samples. No peaks of interfering compounds were observed within the intervals of the retention time of the analytes in any of these samples.

Linearity.
Linearity was evaluated from the calibration curves by triplicate analyses of blank milk samples fortified with the analytes at six (0.0, 0.01, 0.025, 0.05, 0.1, and 0.2 mg/kg) concentration levels. Linearity was expressed as the coefficient of linear correlation ( ) and from the slope of the calibration curve. The linearity of the analytical response across the studied range was excellent, with correlation coefficients higher than 0.997 for all analytes, which was similar to the findings in [19]. The authors [20] found correlation coefficients higher than 0.992 for all analytes, which was a lower score than ours.

Decision Limit and Detection
Capability. CC is defined as the limit at and above which it can be concluded with an error probability of that a sample is noncompliant. CC is defined as the smallest content of the substance that may be detected, identified, and/or quantified in a sample with an error probability of . The CC and CC were  determined by analysis of 10 blank milk samples and the signal-to-noise (S/N) ratio is calculated at the time window in which the analyte is expected. The CC values were calculated as three times the S/N ratio. The CC was calculated by analyzing 10 blank samples spiked with concentration at CC . Then the CC value was added up to 1.64 times the corresponding standard deviation. Then, a preliminary experiment was conducted to check if all compounds were detected when spiked at their CC level ( Table 2).
In Figure 3, very satisfactory S/N ratios were obtained for all analytes at LOQ level. The lowest LOQ value was 50 g/kg for tetracyclines and for sulfonamides 20 g/kg in veterinary drugs in [20] while it was 10 g/kg for both of them in our study except ciprofloxacin and quinolone. Figure 3 shows MRM chromatograms of milk samples at the lowest validation concentration at LOQ level.

Accuracy and Precision.
The accuracy was evaluated by recovery tests, analyzing fortified blank samples at the same concentration levels used in the precision tests (0.01, 0.025, and 0.05 mg/kg). The accuracy and precision of the method results (Table 2) confirmed the values given in Decision 2002/657/EC [16]. Thus, the mean accuracy values obtained in the recovery tests were between 61 and 130%. The precision of the method was determined in two stages: repeatability (intraday) and intermediate precision (interday). Repeatability was expressed by the RSD of the results from six replicates analyzed on the same day by the same analyst using the same instrument. The intermediate precision was expressed by the RSD of the results of eighteen analyses performed on three different days ( = 3), six analyses/day, by the same analyst using the same instrument. The relative standard deviation (RSD) of interday values of veterinary drugs and pesticides analyzed by the present method was 2 to 13% and for the intraday test 5-19% (Table 2), while relative standard deviation (RSDr) of intraday values was 4-26% in [20].

Matrix Effects.
Evaluation of matrix effect is important during validation of analytical methods using the LC-MS/MS technique. The ionization efficiency of the analytes in ESI source may be affected by matrix interference. In order to evaluate the degree of ion suppression or signal enhancement,            calibration curves were established with and without matrix. Matrix-induced effects were assessed by comparing the slopes of these calibration curves using the following formula: matrix effect (ME) = 1 − ( matrix / standard ) × 100, where matrix and standard are the slopes of calibration straight lines for standard and matrix-matched calibration graphs. The matrixmatched calibration curves were constructed using milk samples (5 g/mL matrix equivalent) prepared in MeOH-water solution with 0.1% acetic acid and spiked with veterinary drug and pesticides at concentration levels of 0.01, 0.025, and 0.05 mg/kg. Matrix effect was further evaluated for ion suppression between the standards prepared in pure solvent and standards prepared in matrix and the matrix effect was found to be in a range of 15-25%. These results showed that standard calibration which was simpler and less timeconsuming compared with matrix-matched calibration can effectively be used for quantitation of veterinary drug and pesticides in milk ( Table 2).

Real Samples.
The method used analyzed more than 220 milk samples submitted to the laboratory for veterinary drug and pesticide residues by the local markets. Two transition ion pairs were monitored for each of the analytes and the ion ratios of detected samples were compared well with those of standards. Retention times of analytes were also confirmed by addition of known standards in detected samples. Eight samples out of 220 milk samples were found to contain residues of veterinary drug and pesticide residues (4% incidence was positive). Sulfadiazine (veterinary drug) residue amount was found between 0.075 and 0.125 mg/L in 2 samples and tetracycline (veterinary drug) amount was found to be 0.015-0.100 mg/L in 4 samples. Carbaryl (pesticide) residue concentration level was 0.005-0.025 mg/L in 2 samples.

Conclusions
A multiclass/multiresidue procedure with LC-MS/MS detection has been developed and validated to determine and quantify veterinary and pesticide residues in milk. A simple sample preparation method involved liquid extraction salting out procedures in ethyl acetate system, without cleanup steps, and shortening the sample preparation time. Validation of the method was performed according to Commission Decision 2002/657/EC. The method was characterized by good results in terms of recovery, reproducibility, and repeatability allowing the detection of veterinary drug and pesticide residues below the recommended analytical level. Based on these results, LC-MS/MS method with ethyl acetate extraction showed the suitability for sensitive quantification of veterinary and pesticide residues in milk samples for food safety applications. The validated method was applied on 220 real commercial samples. This short protocol can be applicable to a large number of samples for routine analysis and rapid detection.