Journal of Chromatography B, 945–946 (2014) 31–38
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Liquid chromatography and ion trap mass spectrometry for simultaneous and multiclass analysis of antimicrobial residues in feed water Chusak Ardsoongnearn a , Ongart Boonbanlu a , Sunan Kittijaruwattana a , Leena Suntornsuk b,c,∗ a
Bureau of Quality Control of Livestock Products, Department of Livestock Development, Bangkok 10400, Thailand Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Mahidol University, Bangkok 10400, Thailand c Center of Excellence for Innovation in Drug Design and Discovery, Faculty of Pharmacy, Mahidol University, Bangkok 10400, Thailand b
a r t i c l e
i n f o
Article history: Received 17 July 2013 Accepted 17 November 2013 Available online 25 November 2013 Keywords: Nitrofurans Nitroimidazoles Chloramphenicol HPLC-MS Feed water
a b s t r a c t This work firstly reported the development of liquid chromatography coupled to an ion trap mass spectrometer (LC–MS ion trap) for the simultaneous determination of nitrofurans (e.g. nitrofurazone (NFZ), nitrofurantoin (NFT), furazolidone (FZD) and furaltadone (FTD)), nitroimidazoles (e.g. metronidazole (MNZ), ronidazole (RNZ) and dimetridazole (DMZ)) and chloramphenicol (CAP) in feed water. Isotope-labeled internal standards for the corresponding target analytes were employed to prevent matrix effects that might lead to signal suppression/enhancement. High performance liquid chromatography (HPLC) analysis was performed on a Prodigy ODS-3 column, 2.0 mm × 150 mm, 5 m with a guard cartridge at a flow rate of 0.2 mL/min, column oven temperature of 40 ◦ C, and an injection volume of 10 L. Solid phase extraction (SPE) procedures, factors affecting HPLC separation (e.g. buffer pH and concentrations) and mass spectrometry (MS) parameters were optimized. After an off-line SPE by the OASIS HLB cartridges (with an enrichment factor of 400), the eight antimicrobial agents were separated in 18 min using a gradient elution of acetonitrile in acidified water (pH 5.0). MS detection was by an ion trap MS coupled with electrospray ionization (ESI) in tandem mass spectrometry mode (MS/MS) using the nebulizer gas at 35 psi, drying gas at 9 L/min and drying temperature of 325 ◦ C. Method linearity was good (r2 = 0.979–0.999) with acceptable precision (% RSDs = 3.4–26.6%) and accuracy (%recovery = 88.4–110.1%). Very low limits of detection (LOD) and quantitation (LOQ) were achieved in ranges of 0.002–0.06 g/L and 0.005–0.25 g/L, respectively. The established method is successfully employed by the Department of Livestock Development of Thailand for the monitoring of the drug residues in feed waterbecause of its convenience, reliability and high sensitivity. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Presently, food safety is an issue of public concern, since many prohibited substances are still detected in food products. According to Commission Regulation (EU) No. 37/2010, these substances include Aristolochia spp. and preparation thereof, chloroform, chlorpromazine, colchicines, dapsone and antimicrobials (e.g. chloramphenicol, dimetridazole, metronidazole, nitrofurans (including furazolidone), and ronidazole) [1]. Aristolochia spp. and
∗ Corresponding author at: Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Mahidol University, Bangkok 10400, Thailand. Tel.: +66 2 644 8695; fax: +66 2 644 8695. E-mail address: [email protected] (L. Suntornsuk). 1570-0232/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jchromb.2013.11.034
preparation thereof, chloroform, chlorpromazine, colchicines, dapsone rarely have any uses in livestock, whereas the antimicrobials were previously used for treatments of infection in livestock. Nitrofurans (e.g. furazolidone (FZD), furaltadone (FTD), nitrofurantoin (NFT) and nitrofurazone (NFZ) are antibiotics that were previously used in livestocks as growth promoters and for the prevention and treatment of gastrointestinal infections caused by Escherichia coli, Salmonella spp., Mycoplasma spp., Coccidia spp., coliforms and some other protozoa [2]. Nitroimidazoles are veterinary drugsthat were used for curing and prevention of certainbacterial and protozoal diseases in poultry as well as for swine dysentery, thus they are classified as coccidiostats. The most commonly used nitroimidazoles are metronidazole (MNZ), dimetridazole (DMZ), and ronidazole (RNZ). Both nitrofurans and nitroimidazoles are mutagenic and carcinogenic [3–5]. For this reason, the EU has prohibited their uses
32
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
in feed for food-producing species as described in Annex IV of the Council Regulation (EC) 2377/90 [6]. Chloramphenicol (CAP) is considered as a potent antibiotic for treating pneumonic and enteric disorders. CAP has been recommended for prevention of secondary infections associated with chronic respiratory diseases in poultry. In certain susceptible individuals, CAP is associated with serious toxic effects in human such as bone marrow depression, and fatal a plastic anemia [7,8]. Since this condition is dose independent, CAP has been banned for uses in food producing animals in many countries including EU and Thailand. The incidents of nitrofuran contamination in poultry and seafood exported from Thailand [9] and Brazil [10] to EU during 2002 have prompted more than 150 rapid alerts issued by the EU [11]. Since then, the Thai government has implemented several restricting measures from farm to slaughter houses to ensure the zero tolerance level of these residues. According to Commission Regulation (EU) No. 37/2010, the maximum residue levels (MRL) of these residues cannot be established [1]. Among these measures, regular monitoring of the prohibited substances in feed, feed water, seafood, poultry and meat products has been enforced for registered farms throughout the country. Many analytical methods have been used for monitoring of nitrofurans, nitroimidazoles and CAP residues in various matrices. For example, high performance liquid chromatography (HPLC) and ultra HPLC coupled with photodiode array and mass spectrometric (MS) or tandem MS detectors were employed for determination of nitrofurans and their metabolites in animal feed [2,12–14], poultry muscle and shrimp [15], eyes of broiler chickens [16], animal plasma samples [17], salt [18] and honey [19]. LC-UV, MS or tandem MS of nitroimidazoles and their metabolites in animal feed [20], egg and egg-based samples [21,22], swine kidney [23], poultry muscle [24], pork [25,26], meat [27], animal plasma [28] and waste water [29] were also reported. Furthermore, the techniques were valuable for analysis of CAP in urine [29–31], feed [30], milk [30,32–34], shrimp [35], chicken [36] and environment and food [31,37]. This project was the collaboration between Mahidol University and Veterinary Drugs Assay Division (VDA), Department of Livestock Development, Thailand. The VDA division focuses on quality control of veterinary drugs and monitoring the misuses/abuse of such drugs. In the past, problems of veterinary drug abuses (e.g. the addition of some antimicrobials or growth promoters into feed or feed water) were found in many farms. Previously, HPLC-DAD was employed as a screening method in the monitoring programs of nitrofurans, nitroimidazoles and chloramphenicol residue in feed water. However, there is an urgent need for a confirmatory method with higher sensitivity for simultaneous analysis these residues. This work aimed to develop a LC–MS method for quantitation of nitrofurans (e.g. NFZ, NFT, FZD and FTD), nitroimidazoles (e.g. MNZ, RNZ and DMZ) and CAP (Fig. 1) in feed water since, to the best of our knowledge, there is no report on the multi-residue analysis of these drugs in this matrix. Optimization of solid phase extraction (SPE) procedures, HPLC conditions and MS parameters were performed. The method was validated and applied to analyze the drug residues in forty feed water samples collected from animal farms in Thailand. The method can be routinely applied to quality control of these banned substances in feed water by the Thai officials.
2. Materials and methods 2.1. Reagent Standards of nitrofurans (e.g. FZD, FTD, NFT and NFZ) nitroimidazoles (e.g. MNZ, DMZ and RNZ) and CAP were obtained from Sigma (St. Louis, USA). Isotope-labeled standards (NFZ-13 C-15 N2 ,
NFT-13 C3 , FZD-D4 , FTD-D5 , MNZ-13 C2 -15 N2 , RNZ-D3 and DMZ-D3 ) were used as internal standards (IS) and were from Witega (Berlin, Germany), except CAP-D5 was from Dr. Ehrenstorfer (Augsburg, Germany). Glacial acetic acid and formic acid were AR grade from Merck (Darmstadt, Germany), ammonium acetate, ammonium formate and acetonitrile AR grade were purchased from BDH (Leicestershire, United Kingdom). High purity water was obtained by a MilliQ system from Millipore (Bedford, USA). Other solvents were of HPLC grade from Labs can Asia (Bangkok, Thailand). Adjustment of mobile phase pH was done by adding 0.1% formic acid or acetic acid. Acidified water was prepared by addition of dilute acetic acid to water to obtain pH 5.0. 2.2. Standard and sample preparation Standards and isotope-labeled internal standards were accurately weighed and diluted with acetonitrile to concentrations of 0.5 and 0.2 mg/mL, respectively. The stock solutions were protected from light and kept in a freezer (−20 ◦ C). Intermediate solutions were subsequently diluted with acetonitrile to a concentration of 25 g/mL and kept at 4 ◦ C. Working solutions were prepared daily by diluting the stock standard solutions with acetonitrile to appropriate concentrations for linearity and range experiments. The stocks of standard and internal standard solutions were stable at 4 ◦ C for one month. Forty feed water samples from chicken and pig farms, collected from various provinces of Thailand, were sampled by regional Department of Livestock Development officers. The samples were analyzed without pH adjustment since their pH was between 6.0 and 8.0, which was suitable for SPE using Oasis HLB cartridges. Sample blank solutions were pooled feed water sample (20 L) that was previously analyzed by HPLC-DAD to prove they were antimicrobial-free. Then, they were used for method validation experiments and for preparing blind samples. Blind samples were prepared by spiking sample blank solutions with different concentrations of standards and fixed amounts of the corresponding internal standards. Prior to SPE, two hundred milliliters of the water were transferred into a volumetric flask, 1 mL of 2 M sodium thiosulphate was added into the flask, mixed and left for 30 min before being spiked with 0.5 mL and 0.2 mL of mixtures of the eight IS and known amount of working standard (control samples). Samples were mixed thoroughly and stood for 10 min before loading onto SPE cartridges. Each sample was analyzed by HPLC-DAD prior by HPLC-MS (n = 2). If HPLC-DAD showed any positive results, the samples would be confirmed by HPLC-MS in six replicates. One round of each batch analysis consisted of standard solutions at eight different concentrations, 20 feed water samples and six blind samples. The spiking procedure for the blind samples was randomly predetermined using the software available at www.random.org. 2.3. SPE procedure Off-line SPE using Oasis HLB cartridges (200 mg/6 mL) from waters (Milford, USA) was employed for sample pre-concentration. The cartridges were preconditioned with 5 mL of methanol and 5 mL of water. Two hundred milliliters of feed water sample were loaded onto the cartridges at a rate of 4 mL/min using a vacuum manifold. The cartridges were washed with 5 mL of 5% methanol and dried under vacuum to remove water residues and eluted with 4 mL of eluent (2 mL × 2). Eight different eluents consisting of methanol, acetonitrile, and their mixture with different amounts of acetic acid (0.5–2%) were optimized for the appropriate elution of all eight antimicrobial drugs. The eluate was dried using a rotary evaporator, reconstituted with 0.5 mL of 20 mM ammonium acetate
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
33
Fig. 1. Structures of the investigated compounds.
(pH 4.5): methanol (80:20, v/v), vortexed for 30 sand sonicated for 15 s. Then, it was filtered through a 0.45-m syringe nylon filters prior analysis. 2.4. Liquid chromatography-ion trap mass spectrometry HPLC analysis was performed on a HP1200 series from Agilent Technologies (Santa Clara, USA), equipped with an automatic sampler/injector, a degasser, a binary pump and a column oven. The LC instrument was coupled with a Bruker Daltonicion trap mass spectrometer, equipped with an electrospray (ESI) interface (Bremen, Germany). The LC–MS system was controlled by HP workstation XW 6200 using Bruker Daltonic Esquire 6.1 software comprised of Esquire Control version 6.1, Data Analysis version 3.4, HyStar version 3.2 and Quant Analysis version 1.8 for data analysis. Chromatographic separation was in a reversed phase mode using a Prodigy ODS-3 (15 cm × 0.20 cm with a 5 m particle size) column from Phenomenex (Torrance, USA) coupled with a guard cartridge. The column temperature was set at 40 ◦ C and the injection volume was 10 L. The flow rate of the binary mobile phase was 0.2 mL/minusing either isocratic or gradient profiles. Various mobile phases, including solvent A (aqueous buffer of ammonium formate and formic acid, ammonium acetate and acetic acid at various concentrations and pH or acetonitrile in acidified water) and
solvent B (acetonitrile, methanol, acetonitrile in acidified water), were optimized to obtain reasonable resolution, peak shape and run time. Ionization for MS of MNZ, RNZ, DMZ, FTD and FZD was carried out in a positive mode, while that of NFZ, NFT and CAP was in negative mode [2,38,39]. Precursor ions, product ions from MS/MS reactions and fragmentation amplitudes were studied by direct infusion of 1 mg/Lof each standard solution into an electrospray ionization source at a flow rate of 4 L/min, using ultra-high pure nitrogen as nebulizer and drying gas at 35 psi with a flow of9 L/min, and a drying temperature of 325 ◦ C. Various parameters such as capillary voltage, ion optics (e.g. skimmer, octapole, lens, trap drive voltage), and fragmentation voltage were varied to obtain the highest intensity of precursor and product ions of each compound. For peak identification, one precursor and two product ions must be detected with at least 4 identification points according to Commission Decision 2002/657/EC [40]. The four points are required for Group A substances, which include the current investigated drugs (e.g. MNZ, RNZ, DMZ, NFZ, NFT, FZD, FTD and CAP). For mass spectrometric detection, the relative intensities of the detected ions, expressed as a percentage of the intensity of the most intense ion, shall correspond to those of the calibration standard, at comparable concentration within the reported tolerances.
34
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
2.5. Method validation Validation protocol for the analysis of the target analytes in animal feed water is not yet available in EU guideline. Thus, a validation protocol was established to prove that the method performance was fit for the purpose. 2.5.1. Linearity For the calibration procedure, the internal standard calibration method was applied to this study to overcome matrix effects. These effects are usually unknown or variable and are major limitations for trace analysis by LC-ESI-MS. NFZ-13 C-15 N2 , NFT-13 C3 , FZD-D4 , FTD-D5 , MNZ-13 C2 -15 N2 , RNZ-D3 , DMZ-D3 and CAP-D5 were used as internal standards for NFZ, NFT, FZD, FTD, MNZ, RNZ, DMZ and CAP, respectively. Linearity of method was performed from blank samples spiked with eight different concentrations of the standards and fixed amounts of the internal standards. The spiked solutions were subjected to SPE, in the same manner as feed water samples (Section 2.3), prior to HPLC-MS analysis. Linearity of system was evaluated by spiking eight different concentrations of the standards and fixed amounts of the internal standards into 4.0 mL of eluent 6 (1% acetic acid:methanol, 80:20), which was the optimal eluent in SPE. The solutions were dried and reconstituted as previously described (Section 2.3). Linearity of method and system was examined by analyzing standard solutions in triplicate at eight concentrations on three different days. Calibration curves were plots of peak area ratios against concentration ratios of the target compounds and IS. Slopes, intercepts and correlation of determinations (r2 ) were calculated using Excel® . 2.5.2. Precision and accuracy Method precision and accuracy was evaluated from intra-day and inter-day measurements by analyzing the spiked sample blank solutions (n = 10) on the same and three different days. The sample blank solutions were spiked with three concentrations of the standards (low, mid and high points of the calibration curve). Precision was reported as % relative standard deviations (% RSDs) and accuracy was represented as % recovery (% R) of the assays. 2.5.3. LOD and LOQ Method sensitivity was evaluated by limits of detection (LOD) and quantitation (LOQ). LOD is defined as the lowest concentration at which an analyte can be differentiated from the background noise, whereas LOQ is the lowest concentration at which an analyte can be accurately quantified. In this study, LOD and LOQ were based on the variability of the blank [41]. Ten blank samples were spiked with the lowest acceptable concentration (0.013 g/L for MNZ, 0.005 g/L for RNZ and CAP, 0.015 g/L for DMZ, 0.035 g/L for NFZ, 0.025 g/L for NFT and 0.001 g/L for FZD and FTD) and analyzed by HPLC-MS. Based on the standard curve, the calculated LOD and LOQ were expressed as the analyte concentrations corresponding to peak area ratios of the mean value +3SD and +10SD, respectively. Then, standards at calculated LODs and LOQs were spiked into the blank samples. Signal to noise ratios (S/N) of peak height ratios were monitored and LOD and LOQ were adjusted till the S/N of not less than 3 and 10, respectively, were obtained. 2.5.4. Robustness Method robustness is an evaluation of the method capacity to remain unaffected by small, but deliberate changes in analytical conditions. Currently, pH of the mobile phase greatly influenced the analyte retention. Thus, robustness testing was carried out by varying the mobile phase pH ± 0.5 around the optimal value and % RSDs of retention times were determined.
Fig. 2. Recovery of the analytes from different eluents in SPE.
3. Results and discussion 3.1. SPE optimization Off-line SPE was required in this study to achieve the desirable sensitivity. Oasis HLB cartridges were chosen for water sample pretreatment and pre-concentration (enrichment factor of 400) since the cartridges was efficient for analytes with different polarity. Eluent 1 (methanol) provided good recovery for MNZ, RNZ, DMZ and NFT (% R between 81.5 and 105.4%), but gave low recovery for NFZ and CAP (% R ∼ 19.4%) and could not elute FZD and FTD. Interestingly, eluent 2 (acetonitrile) still gave acceptable recovery for MNZ, RNZ, DMZ and NFT, and significantly improved the recovery of NFZ, FZD, FTD and CAP to 61.5, 106.5, 1002 and 53.2%, respectively. Subsequent work focused on increasing the recovery for NFZ and CAP. A mixture of methanol and acetonitrile at 50:50 and 80:20 (eluents 3 and 4, respectively) did not enhance the recovery of NFZ and CAP, thus addition of acetic acid at 0.5 (eluent 5), 1.0 (eluent 6), 1.5 (eluent 7) and 2.0% (eluent 8) to the mixture of methanol and acetonitrile (80:20, eluent 4) was investigated as eluents. Among eluents 5 to 8, the highest recovery for CAP (∼67.5%) was observed in eluent 6. In addition, eluent 5 gave the lowest recovery for MNZ (57.0%). Thus, the eluent 6 was selected and the recovery for all analytes in this eluent ranged from 60.1 to 106.0% (Fig. 2). The eluate was dried and reconstituted with 20 mM ammonium acetate (pH 4.5): methanol (80:20, v/v) since it provided higher sensitivity than that from buffer of pH 5.0.
3.2. HPLC optimization Initial experiments were performed by isocratic elution using methanol or acetonitrile together with aqueous ammonium formate or ammonium acetate (5–40 mM, pH between 3.9 and 4.8). Results revealed that the resolving power of isocratic elution was not sufficient to separate all of the analytes, especially MNZ, RNZ and DMZ, which were overlapping under most mobile phase compositions. Moreover, run time in isocratic elution was unsatisfactory (up to 50 min per injection). Thus, gradient elution of the investigated substances using ammonium formate or ammonium acetate (10 and 20 mM, pH 3.5 and 5.0) and acetonitrile was optimized. Improvement of resolution and reduced retention time were observed in buffer at low concentrations. Ammonium acetate buffer (pH 5.0) was preferred since reasonable separation between DMZ, FTD and NFZ could be obtained. Separation of these analytes in ammonium formate buffer (pH 3.6) was more difficult due to the narrow separating window FTD and the retention time shift during long batch run.
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
35
Table 1 MS/MS reactions monitored by LC-ESI MSn . Analyte Positive mode [M+H]+ MNZ MNZ-13 C2 -15 N2 RNZ RNZ-D3 DMZ DMZ-D3 FZD FZD-D4 FTD FTD-D5 Negative mode [M−H]− NFZ NFZ-13 C-15 N2 NFT NFT-13 C3 CAP
Retention time (min)
Exact mass
MS/MS reactions (m/z)
5.89
171.06
6.74
200.05
9.32
141.05
13.69
225.04
16.43
324.11
172.0 > 128.0 176.0 > 132.0 201.0 > 140.0 204.2 > 143.0 142.0 > 96.2 145.2 > 99.2 226.0 > 113.1 229.8 > 117.1 325.2 > 281.1 330.2 > 286.2
10.96
198.04
12.11
238.03
18.63
322.01
Finally, the gradient elution of buffer-free mobile phase (solvent A: 10% ACN, solvent B: 30% ACN in acidified water, pH 5.0) was employed. The gradient was as follows: 0 min (0% B), 12 min (100% B) and 15 min (100% B), then the mobile phase was returned to the initial condition in 1 min and equilibrated for another 6 min. The mobile phase offered the simple and convenient column equilibration and clean up. Reasonable run time of 18.7 min with a resolution of >3.0 could be achieved for the target analytes (Fig. 3a).
196.8 > 149.9 199.8 > 152.9 236.8 > 151.9 239.8 > 151.9 321.0 > 257.0
Capillary (volt)
−3950 −3950 −3950 −4183 −4242
3775 3075 3717
Fragmentation amplitude (volt) 0.75 0.67 0.73 0.70 0.66 0.74 0.63 0.63 0.67 0.67 0.80 0.77 0.84 0.73 1.21
3.3. MS optimization Electrospray was used as the ionization source. Selection of an ionization mode for individual analyte was based on the MS optimization results and previous reports [2,38,39]. Tuning of MS parameters by direct infusion of the target compounds and IS was performed using the optimization function in Esquire Control version 6.1. Fine tune MS parameters were listed in Table 1. These parameters were specified for each time segment and were used for all experiments. Chromatograms with MS detection of each compound were accomplished with the previously optimized HPLC conditions (Fig. 3b). 3.4. Validation of the LC–MS method Method validation experiments were performed using real feed water samples that were previously analyzed to prove they were blank samples (antimicrobial-free samples). The matrixmatch standards (spiked blank samples) and internal standards were employed to minimize matrix effects, hence improvement of method precision and accuracy.
Fig. 3. (a) HPLC-MS chromatograms of the investigated compounds (1 mg/L) and (b) time segments for measuring specific transitions of analytes are indicated in the chromatogram. HPLC condition: prodigy ODS (3), 2.0 × 150 mm, 5 m with guard cartridge, gradient elution of A: B (10%: 30% acetonitrile in acidified water (pH 5.0), 0 min B 0%, 12 min B 100%, 15 min B 100%, 16 min B 0%, 22 min B 0%, 40 ◦ C, flow rate: 0.2 mL/min. MS condition; see Section 2.4 and Table 1.
3.4.1. Linearity The proposed method showed good linearity of method (matrixmatched standard) and system (matrix-free standard) in the investigated ranges with r2 greater than 0.993 for all target analytes (Table 2). Percent differences of slopes between matrix-matched standards and matrix-free calibration curves of greater than 10% indicated the presence of matrix effects that normally occur in electrospay ionization MS. Matrix effects can also affect analysis of residues in trace amounts, such as in environmental samples, which results in unsatisfactory accuracy and precision. In the present study, % difference between matrix-matched standards and matrixfree standard calibration curves were within 10% for most analytes. Additionally, matrix effects could be pronounced since off-line SPE was performed for sample enrichment. To investigate this, ratios of slopes between the matrix-matched and matrix-free standards were calculated and degrees of signal suppression or enhancement could be identified. For example, the slope ratios between 0.5 and 0.8 indicated moderate signal suppression, between 0.8 and 1.2 corresponded to signal suppression or enhancement of less than 20% and from 1.2 to 1.5 represented moderate signal enhancement [42]. Our results revealed the ratios from 1.0 to 1.2 (Table 2), which indicated that no signal suppression was observed and moderate signal enhancement was found only in FTD. The sample matrices did not show any strong signal suppression or enhancement.
36
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
Table 2 Linearity, LOD and LOQ dataa . Range (g/L) 0.1–0.7 0.03–0.21 0.2–1.4 0.25–1.75 0.2–1.4 0.005–0.035 0.01–0.07 0.025–0.176
MNZ RNZ DMZ NFZ NFT FZD FTD CAP a
Method linearity
System linearity
RS
LOD (g/L)
LOQ (g/L)
y = 0.779x + 0.042 (r2 = 0.995) y = 0.993x + 0.013 (r2 = 0.997) y = 0.778x + 0.063 (r2 = 0.997) y = 0.347x − 0.028 (r2 = 0.996) y = 0.913x − 0.064 (r2 = 0.995) y = 0.975x + 0.035 (r2 = 0.996) y = 1.148x + 0.001 (r2 = 0.995) y = 1.112x − 0.013 (r2 = 0.996)
y = 0.729x + 0.045 (r2 = 0.993) y = 0.936x + 0.008 (r2 = 0.997) y = 0.775x + 0.027 (r2 = 0.996) y = 0.338x − 0.020 (r2 = 0.998) y = 0.809x − 0.031 (r2 = 0.999) y = 0.937x + 0.036 (r2 = 0.997) y = 0.939x − 0.016 (r2 = 0.996) y = 1.017x + 0.005 (r2 = 0.997)
1.07 1.06 1.00 1.03 1.13 1.04 1.22 1.09
0.03 0.01 0.05 0.06 0.04 0.002 0.002 0.01
0.10 0.03 0.20 0.25 0.20 0.005 0.01 0.025
RS is ratios of slope from the matrix-matched and matrix-free standard, LOD and LOQ are limits of detection and quantitation, respectively.
Table 3 Precision dataa . MNZ Conc. (g/L) Intra-day Inter-day
0.1 13.21 14.14
RNZ 0.4 9.52 10.00
0.7 8.35 9.40
NFT Conc. (g/L) Intra-day Inter-day a
0.2 6.26 7.23
DMZ
0.03 13.79 14.31
0.12 8.53 8.76
FZD 0.8 7.91 9.14
1.4 7.32 7.54
0.005 10.00 10.71
0.2 13.10 13.94
0.8 9.01 9.20
1.4 9.77 9.93
FTD 0.02 7.33 7.44
0.035 7.47 7.40
0.01 18.25 19.61
0.25 9.33 9.39
1.0 10.18 10.11
1.75 9.88 10.44
CAP 0.04 12.78 14.44
0.07 12.90 14.15
0.025 11.70 11.60
0.1 10.41 10.60
0.175 10.11 10.26
Data is represented as percent relative standard deviation of the assay.
3.4.2. Precision and accuracy Method precision was represented by % RSDs from intra- and inter-day analysis. The uses of suitable isotope-labeled internal standards offered acceptable precision for all target analytes with % RSDs between 6.26 and 18.25% for intra-day precision and between 7.23 and 19.61% for inter-day precision (Table 3). Generally, % RSDs from intra-day precision were slightly lower than those from interday precision. For most analytes, higher % RSDs were obtained at low concentrations (the lowest points of calibration curves), which were close to LOQs of the analytes. High % RSDs were obtained for both intra- and inter-day assay of FTD (between 12.78 and 19.61) due to the moderate signal enhancement from matrix effects, whereas % RSDs from NFT were in lower ranges (6.26–9.14). Accuracy of the method was represented by recovery of the spiked blank samples from the same and different day analysis. Recoveries were higher than 88% with % RSDs of less than 26.6% for all target analytes. The intra-day and inter-day recoveries were between 88.4 and 110.1% and between 94.1 and 103.2%, respectively (Table 4). The lowest recovery with highest %RSD found in FTD might stem from overlapping of recovery with matrix effects [42]. However, these values were within acceptable ranges (50–120%) recommended by AOAC [41]. 3.4.3. LOD and LOQ LODs and LOQs were estimated from feed water samples spiked at very low concentrations, which provided S/N of 3 and 10, Table 4 Accuracy dataa .
%R MNZ RNZ DMZ NFZ NFT FZD FTD CAP
90.1–103.6 91.3–103.7 96.0–106.8 94.8–110.1 97.6–106.3 93.5–104.2 88.4–104.1 96.1–105.6
respectively, based on visualization of peak height. LODs of all target analytes were from 0.002 to 0.06 g/L (Table 2). The lowest LODs were measured in FZD and FTD (0.002 g/L) and the highest in NFZ (0.06 g/L). Comparing to the reported LODs, which varied from 0.005 to 1.6 g/L [4,11,26,28,35], the proposed technique could provide lower LODs indicating the high sensitivity of the method. LOQs were found between 0.005 and 0.25 g/L, which were the lowest concentration levels for each analyte in recovery studies. The lowest LOQ corresponded to FZD (0.005 g/L) and the highest to NFZ (0.25 g/L). Fig. 4 represents the HPLC–MS chromatograms of the feed water sample and the sample spiked with the standards at LOQ concentrations (Table 2). Although no MRL can be established for the investigated analytes, the LOD and LOQ were considered to be in very low levels.
3.4.4. Robustness Modification of mobile phase pH could affect retention of the analytes and their resolution. Results showed that the method was robust upon the alteration of mobile phase pH (4.5, 5.0 and 5.5), which gave the % RSDs of the retention time within 1.25% (Table 5). Validation data confirms that the method is precise, accurate and sensitive for trace analysis of NFZ, NFT, FZD, FTD, MNZ, RNZ, DMZ and CAP in feed water.
Table 5 Robustness data*.
Intra-day
a
0.21 9.66 9.54
NFZ
Inter-day % RSD 5.3–16.1 7.6–16.0 4.9–17.2 6.3–14.3 3.4–11.3 5.5–11.6 9.2–26.6 8.4–15.1
%R 94.4–99.8 95.1–100.8 100.1–102.7 99.7–107.0 100.6–102.2 96.4–101.3 94.1–95.5 99.4–103.2
Analyte % RSD 9.4–14.1 8.8–14.3 9.2–13.9 9.4–10.4 7.2–9.1 7.4–10.7 14.1–19.6 10.3–11.6
% R is percent recovery, % RSD is percent relative standard deviation.
MNZ RNZ DMZ NFZ NFT FZD FTD CAP
Retention time from various pH of water in mobile phase (min) pH 4.5
pH 5.0
pH 5.5
% RSD
6.1 7.0 9.5 11.1 12.3 13.8 16.8 18.9
6.1 6.9 9.4 11.1 12.2 13.7 16.5 18.8
6.1 7.0 9.4 11.1 12.0 13.8 16.4 19.1
0.32 1.00 0.20 0.35 1.25 0.42 1.20 0.87
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
37
Fig. 4. HPLC-MS chromatograms of (a) the feed water sample and (b) the sample spiked with the standards at LOQ levels. HPLC and MS condition: see Fig. 3.
3.4.5. Application Forty feed water samples from different farms were tested for the drug residues. In each batch analysis, six blind samples were simultaneously analyzed as quality control samples. Recoveries of the blind sample batches were between 68.1 and 117.6%. The low recovery of RNZ (68.1%) in the batch analysis might due to the low spiking concentration that was close to the LOQ of RNZ (0.03 g/L) and the high %RSD (about 14%). However, this value was acceptable according to AOAC Requirements for Single Laboratory Validation of Chemical Methods, which allows the recovery of 50–120% for a concentration of ≤1 ppb [41]. This low recovery had no negative effects on the quantitation of the antimicrobials in real samples since other validation parameters were within acceptable ranges. In the two batch analysis, no residues of NFZ, NFT, FZD, FTD, MNZ, RNZ, DMZ and CAP were found in the forty feed water samples. This finding may reflect the efficiency of residue control measures in livestock production by the Department of Livestock Development of Thailand. These preventive measures have been strictly applied since 2002 to assure the food safety concept (from Farm to Table). 4. Conclusions A HPLC-MS method has been established for residue analysis of eight multiclass antimicrobial drugs, including nitrofurans, nitroimidazoles and chloramphenicol, at sub-ppb levels in feed water. The simple gradient elution of acetonitrile in acidified water in 22 min was efficient for separation of the drugs. The SPE procedure using 1% acetic acid in a mixture of acetonitrile and methanol (80:20) as the eluent could provide good recoveries of the drug residues from sample matrices with an enrichment of 400. All analytical performance characteristics meet the requirements and the method is suitable for the intended purposes. The SPE scheme and MS detection were valuable for the current work enabling the LOD at trace levels (0.002–0.06 g/L), which are lower than reported
values. The use of matrix-match standards and internal standards enhanced method precision and accuracy, while minimized matrix effects. Novelty of this work includes the high sensitivity for simultaneous and multiclass analysis of drug residues in feed water. In addition, the use of buffer-free mobile phase, resulting in fast and convenient column equilibration and clean up and reduced analysis time and cost. The method can be applied to routine analysis in regulatory departments, where sample throughput and sensitivity are priority.
Acknowledgements This project is supported by the Office of the High Education Commission and Mahidol University under the National Research Universities Initiative. The authors thank the Bureau of Quality Control of Livestock Products, Department of Livestock Development for providing all instruments and most chemicals. Thanks also go to Prof. Wolfgang Buchberger, Institute of Analytical Chemistry, Johannes Kepler University, Linz, Austria, for suggestions on the manuscript. Thanks also go to Mr. Brompoj Prutthiwanasan for art works for all figures.
References [1] Commission Regulation (EU) No. 37/2010 of 22 December 2009. [2] J. Barbosa, S. Moura, R. Barbosa, F. Ramos, M.I.N. Silveira, Anal. Chim. Acta 586 (2007) 359. [3] JECFA, Toxicological Evaluation of Certain Veterinary Drug Residues in Food: The Joint FAO/WHO Expert Committee on Food Additives, vol. 31, WHO Food Additives Series, 1993, pp. 85. [4] L.F. Capitan-Vallvey, A. Ariza, R. Checa, N. Navas, J. Chromatogr. A 978 (2002) 243. [5] C. Mahugo-Santana, Z. Sosa-Ferrera, M.E. Torres-Padrón, J.J. Santana-Rodríguez, Anal. Chim. Acta 665 (2010) 113. [6] EMEA, Status of MRL Procedures. MRL Assessments in the Context of Council Regulation (EEC) No. 2377/90, European Medicines
38
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38 Agency. Veterinary Medicines and Inspections, 2009, Available: http://www.emea.europa.eu/pdfs/vet/mrls/076599en.pdf (internet). G. Biancotto, L. Contiero, C. Benetti, M. Calligaris, E. Tibaldi, L. Cerni, M. Francese, Anal. Chim. Acta 637 (2009) 173. K. Li, L. Liu, C. Xu, X. Chu, Anal. Sci. 23 (2007) 1281. Commission Decision 2002/251/EC, Off. J. Eur. Commun. L84 (2002) 77. Commission Decision 2002/794/EC, Off. J. Eur. Commun. L276 (2002) 66. European Commission Rapid Alert System for Food and Feed, 2003, Available: http://europa.eu.int/comm/food/fs/sfp/ras index en.html (internet). X.Z. Hu, Y. Xu, A. Yediler, J. Agric. Food Chem. 55 (2007) 1144. ˜ P. Vinas, N. Campillo, L. Carrasco, M. Hernández-Córdoba, Chromatographia 65 (2007) 85. R.J. McCracken, D. Glenn Kennedy, J. Chromatogr. A. 771 (1997) 349. C. Bock, P. Gowik, C. Stachel, J. Chromatogr. B 856 (2007) 178. K.M. Cooper, R.J. McCracken, M.D.G. Buurman, Chem. Anal. Control Expo. Risk Assess. 25 (2008) 548. A. Radovinkovic, M. Moloney, P. Byrne, J. Chromatogr. B 879 (2011) 159. A.S. Pereira, L.C. Pampana, J.L. Donato, G. De Nucci, Anal. Chim. Acta 514 (2004) 9. M.I. Lopez, M.F. Feldlaufer, A.D. Williams, P.S. Chu, J. Agric. Food Chem. 55 (2007) 1103. L.F. Capitan-Vallvey, A. Ariza, R. Checa, N. Navas, Chromatographia 65 (2007) 283. R. Mohamed, P. Mottier, L. Treguier, J. Richoz-Payot, E. Yilmaz, J.C. Tabet, P.A. Guy, J. Agric. Food Chem. 56 (2008) 3500. M. Cronly, P. Behan, B. Foley, E. Malone, L. Regan, J. Chromatogr. A 1216 (2009) 8101. X. Xia, X. Li, S. Ding, S. Zhang, H. Jiang, J. Li, J. Shen, Anal. Chim. Acta 637 (2009) 79.
[24] K. Mitrowska, A. Posyniak, J. Zmudzki, Talanta 81 (2010) 1273. [25] R. Zeleny, S. Harbeck, H. Schimmel, J. Chromatogr. A 1216 (2009) 249. [26] X. Xia, X. Li, S. Zhang, S. Ding, H. Jiang, J. Li, J. Shen, J. Chromatogr. A 1208 (2008) 101. [27] H.W. Sun, F.C. Wang, L.F. Ai, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 857 (2007) 296. [28] M. Cronly, P. Behan, B. Foley, E. Malone, L. Regan, J Chromatogr. B 877 (2009) 1494. [29] S. Teixeira, C. Delerue-Matos, A. Alves, L. Santos, J. Sep. Sci. 31 (2008) 2924. [30] M. Rejtharova, L. Rejthar, J. Chromatogr. A 1216 (2009) 8246. [31] A. Aresta, D. Bianchi, C.D. Calvano, C.G. Zambonin, J. Pharm. Biomed. Anal. 53 (2010) 440. [32] H. Tian, Chemosphere 83 (2011) 349. [33] X.B. Chen, Y.L. Wu, T. Yang, J. Chromatogr. B 879 (2011) 799. [34] L. Rodziewicz, I. Zawadzka, Talanta 75 (2008) 846. [35] A. Tyagi, P. Vernekar, I. Karunasagar, I. Karunasagar, Food Addit. Contam. 25 (2008) 432. [36] S. Zhang, Z. Liu, X. Guo, L. Cheng, Z. Wang, J. Shen, J. Chromatogr. B 875 (2008) 399. [37] J. Han, Y. Wang, C. Yu, C. Li, Y. Yan, Y. Liu, L. Wang, Anal. Chim. Acta 685 (2011) 138. [38] K.M. Cooper, R.J. McCraken, D.G. Kennedy, Analyst 130 (2005) 824. [39] R.R. McCracken, D.G. Kennedy, Food Addit. Contam. 24 (2007) 26. [40] Commission Decision (EEC) No.657, Off. J. Eur. Commun. (2002) 8. [41] AOAC Requirements for Single Laboratory Validation of Chemical Methods, The Association of Official Agricultural Chemists, 2002, Available: http://www.aoac.org/Ag Materials/additives/aoac slv.pdf (internet). [42] A. Masia, M. Ibanez, C. Blasco, J.V. Sancho, Y. Pico, F. Hernandez, Anal. Chim. Acta 761 (2013) 117.
Contents lists available at ScienceDirect
Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb
Liquid chromatography and ion trap mass spectrometry for simultaneous and multiclass analysis of antimicrobial residues in feed water Chusak Ardsoongnearn a , Ongart Boonbanlu a , Sunan Kittijaruwattana a , Leena Suntornsuk b,c,∗ a
Bureau of Quality Control of Livestock Products, Department of Livestock Development, Bangkok 10400, Thailand Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Mahidol University, Bangkok 10400, Thailand c Center of Excellence for Innovation in Drug Design and Discovery, Faculty of Pharmacy, Mahidol University, Bangkok 10400, Thailand b
a r t i c l e
i n f o
Article history: Received 17 July 2013 Accepted 17 November 2013 Available online 25 November 2013 Keywords: Nitrofurans Nitroimidazoles Chloramphenicol HPLC-MS Feed water
a b s t r a c t This work firstly reported the development of liquid chromatography coupled to an ion trap mass spectrometer (LC–MS ion trap) for the simultaneous determination of nitrofurans (e.g. nitrofurazone (NFZ), nitrofurantoin (NFT), furazolidone (FZD) and furaltadone (FTD)), nitroimidazoles (e.g. metronidazole (MNZ), ronidazole (RNZ) and dimetridazole (DMZ)) and chloramphenicol (CAP) in feed water. Isotope-labeled internal standards for the corresponding target analytes were employed to prevent matrix effects that might lead to signal suppression/enhancement. High performance liquid chromatography (HPLC) analysis was performed on a Prodigy ODS-3 column, 2.0 mm × 150 mm, 5 m with a guard cartridge at a flow rate of 0.2 mL/min, column oven temperature of 40 ◦ C, and an injection volume of 10 L. Solid phase extraction (SPE) procedures, factors affecting HPLC separation (e.g. buffer pH and concentrations) and mass spectrometry (MS) parameters were optimized. After an off-line SPE by the OASIS HLB cartridges (with an enrichment factor of 400), the eight antimicrobial agents were separated in 18 min using a gradient elution of acetonitrile in acidified water (pH 5.0). MS detection was by an ion trap MS coupled with electrospray ionization (ESI) in tandem mass spectrometry mode (MS/MS) using the nebulizer gas at 35 psi, drying gas at 9 L/min and drying temperature of 325 ◦ C. Method linearity was good (r2 = 0.979–0.999) with acceptable precision (% RSDs = 3.4–26.6%) and accuracy (%recovery = 88.4–110.1%). Very low limits of detection (LOD) and quantitation (LOQ) were achieved in ranges of 0.002–0.06 g/L and 0.005–0.25 g/L, respectively. The established method is successfully employed by the Department of Livestock Development of Thailand for the monitoring of the drug residues in feed waterbecause of its convenience, reliability and high sensitivity. © 2013 Elsevier B.V. All rights reserved.
1. Introduction Presently, food safety is an issue of public concern, since many prohibited substances are still detected in food products. According to Commission Regulation (EU) No. 37/2010, these substances include Aristolochia spp. and preparation thereof, chloroform, chlorpromazine, colchicines, dapsone and antimicrobials (e.g. chloramphenicol, dimetridazole, metronidazole, nitrofurans (including furazolidone), and ronidazole) [1]. Aristolochia spp. and
∗ Corresponding author at: Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Mahidol University, Bangkok 10400, Thailand. Tel.: +66 2 644 8695; fax: +66 2 644 8695. E-mail address: [email protected] (L. Suntornsuk). 1570-0232/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jchromb.2013.11.034
preparation thereof, chloroform, chlorpromazine, colchicines, dapsone rarely have any uses in livestock, whereas the antimicrobials were previously used for treatments of infection in livestock. Nitrofurans (e.g. furazolidone (FZD), furaltadone (FTD), nitrofurantoin (NFT) and nitrofurazone (NFZ) are antibiotics that were previously used in livestocks as growth promoters and for the prevention and treatment of gastrointestinal infections caused by Escherichia coli, Salmonella spp., Mycoplasma spp., Coccidia spp., coliforms and some other protozoa [2]. Nitroimidazoles are veterinary drugsthat were used for curing and prevention of certainbacterial and protozoal diseases in poultry as well as for swine dysentery, thus they are classified as coccidiostats. The most commonly used nitroimidazoles are metronidazole (MNZ), dimetridazole (DMZ), and ronidazole (RNZ). Both nitrofurans and nitroimidazoles are mutagenic and carcinogenic [3–5]. For this reason, the EU has prohibited their uses
32
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
in feed for food-producing species as described in Annex IV of the Council Regulation (EC) 2377/90 [6]. Chloramphenicol (CAP) is considered as a potent antibiotic for treating pneumonic and enteric disorders. CAP has been recommended for prevention of secondary infections associated with chronic respiratory diseases in poultry. In certain susceptible individuals, CAP is associated with serious toxic effects in human such as bone marrow depression, and fatal a plastic anemia [7,8]. Since this condition is dose independent, CAP has been banned for uses in food producing animals in many countries including EU and Thailand. The incidents of nitrofuran contamination in poultry and seafood exported from Thailand [9] and Brazil [10] to EU during 2002 have prompted more than 150 rapid alerts issued by the EU [11]. Since then, the Thai government has implemented several restricting measures from farm to slaughter houses to ensure the zero tolerance level of these residues. According to Commission Regulation (EU) No. 37/2010, the maximum residue levels (MRL) of these residues cannot be established [1]. Among these measures, regular monitoring of the prohibited substances in feed, feed water, seafood, poultry and meat products has been enforced for registered farms throughout the country. Many analytical methods have been used for monitoring of nitrofurans, nitroimidazoles and CAP residues in various matrices. For example, high performance liquid chromatography (HPLC) and ultra HPLC coupled with photodiode array and mass spectrometric (MS) or tandem MS detectors were employed for determination of nitrofurans and their metabolites in animal feed [2,12–14], poultry muscle and shrimp [15], eyes of broiler chickens [16], animal plasma samples [17], salt [18] and honey [19]. LC-UV, MS or tandem MS of nitroimidazoles and their metabolites in animal feed [20], egg and egg-based samples [21,22], swine kidney [23], poultry muscle [24], pork [25,26], meat [27], animal plasma [28] and waste water [29] were also reported. Furthermore, the techniques were valuable for analysis of CAP in urine [29–31], feed [30], milk [30,32–34], shrimp [35], chicken [36] and environment and food [31,37]. This project was the collaboration between Mahidol University and Veterinary Drugs Assay Division (VDA), Department of Livestock Development, Thailand. The VDA division focuses on quality control of veterinary drugs and monitoring the misuses/abuse of such drugs. In the past, problems of veterinary drug abuses (e.g. the addition of some antimicrobials or growth promoters into feed or feed water) were found in many farms. Previously, HPLC-DAD was employed as a screening method in the monitoring programs of nitrofurans, nitroimidazoles and chloramphenicol residue in feed water. However, there is an urgent need for a confirmatory method with higher sensitivity for simultaneous analysis these residues. This work aimed to develop a LC–MS method for quantitation of nitrofurans (e.g. NFZ, NFT, FZD and FTD), nitroimidazoles (e.g. MNZ, RNZ and DMZ) and CAP (Fig. 1) in feed water since, to the best of our knowledge, there is no report on the multi-residue analysis of these drugs in this matrix. Optimization of solid phase extraction (SPE) procedures, HPLC conditions and MS parameters were performed. The method was validated and applied to analyze the drug residues in forty feed water samples collected from animal farms in Thailand. The method can be routinely applied to quality control of these banned substances in feed water by the Thai officials.
2. Materials and methods 2.1. Reagent Standards of nitrofurans (e.g. FZD, FTD, NFT and NFZ) nitroimidazoles (e.g. MNZ, DMZ and RNZ) and CAP were obtained from Sigma (St. Louis, USA). Isotope-labeled standards (NFZ-13 C-15 N2 ,
NFT-13 C3 , FZD-D4 , FTD-D5 , MNZ-13 C2 -15 N2 , RNZ-D3 and DMZ-D3 ) were used as internal standards (IS) and were from Witega (Berlin, Germany), except CAP-D5 was from Dr. Ehrenstorfer (Augsburg, Germany). Glacial acetic acid and formic acid were AR grade from Merck (Darmstadt, Germany), ammonium acetate, ammonium formate and acetonitrile AR grade were purchased from BDH (Leicestershire, United Kingdom). High purity water was obtained by a MilliQ system from Millipore (Bedford, USA). Other solvents were of HPLC grade from Labs can Asia (Bangkok, Thailand). Adjustment of mobile phase pH was done by adding 0.1% formic acid or acetic acid. Acidified water was prepared by addition of dilute acetic acid to water to obtain pH 5.0. 2.2. Standard and sample preparation Standards and isotope-labeled internal standards were accurately weighed and diluted with acetonitrile to concentrations of 0.5 and 0.2 mg/mL, respectively. The stock solutions were protected from light and kept in a freezer (−20 ◦ C). Intermediate solutions were subsequently diluted with acetonitrile to a concentration of 25 g/mL and kept at 4 ◦ C. Working solutions were prepared daily by diluting the stock standard solutions with acetonitrile to appropriate concentrations for linearity and range experiments. The stocks of standard and internal standard solutions were stable at 4 ◦ C for one month. Forty feed water samples from chicken and pig farms, collected from various provinces of Thailand, were sampled by regional Department of Livestock Development officers. The samples were analyzed without pH adjustment since their pH was between 6.0 and 8.0, which was suitable for SPE using Oasis HLB cartridges. Sample blank solutions were pooled feed water sample (20 L) that was previously analyzed by HPLC-DAD to prove they were antimicrobial-free. Then, they were used for method validation experiments and for preparing blind samples. Blind samples were prepared by spiking sample blank solutions with different concentrations of standards and fixed amounts of the corresponding internal standards. Prior to SPE, two hundred milliliters of the water were transferred into a volumetric flask, 1 mL of 2 M sodium thiosulphate was added into the flask, mixed and left for 30 min before being spiked with 0.5 mL and 0.2 mL of mixtures of the eight IS and known amount of working standard (control samples). Samples were mixed thoroughly and stood for 10 min before loading onto SPE cartridges. Each sample was analyzed by HPLC-DAD prior by HPLC-MS (n = 2). If HPLC-DAD showed any positive results, the samples would be confirmed by HPLC-MS in six replicates. One round of each batch analysis consisted of standard solutions at eight different concentrations, 20 feed water samples and six blind samples. The spiking procedure for the blind samples was randomly predetermined using the software available at www.random.org. 2.3. SPE procedure Off-line SPE using Oasis HLB cartridges (200 mg/6 mL) from waters (Milford, USA) was employed for sample pre-concentration. The cartridges were preconditioned with 5 mL of methanol and 5 mL of water. Two hundred milliliters of feed water sample were loaded onto the cartridges at a rate of 4 mL/min using a vacuum manifold. The cartridges were washed with 5 mL of 5% methanol and dried under vacuum to remove water residues and eluted with 4 mL of eluent (2 mL × 2). Eight different eluents consisting of methanol, acetonitrile, and their mixture with different amounts of acetic acid (0.5–2%) were optimized for the appropriate elution of all eight antimicrobial drugs. The eluate was dried using a rotary evaporator, reconstituted with 0.5 mL of 20 mM ammonium acetate
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
33
Fig. 1. Structures of the investigated compounds.
(pH 4.5): methanol (80:20, v/v), vortexed for 30 sand sonicated for 15 s. Then, it was filtered through a 0.45-m syringe nylon filters prior analysis. 2.4. Liquid chromatography-ion trap mass spectrometry HPLC analysis was performed on a HP1200 series from Agilent Technologies (Santa Clara, USA), equipped with an automatic sampler/injector, a degasser, a binary pump and a column oven. The LC instrument was coupled with a Bruker Daltonicion trap mass spectrometer, equipped with an electrospray (ESI) interface (Bremen, Germany). The LC–MS system was controlled by HP workstation XW 6200 using Bruker Daltonic Esquire 6.1 software comprised of Esquire Control version 6.1, Data Analysis version 3.4, HyStar version 3.2 and Quant Analysis version 1.8 for data analysis. Chromatographic separation was in a reversed phase mode using a Prodigy ODS-3 (15 cm × 0.20 cm with a 5 m particle size) column from Phenomenex (Torrance, USA) coupled with a guard cartridge. The column temperature was set at 40 ◦ C and the injection volume was 10 L. The flow rate of the binary mobile phase was 0.2 mL/minusing either isocratic or gradient profiles. Various mobile phases, including solvent A (aqueous buffer of ammonium formate and formic acid, ammonium acetate and acetic acid at various concentrations and pH or acetonitrile in acidified water) and
solvent B (acetonitrile, methanol, acetonitrile in acidified water), were optimized to obtain reasonable resolution, peak shape and run time. Ionization for MS of MNZ, RNZ, DMZ, FTD and FZD was carried out in a positive mode, while that of NFZ, NFT and CAP was in negative mode [2,38,39]. Precursor ions, product ions from MS/MS reactions and fragmentation amplitudes were studied by direct infusion of 1 mg/Lof each standard solution into an electrospray ionization source at a flow rate of 4 L/min, using ultra-high pure nitrogen as nebulizer and drying gas at 35 psi with a flow of9 L/min, and a drying temperature of 325 ◦ C. Various parameters such as capillary voltage, ion optics (e.g. skimmer, octapole, lens, trap drive voltage), and fragmentation voltage were varied to obtain the highest intensity of precursor and product ions of each compound. For peak identification, one precursor and two product ions must be detected with at least 4 identification points according to Commission Decision 2002/657/EC [40]. The four points are required for Group A substances, which include the current investigated drugs (e.g. MNZ, RNZ, DMZ, NFZ, NFT, FZD, FTD and CAP). For mass spectrometric detection, the relative intensities of the detected ions, expressed as a percentage of the intensity of the most intense ion, shall correspond to those of the calibration standard, at comparable concentration within the reported tolerances.
34
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
2.5. Method validation Validation protocol for the analysis of the target analytes in animal feed water is not yet available in EU guideline. Thus, a validation protocol was established to prove that the method performance was fit for the purpose. 2.5.1. Linearity For the calibration procedure, the internal standard calibration method was applied to this study to overcome matrix effects. These effects are usually unknown or variable and are major limitations for trace analysis by LC-ESI-MS. NFZ-13 C-15 N2 , NFT-13 C3 , FZD-D4 , FTD-D5 , MNZ-13 C2 -15 N2 , RNZ-D3 , DMZ-D3 and CAP-D5 were used as internal standards for NFZ, NFT, FZD, FTD, MNZ, RNZ, DMZ and CAP, respectively. Linearity of method was performed from blank samples spiked with eight different concentrations of the standards and fixed amounts of the internal standards. The spiked solutions were subjected to SPE, in the same manner as feed water samples (Section 2.3), prior to HPLC-MS analysis. Linearity of system was evaluated by spiking eight different concentrations of the standards and fixed amounts of the internal standards into 4.0 mL of eluent 6 (1% acetic acid:methanol, 80:20), which was the optimal eluent in SPE. The solutions were dried and reconstituted as previously described (Section 2.3). Linearity of method and system was examined by analyzing standard solutions in triplicate at eight concentrations on three different days. Calibration curves were plots of peak area ratios against concentration ratios of the target compounds and IS. Slopes, intercepts and correlation of determinations (r2 ) were calculated using Excel® . 2.5.2. Precision and accuracy Method precision and accuracy was evaluated from intra-day and inter-day measurements by analyzing the spiked sample blank solutions (n = 10) on the same and three different days. The sample blank solutions were spiked with three concentrations of the standards (low, mid and high points of the calibration curve). Precision was reported as % relative standard deviations (% RSDs) and accuracy was represented as % recovery (% R) of the assays. 2.5.3. LOD and LOQ Method sensitivity was evaluated by limits of detection (LOD) and quantitation (LOQ). LOD is defined as the lowest concentration at which an analyte can be differentiated from the background noise, whereas LOQ is the lowest concentration at which an analyte can be accurately quantified. In this study, LOD and LOQ were based on the variability of the blank [41]. Ten blank samples were spiked with the lowest acceptable concentration (0.013 g/L for MNZ, 0.005 g/L for RNZ and CAP, 0.015 g/L for DMZ, 0.035 g/L for NFZ, 0.025 g/L for NFT and 0.001 g/L for FZD and FTD) and analyzed by HPLC-MS. Based on the standard curve, the calculated LOD and LOQ were expressed as the analyte concentrations corresponding to peak area ratios of the mean value +3SD and +10SD, respectively. Then, standards at calculated LODs and LOQs were spiked into the blank samples. Signal to noise ratios (S/N) of peak height ratios were monitored and LOD and LOQ were adjusted till the S/N of not less than 3 and 10, respectively, were obtained. 2.5.4. Robustness Method robustness is an evaluation of the method capacity to remain unaffected by small, but deliberate changes in analytical conditions. Currently, pH of the mobile phase greatly influenced the analyte retention. Thus, robustness testing was carried out by varying the mobile phase pH ± 0.5 around the optimal value and % RSDs of retention times were determined.
Fig. 2. Recovery of the analytes from different eluents in SPE.
3. Results and discussion 3.1. SPE optimization Off-line SPE was required in this study to achieve the desirable sensitivity. Oasis HLB cartridges were chosen for water sample pretreatment and pre-concentration (enrichment factor of 400) since the cartridges was efficient for analytes with different polarity. Eluent 1 (methanol) provided good recovery for MNZ, RNZ, DMZ and NFT (% R between 81.5 and 105.4%), but gave low recovery for NFZ and CAP (% R ∼ 19.4%) and could not elute FZD and FTD. Interestingly, eluent 2 (acetonitrile) still gave acceptable recovery for MNZ, RNZ, DMZ and NFT, and significantly improved the recovery of NFZ, FZD, FTD and CAP to 61.5, 106.5, 1002 and 53.2%, respectively. Subsequent work focused on increasing the recovery for NFZ and CAP. A mixture of methanol and acetonitrile at 50:50 and 80:20 (eluents 3 and 4, respectively) did not enhance the recovery of NFZ and CAP, thus addition of acetic acid at 0.5 (eluent 5), 1.0 (eluent 6), 1.5 (eluent 7) and 2.0% (eluent 8) to the mixture of methanol and acetonitrile (80:20, eluent 4) was investigated as eluents. Among eluents 5 to 8, the highest recovery for CAP (∼67.5%) was observed in eluent 6. In addition, eluent 5 gave the lowest recovery for MNZ (57.0%). Thus, the eluent 6 was selected and the recovery for all analytes in this eluent ranged from 60.1 to 106.0% (Fig. 2). The eluate was dried and reconstituted with 20 mM ammonium acetate (pH 4.5): methanol (80:20, v/v) since it provided higher sensitivity than that from buffer of pH 5.0.
3.2. HPLC optimization Initial experiments were performed by isocratic elution using methanol or acetonitrile together with aqueous ammonium formate or ammonium acetate (5–40 mM, pH between 3.9 and 4.8). Results revealed that the resolving power of isocratic elution was not sufficient to separate all of the analytes, especially MNZ, RNZ and DMZ, which were overlapping under most mobile phase compositions. Moreover, run time in isocratic elution was unsatisfactory (up to 50 min per injection). Thus, gradient elution of the investigated substances using ammonium formate or ammonium acetate (10 and 20 mM, pH 3.5 and 5.0) and acetonitrile was optimized. Improvement of resolution and reduced retention time were observed in buffer at low concentrations. Ammonium acetate buffer (pH 5.0) was preferred since reasonable separation between DMZ, FTD and NFZ could be obtained. Separation of these analytes in ammonium formate buffer (pH 3.6) was more difficult due to the narrow separating window FTD and the retention time shift during long batch run.
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
35
Table 1 MS/MS reactions monitored by LC-ESI MSn . Analyte Positive mode [M+H]+ MNZ MNZ-13 C2 -15 N2 RNZ RNZ-D3 DMZ DMZ-D3 FZD FZD-D4 FTD FTD-D5 Negative mode [M−H]− NFZ NFZ-13 C-15 N2 NFT NFT-13 C3 CAP
Retention time (min)
Exact mass
MS/MS reactions (m/z)
5.89
171.06
6.74
200.05
9.32
141.05
13.69
225.04
16.43
324.11
172.0 > 128.0 176.0 > 132.0 201.0 > 140.0 204.2 > 143.0 142.0 > 96.2 145.2 > 99.2 226.0 > 113.1 229.8 > 117.1 325.2 > 281.1 330.2 > 286.2
10.96
198.04
12.11
238.03
18.63
322.01
Finally, the gradient elution of buffer-free mobile phase (solvent A: 10% ACN, solvent B: 30% ACN in acidified water, pH 5.0) was employed. The gradient was as follows: 0 min (0% B), 12 min (100% B) and 15 min (100% B), then the mobile phase was returned to the initial condition in 1 min and equilibrated for another 6 min. The mobile phase offered the simple and convenient column equilibration and clean up. Reasonable run time of 18.7 min with a resolution of >3.0 could be achieved for the target analytes (Fig. 3a).
196.8 > 149.9 199.8 > 152.9 236.8 > 151.9 239.8 > 151.9 321.0 > 257.0
Capillary (volt)
−3950 −3950 −3950 −4183 −4242
3775 3075 3717
Fragmentation amplitude (volt) 0.75 0.67 0.73 0.70 0.66 0.74 0.63 0.63 0.67 0.67 0.80 0.77 0.84 0.73 1.21
3.3. MS optimization Electrospray was used as the ionization source. Selection of an ionization mode for individual analyte was based on the MS optimization results and previous reports [2,38,39]. Tuning of MS parameters by direct infusion of the target compounds and IS was performed using the optimization function in Esquire Control version 6.1. Fine tune MS parameters were listed in Table 1. These parameters were specified for each time segment and were used for all experiments. Chromatograms with MS detection of each compound were accomplished with the previously optimized HPLC conditions (Fig. 3b). 3.4. Validation of the LC–MS method Method validation experiments were performed using real feed water samples that were previously analyzed to prove they were blank samples (antimicrobial-free samples). The matrixmatch standards (spiked blank samples) and internal standards were employed to minimize matrix effects, hence improvement of method precision and accuracy.
Fig. 3. (a) HPLC-MS chromatograms of the investigated compounds (1 mg/L) and (b) time segments for measuring specific transitions of analytes are indicated in the chromatogram. HPLC condition: prodigy ODS (3), 2.0 × 150 mm, 5 m with guard cartridge, gradient elution of A: B (10%: 30% acetonitrile in acidified water (pH 5.0), 0 min B 0%, 12 min B 100%, 15 min B 100%, 16 min B 0%, 22 min B 0%, 40 ◦ C, flow rate: 0.2 mL/min. MS condition; see Section 2.4 and Table 1.
3.4.1. Linearity The proposed method showed good linearity of method (matrixmatched standard) and system (matrix-free standard) in the investigated ranges with r2 greater than 0.993 for all target analytes (Table 2). Percent differences of slopes between matrix-matched standards and matrix-free calibration curves of greater than 10% indicated the presence of matrix effects that normally occur in electrospay ionization MS. Matrix effects can also affect analysis of residues in trace amounts, such as in environmental samples, which results in unsatisfactory accuracy and precision. In the present study, % difference between matrix-matched standards and matrixfree standard calibration curves were within 10% for most analytes. Additionally, matrix effects could be pronounced since off-line SPE was performed for sample enrichment. To investigate this, ratios of slopes between the matrix-matched and matrix-free standards were calculated and degrees of signal suppression or enhancement could be identified. For example, the slope ratios between 0.5 and 0.8 indicated moderate signal suppression, between 0.8 and 1.2 corresponded to signal suppression or enhancement of less than 20% and from 1.2 to 1.5 represented moderate signal enhancement [42]. Our results revealed the ratios from 1.0 to 1.2 (Table 2), which indicated that no signal suppression was observed and moderate signal enhancement was found only in FTD. The sample matrices did not show any strong signal suppression or enhancement.
36
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
Table 2 Linearity, LOD and LOQ dataa . Range (g/L) 0.1–0.7 0.03–0.21 0.2–1.4 0.25–1.75 0.2–1.4 0.005–0.035 0.01–0.07 0.025–0.176
MNZ RNZ DMZ NFZ NFT FZD FTD CAP a
Method linearity
System linearity
RS
LOD (g/L)
LOQ (g/L)
y = 0.779x + 0.042 (r2 = 0.995) y = 0.993x + 0.013 (r2 = 0.997) y = 0.778x + 0.063 (r2 = 0.997) y = 0.347x − 0.028 (r2 = 0.996) y = 0.913x − 0.064 (r2 = 0.995) y = 0.975x + 0.035 (r2 = 0.996) y = 1.148x + 0.001 (r2 = 0.995) y = 1.112x − 0.013 (r2 = 0.996)
y = 0.729x + 0.045 (r2 = 0.993) y = 0.936x + 0.008 (r2 = 0.997) y = 0.775x + 0.027 (r2 = 0.996) y = 0.338x − 0.020 (r2 = 0.998) y = 0.809x − 0.031 (r2 = 0.999) y = 0.937x + 0.036 (r2 = 0.997) y = 0.939x − 0.016 (r2 = 0.996) y = 1.017x + 0.005 (r2 = 0.997)
1.07 1.06 1.00 1.03 1.13 1.04 1.22 1.09
0.03 0.01 0.05 0.06 0.04 0.002 0.002 0.01
0.10 0.03 0.20 0.25 0.20 0.005 0.01 0.025
RS is ratios of slope from the matrix-matched and matrix-free standard, LOD and LOQ are limits of detection and quantitation, respectively.
Table 3 Precision dataa . MNZ Conc. (g/L) Intra-day Inter-day
0.1 13.21 14.14
RNZ 0.4 9.52 10.00
0.7 8.35 9.40
NFT Conc. (g/L) Intra-day Inter-day a
0.2 6.26 7.23
DMZ
0.03 13.79 14.31
0.12 8.53 8.76
FZD 0.8 7.91 9.14
1.4 7.32 7.54
0.005 10.00 10.71
0.2 13.10 13.94
0.8 9.01 9.20
1.4 9.77 9.93
FTD 0.02 7.33 7.44
0.035 7.47 7.40
0.01 18.25 19.61
0.25 9.33 9.39
1.0 10.18 10.11
1.75 9.88 10.44
CAP 0.04 12.78 14.44
0.07 12.90 14.15
0.025 11.70 11.60
0.1 10.41 10.60
0.175 10.11 10.26
Data is represented as percent relative standard deviation of the assay.
3.4.2. Precision and accuracy Method precision was represented by % RSDs from intra- and inter-day analysis. The uses of suitable isotope-labeled internal standards offered acceptable precision for all target analytes with % RSDs between 6.26 and 18.25% for intra-day precision and between 7.23 and 19.61% for inter-day precision (Table 3). Generally, % RSDs from intra-day precision were slightly lower than those from interday precision. For most analytes, higher % RSDs were obtained at low concentrations (the lowest points of calibration curves), which were close to LOQs of the analytes. High % RSDs were obtained for both intra- and inter-day assay of FTD (between 12.78 and 19.61) due to the moderate signal enhancement from matrix effects, whereas % RSDs from NFT were in lower ranges (6.26–9.14). Accuracy of the method was represented by recovery of the spiked blank samples from the same and different day analysis. Recoveries were higher than 88% with % RSDs of less than 26.6% for all target analytes. The intra-day and inter-day recoveries were between 88.4 and 110.1% and between 94.1 and 103.2%, respectively (Table 4). The lowest recovery with highest %RSD found in FTD might stem from overlapping of recovery with matrix effects [42]. However, these values were within acceptable ranges (50–120%) recommended by AOAC [41]. 3.4.3. LOD and LOQ LODs and LOQs were estimated from feed water samples spiked at very low concentrations, which provided S/N of 3 and 10, Table 4 Accuracy dataa .
%R MNZ RNZ DMZ NFZ NFT FZD FTD CAP
90.1–103.6 91.3–103.7 96.0–106.8 94.8–110.1 97.6–106.3 93.5–104.2 88.4–104.1 96.1–105.6
respectively, based on visualization of peak height. LODs of all target analytes were from 0.002 to 0.06 g/L (Table 2). The lowest LODs were measured in FZD and FTD (0.002 g/L) and the highest in NFZ (0.06 g/L). Comparing to the reported LODs, which varied from 0.005 to 1.6 g/L [4,11,26,28,35], the proposed technique could provide lower LODs indicating the high sensitivity of the method. LOQs were found between 0.005 and 0.25 g/L, which were the lowest concentration levels for each analyte in recovery studies. The lowest LOQ corresponded to FZD (0.005 g/L) and the highest to NFZ (0.25 g/L). Fig. 4 represents the HPLC–MS chromatograms of the feed water sample and the sample spiked with the standards at LOQ concentrations (Table 2). Although no MRL can be established for the investigated analytes, the LOD and LOQ were considered to be in very low levels.
3.4.4. Robustness Modification of mobile phase pH could affect retention of the analytes and their resolution. Results showed that the method was robust upon the alteration of mobile phase pH (4.5, 5.0 and 5.5), which gave the % RSDs of the retention time within 1.25% (Table 5). Validation data confirms that the method is precise, accurate and sensitive for trace analysis of NFZ, NFT, FZD, FTD, MNZ, RNZ, DMZ and CAP in feed water.
Table 5 Robustness data*.
Intra-day
a
0.21 9.66 9.54
NFZ
Inter-day % RSD 5.3–16.1 7.6–16.0 4.9–17.2 6.3–14.3 3.4–11.3 5.5–11.6 9.2–26.6 8.4–15.1
%R 94.4–99.8 95.1–100.8 100.1–102.7 99.7–107.0 100.6–102.2 96.4–101.3 94.1–95.5 99.4–103.2
Analyte % RSD 9.4–14.1 8.8–14.3 9.2–13.9 9.4–10.4 7.2–9.1 7.4–10.7 14.1–19.6 10.3–11.6
% R is percent recovery, % RSD is percent relative standard deviation.
MNZ RNZ DMZ NFZ NFT FZD FTD CAP
Retention time from various pH of water in mobile phase (min) pH 4.5
pH 5.0
pH 5.5
% RSD
6.1 7.0 9.5 11.1 12.3 13.8 16.8 18.9
6.1 6.9 9.4 11.1 12.2 13.7 16.5 18.8
6.1 7.0 9.4 11.1 12.0 13.8 16.4 19.1
0.32 1.00 0.20 0.35 1.25 0.42 1.20 0.87
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38
37
Fig. 4. HPLC-MS chromatograms of (a) the feed water sample and (b) the sample spiked with the standards at LOQ levels. HPLC and MS condition: see Fig. 3.
3.4.5. Application Forty feed water samples from different farms were tested for the drug residues. In each batch analysis, six blind samples were simultaneously analyzed as quality control samples. Recoveries of the blind sample batches were between 68.1 and 117.6%. The low recovery of RNZ (68.1%) in the batch analysis might due to the low spiking concentration that was close to the LOQ of RNZ (0.03 g/L) and the high %RSD (about 14%). However, this value was acceptable according to AOAC Requirements for Single Laboratory Validation of Chemical Methods, which allows the recovery of 50–120% for a concentration of ≤1 ppb [41]. This low recovery had no negative effects on the quantitation of the antimicrobials in real samples since other validation parameters were within acceptable ranges. In the two batch analysis, no residues of NFZ, NFT, FZD, FTD, MNZ, RNZ, DMZ and CAP were found in the forty feed water samples. This finding may reflect the efficiency of residue control measures in livestock production by the Department of Livestock Development of Thailand. These preventive measures have been strictly applied since 2002 to assure the food safety concept (from Farm to Table). 4. Conclusions A HPLC-MS method has been established for residue analysis of eight multiclass antimicrobial drugs, including nitrofurans, nitroimidazoles and chloramphenicol, at sub-ppb levels in feed water. The simple gradient elution of acetonitrile in acidified water in 22 min was efficient for separation of the drugs. The SPE procedure using 1% acetic acid in a mixture of acetonitrile and methanol (80:20) as the eluent could provide good recoveries of the drug residues from sample matrices with an enrichment of 400. All analytical performance characteristics meet the requirements and the method is suitable for the intended purposes. The SPE scheme and MS detection were valuable for the current work enabling the LOD at trace levels (0.002–0.06 g/L), which are lower than reported
values. The use of matrix-match standards and internal standards enhanced method precision and accuracy, while minimized matrix effects. Novelty of this work includes the high sensitivity for simultaneous and multiclass analysis of drug residues in feed water. In addition, the use of buffer-free mobile phase, resulting in fast and convenient column equilibration and clean up and reduced analysis time and cost. The method can be applied to routine analysis in regulatory departments, where sample throughput and sensitivity are priority.
Acknowledgements This project is supported by the Office of the High Education Commission and Mahidol University under the National Research Universities Initiative. The authors thank the Bureau of Quality Control of Livestock Products, Department of Livestock Development for providing all instruments and most chemicals. Thanks also go to Prof. Wolfgang Buchberger, Institute of Analytical Chemistry, Johannes Kepler University, Linz, Austria, for suggestions on the manuscript. Thanks also go to Mr. Brompoj Prutthiwanasan for art works for all figures.
References [1] Commission Regulation (EU) No. 37/2010 of 22 December 2009. [2] J. Barbosa, S. Moura, R. Barbosa, F. Ramos, M.I.N. Silveira, Anal. Chim. Acta 586 (2007) 359. [3] JECFA, Toxicological Evaluation of Certain Veterinary Drug Residues in Food: The Joint FAO/WHO Expert Committee on Food Additives, vol. 31, WHO Food Additives Series, 1993, pp. 85. [4] L.F. Capitan-Vallvey, A. Ariza, R. Checa, N. Navas, J. Chromatogr. A 978 (2002) 243. [5] C. Mahugo-Santana, Z. Sosa-Ferrera, M.E. Torres-Padrón, J.J. Santana-Rodríguez, Anal. Chim. Acta 665 (2010) 113. [6] EMEA, Status of MRL Procedures. MRL Assessments in the Context of Council Regulation (EEC) No. 2377/90, European Medicines
38
[7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]
C. Ardsoongnearn et al. / J. Chromatogr. B 945–946 (2014) 31–38 Agency. Veterinary Medicines and Inspections, 2009, Available: http://www.emea.europa.eu/pdfs/vet/mrls/076599en.pdf (internet). G. Biancotto, L. Contiero, C. Benetti, M. Calligaris, E. Tibaldi, L. Cerni, M. Francese, Anal. Chim. Acta 637 (2009) 173. K. Li, L. Liu, C. Xu, X. Chu, Anal. Sci. 23 (2007) 1281. Commission Decision 2002/251/EC, Off. J. Eur. Commun. L84 (2002) 77. Commission Decision 2002/794/EC, Off. J. Eur. Commun. L276 (2002) 66. European Commission Rapid Alert System for Food and Feed, 2003, Available: http://europa.eu.int/comm/food/fs/sfp/ras index en.html (internet). X.Z. Hu, Y. Xu, A. Yediler, J. Agric. Food Chem. 55 (2007) 1144. ˜ P. Vinas, N. Campillo, L. Carrasco, M. Hernández-Córdoba, Chromatographia 65 (2007) 85. R.J. McCracken, D. Glenn Kennedy, J. Chromatogr. A. 771 (1997) 349. C. Bock, P. Gowik, C. Stachel, J. Chromatogr. B 856 (2007) 178. K.M. Cooper, R.J. McCracken, M.D.G. Buurman, Chem. Anal. Control Expo. Risk Assess. 25 (2008) 548. A. Radovinkovic, M. Moloney, P. Byrne, J. Chromatogr. B 879 (2011) 159. A.S. Pereira, L.C. Pampana, J.L. Donato, G. De Nucci, Anal. Chim. Acta 514 (2004) 9. M.I. Lopez, M.F. Feldlaufer, A.D. Williams, P.S. Chu, J. Agric. Food Chem. 55 (2007) 1103. L.F. Capitan-Vallvey, A. Ariza, R. Checa, N. Navas, Chromatographia 65 (2007) 283. R. Mohamed, P. Mottier, L. Treguier, J. Richoz-Payot, E. Yilmaz, J.C. Tabet, P.A. Guy, J. Agric. Food Chem. 56 (2008) 3500. M. Cronly, P. Behan, B. Foley, E. Malone, L. Regan, J. Chromatogr. A 1216 (2009) 8101. X. Xia, X. Li, S. Ding, S. Zhang, H. Jiang, J. Li, J. Shen, Anal. Chim. Acta 637 (2009) 79.
[24] K. Mitrowska, A. Posyniak, J. Zmudzki, Talanta 81 (2010) 1273. [25] R. Zeleny, S. Harbeck, H. Schimmel, J. Chromatogr. A 1216 (2009) 249. [26] X. Xia, X. Li, S. Zhang, S. Ding, H. Jiang, J. Li, J. Shen, J. Chromatogr. A 1208 (2008) 101. [27] H.W. Sun, F.C. Wang, L.F. Ai, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 857 (2007) 296. [28] M. Cronly, P. Behan, B. Foley, E. Malone, L. Regan, J Chromatogr. B 877 (2009) 1494. [29] S. Teixeira, C. Delerue-Matos, A. Alves, L. Santos, J. Sep. Sci. 31 (2008) 2924. [30] M. Rejtharova, L. Rejthar, J. Chromatogr. A 1216 (2009) 8246. [31] A. Aresta, D. Bianchi, C.D. Calvano, C.G. Zambonin, J. Pharm. Biomed. Anal. 53 (2010) 440. [32] H. Tian, Chemosphere 83 (2011) 349. [33] X.B. Chen, Y.L. Wu, T. Yang, J. Chromatogr. B 879 (2011) 799. [34] L. Rodziewicz, I. Zawadzka, Talanta 75 (2008) 846. [35] A. Tyagi, P. Vernekar, I. Karunasagar, I. Karunasagar, Food Addit. Contam. 25 (2008) 432. [36] S. Zhang, Z. Liu, X. Guo, L. Cheng, Z. Wang, J. Shen, J. Chromatogr. B 875 (2008) 399. [37] J. Han, Y. Wang, C. Yu, C. Li, Y. Yan, Y. Liu, L. Wang, Anal. Chim. Acta 685 (2011) 138. [38] K.M. Cooper, R.J. McCraken, D.G. Kennedy, Analyst 130 (2005) 824. [39] R.R. McCracken, D.G. Kennedy, Food Addit. Contam. 24 (2007) 26. [40] Commission Decision (EEC) No.657, Off. J. Eur. Commun. (2002) 8. [41] AOAC Requirements for Single Laboratory Validation of Chemical Methods, The Association of Official Agricultural Chemists, 2002, Available: http://www.aoac.org/Ag Materials/additives/aoac slv.pdf (internet). [42] A. Masia, M. Ibanez, C. Blasco, J.V. Sancho, Y. Pico, F. Hernandez, Anal. Chim. Acta 761 (2013) 117.
