Food Anal. Methods DOI 10.1007/s12161-017-0922-2

Rapid Simultaneous Screening and Detection of 12 Anticoagulant Rodenticides in Food by Ultra-performance Liquid Chromatography-Triple Quadrupole/Linear Ion Trap Tandem Mass Spectrometry Xiaoli Cao 1 & Xiaoqian Yang 1 & Zhong Liu 1 & Haitao Jiao 1 & Suhua Liu 1 & Lanzheng Liu 1 & Qingfen Meng 2

Received: 9 December 2016 / Accepted: 24 April 2017 # Springer Science+Business Media New York 2017

Abstract A rapid analytical method for the simultaneous screening and detection of 12 anticoagulant rodenticides in food samples was established based on modified QuEChERS (quick, easy, cheap, effective, rugged, and safe) sample preparation method using an ultra-performance liquid chromatography-triple quadrupole/linear ion trap tandem mass spectrometry (UPLC-QTrap-MS/MS). Food samples were extracted and purified with modified QuEChERS method. The 12 anticoagulant rodenticides (warfarin, brodifacoum, difethialone, coumachlor, coumatetralyl, bromadiolone, chlorophacinone, difenacoum, diphacinone, pindone, valone, and flocoumafen) and the internal standard (warfarin-D5) were separated within 6 min using an ACQUITY UPLC BEH-C18 column (1.7 μm, 2.1 mm × 50 mm) and gradient elution with the mobile phase consisting of 5 mM ammonium acetate formate buffer and acetonitrile. The multiple reaction monitoring-information dependent acquisition-enhanced product ion (MRM-IDA-EPI) scanning was employed for detection. The calibration curves were linear (R2 > 0.99) for all the compounds. The mean recoveries for the 12 analytes at three spiked levels (1× LOQ, 5× LOQ, and 10× LOQ) were in the range of 79.5–113.2% with RSDs of 1.8–13.2%.The limits of detection (LOD) for the 12 rodenticides ranged from 0.01 (warfarin) to 0.05 μg/kg (diphacinone). The limit of

* Lanzheng Liu [email protected]

1

Jinan Municipal Center for Disease Control and Prevention, Jinan 250021, Shandong, People’s Republic of China

2

Qinghai Institute of Salt Lakes, Chinese Academy of Sciences, Xining 810008, People’s Republic of China

quantification (LOQ) for the 12 rodenticides was between 0.02 (warfarin) and 0.10 μg/kg (diphacinone). The developed method was more straightforward, less time and labor intensive, and more sensitive, selective, and accurate for screening multiple anticoagulant rodenticides, and it was successfully used in several poisoning cases. Keywords Anticoagulant rodenticides . Multi-residue analysis . Modified QuEChERS . Food . UPLC-QTrap-MS/ MS

Introduction Anticoagulant rodenticides are widely used as pest control agents in agriculture and urban rodent control (GómezCanela et al., 2014a). However, poisoning of human beings and non-targeted animals also often occurs when the anticoagulant rodenticides are applied to feed stocks (Watt et al. 2005; Spahr et al. 2007; Olmos et al. 2007; Berny et al. 2010; Sánchez-Barbudo et al. 2012). There are two classes of anticoagulant rodenticides, hydroxycoumarins and indandione. These rodenticides specially inhibit the action of vitamin K reductase, leading to reduced biosynthesis of the active forms of clotting factors II, VII, IX, and X (Chua et al. 1998). The usual clinical presentation of anticoagulant rodenticides poisoning is coagulation disorders, hemorrhage, and even death. Vitamin K1 is clinically used as an antidote to treat patients poisoned by anticoagulant rodenticides. The vast majority of rodenticide poisoning was foodborne poisoning and may be attributed to accidental exposures, suicide, or intentional poisoning (Chua et al. 1998; Papin et al. 2007; Watson et al. 2004). Thus, a rapid analytical method is desired to certify

Food Anal. Methods

the suspected poisoning of anticoagulant rodenticides for both domestic animals and human beings in accidental and intentional cases. At present, a number of analytical methods are used to detect anticoagulant rodenticides, including thin-layer chromatography (Rengel et al. 1993), high-performance liquid chromatography (HPLC) with ultraviolet (UV) and fluorescence detector (FLD) (Palazoglu et al. 1998; Kuijpers et al. 1995; Chalermchaikit et al. 1993; Felice et al. 1991; Kieboom et al. 1981; Hunter et al. 1998; O’Bryan et al. 1991; Kelly et al. 1993; Jones,1996; Guan et al., 1999a), high-performance thin-layer chromatography (HPTLC) (Berny et al., 1995), ion chromatography (IC) (Jin et al., 2007a), gas chromatography-mass spectrometry (Duffield et al. 1979), and liquid chromatography-mass spectrometry (LC-MS) (Jin et al. 2006; Vindenes et al. 2008; Grobosch et al. 2006; Jin et al., 2007b; Guan et al., 1999b). But these methods suffered from a lack of sensitivity and selectivity. In recent years, because of excellent specificity, speed, and sensitivity, liquid chromatography coupled with electrospray ionization tandem mass spectrometry based on a triplequadrupole configuration (LC-MS/MS) has played an important role for the analyses of anticoagulant rodenticide determination in complex biological matrices such as tissue (Vandenbroucke et al. 2008; Maršálek et al. 2015), blood (Vandenbroucke et al. 2008; Jin et al. 2009; Dong et al. 2015; Jin et al., 2007c; Jin et al., 2007d; Huang et al. 2008; Jin et al., 2007e; Cai et al., 2009a, b; Yan et al. 2012), human hair (Zhu et al. 2013), soil, and water (Gómez-Canela et al., 2014b; Chen et al. 2014). The multiple reaction monitoring (MRM) of the LC-MS/MS methods can obtain the quantitative ions, qualitative ion, and ion ratio to achieve identification of target analytes, but they cannot obtain more secondary structure information of the compound. Because of the ionization matrix effects, the ion ratio of the qualitative ion and quantitative ion may be deviated. And at this time, the false negative or false positive results can occur. In this study, an ultra-performance liquid chromatographytriple quadrupole/linear ion trap tandem mass spectrometry (QTrap UPLC-MS/MS) method was developed for the determination of 12 anticoagulant rodenticides (warfarin, brodifacoum, difethialone, coumachlor, coumatetralyl, bromadiolone, chlorophacinone, difenacoum, diphacinone, pindone, valone, and flocoumafen) in the food samples. The QTrap UPLC-MS/MS method used multiple reaction monitoring-information dependent acquisition-enhanced product ion (MRM-IDA-EPI) scanning and library search mode. In this method, the MRM information and secondary structure information combined with the library search can be realized qualitatively and quantitatively simultaneously. The developed method was successfully used for the determination and the confirmation of suspected anticoagulant rodenticides in poisoning cases.

Experimental Chemicals and Reagents Acetonitrile, acetone, methanol, and ethyl acetate were purchased from Merck Company (Germany). All solvents were HPLC grade. The ammonium acetate (HPLC grade) was purchased from Fluka (Sigma-Aldrich). Ultra-purified water was obtained from a Milli-Q apparatus (Millipore, Bedford, MA, USA). The sorbents of primary-secondary amine (PSA), graphitized carbon black (GCB), octadecylsilane (C18), and EMR-lipid purge tube were purchased from Agilent (USA). Standard and Working Solutions The anticoagulant rodenticides investigated in this work were purchased from J&K Technology (USA). The purities of the standard pesticides were >99% (w/w). Warfarin-D5 in acetonitrile at a concentration of 100 μg/ml was obtained from J&K Technology (USA). The internal standard working solution (1 μg/ml) was prepared by dilution of the warfarin-D5 (100 μg/ml) in acetonitrile. Standard stock solutions (100 μg/ml) of individual anticoagulant rodenticides were prepared in acetonitrile, and the stock solutions stored at −20 °C were stable for at least 6 months. Mixed standard working solutions (1 μg/ml) were prepared by dilution of stock solutions in acetonitrile. The mixed standard working solutions should be stored in dark below 4 °C and can be used for 1 month. Chromatography and Mass Spectrometry Chromatographic separation was performed using an Acquity UPLC system (Waters, USA) with a BEH-C18 column (1.7 μm, 2.1 mm × 50 mm, Waters) equipped with a guard column at 30 °C. The mobile phase consisted of 5 mM ammonium acetate (A) and acetonitrile (B). The gradient elution was as follows: 10% B at 0–0.5 min, 10–95% B at 0.5– 3.0 min, 95% B at 3.0–4.0 min, 95–10% B at 4.0–4.1 min, and re-equilibrated at 10% B for 1.9 min. Flow rate was 0.3 ml/min and sample injection volume was 3 μl. An API 5500 QTrap-MS/MS with an electrospray ionization (ESI) source (AB SCIEX, USA) was used for MS/ MS analysis. An Analyst 1.6.1 software package was used for instrument control and data acquisition. To optimize the MS/MS parameters, the mixed standard solutions of rodenticides (100 ng/ml) were constantly added at a flow rate of 7 μl/ml using a syringe pump in the infusion mode, and at the same time, the declustering potential and collision energies were optimized. A MRM-IDA-EPI scan mode was used in this study. MRM scans were used to quantify the target analytes and obtain the maximum sensitivity. The

Food Anal. Methods

IDA method was employed to trigger the EPI scans by analyzing MRM signals. A total of 26 MRM transitions with addition of IS were used (Table 1), and the dwell time of each MRM transition was 0.02 s. EPI scan was used as a dependent scan for structure identification. The EPI scan was operated from m/z 50 to 600 with a scan speed of 10,000 amu/s and a collision energy of −35 eV with collision energy spread of 15 eV after dynamic background subtraction of the survey scan. The MS scan was performed in negative ion modes and the ESI source was operated at −4500 V and 500 °C. The nebulizer gas (nitrogen) and curtain gas (nitrogen) were set to 50 and 20 psi. The collision gas (nitrogen) was set to medium in MRM scan mode and high in EPI scan mode, respectively. In addition, the high-speed refrigerated centrifuge (Thermo, USA), IKAT18 homogenizer (IKA, Germany), and KQ-300GDV ultrasonic cleaning instrument (China) were used. Further, the omogenizer (IKA Germany) and the Milli-Q ultra-pure water filter (Merck, Germany) were used.

pulverized and homogenized, and then, 10 g of homogenized sample was accurately weighed into a 50-ml centrifuge tube. The samples were spiked 100 μl of internal standard working solution (1 μg/ml) and extracted with 10 ml of acetonitrile to achieve an internal standard concentration of 10 ng/ml (spiked at 10 μg/kg). The solution was vortexed for 1 min. Then, 2.0 g anhydrous MgSO4 and 1.5 g sodium chloride were added and immediately shaken vigorously to prevent formation of MgSO4 agglomerates. Then, the tube was centrifuged at 17,000×g for 2 min at 4 °C. Four milliliters of the upper acetonitrile extract was transferred into a 15-ml tube containing different purification reagents (determined by the matrix of extract). Then, the solution was shaken for 1 min and then centrifuged at 17,000×g for 5 min. One milliliter of the supernatant was evaporated to dryness under a nitrogen flow at 40 °C. The residue was solubilized in 1.0 ml of the initial mobile phase. The solution was ultrasonicated or vortexed to facilitate dissolution. The final extract (containing 1.0 g sample for every 1.0 ml solution) was filtered through 0.2-μm PTFE into an autosampler vial, and 3 μl of the filtrate was injected with the UPLC-MS/MS system.

Sample Preparation Food samples that may cause poisoning were variety. In this study, we studied vegetables (leek, cabbage, cowpea, celery, radish, etc.), fruits (apple, pear, peach, strawberry, grape, etc.), condiments (salt, soy sauce, vinegar, spices, etc.), cooked meat products (sausage, cooked liver, trotters, etc.), and flour. All samples were extracted and purified with modified QuEChERS method. First, the samples were

Results and Discussion Optimization of UPLC-MS/MS Condition Because of the different molecular structure and properties of rodenticides, considerable number of chromatographic

Table 1 UPLC-MS/MS in ESI—acquisition parameters for 12 anticoagulant rodenticides (ordered by retention time), declustering potential, collision energies, and ion ratio used for quantification and identification transitions Peak number

Target compound

Rt (min)

Precursor ion (m/z) [M-H]−

D.P. (V)

Quantification Transition (C.E., eV)

Identification Transition (C.E., eV)

Ion ratioa (%)

1 2 3 4 5 6 7 8 9 10 11 12 (IS)

Valone Pindone Warfarin Coumatetralyl Coumachlor Diphacinone Chlorophacinone Bromadiolone Difenacoum Flocoumafen Brodifacoum Difethialone Warfarin-D5

3.06 3.13 3.15 3.17 3.29 3.31 3.45 3.56 3.61 3.65 3.74 3.76 3.14

229.0 229.0 307.0 291.0 340.9 339.0 373.0 525.0 443.1 541.0 521.2 537.0 312.0

−70 −70 −60 −100 −80 −50 −40 −40 −20 −180 −20 −50 −120

229.0→145.0 (−32) 229.0→171.9 (−28) 307.0→161.0 (−27) 291.0→141.0 (−35) 340.9→161.0 (−30) 339.0→167.0 (−30) 373.0→201.0 (−30) 525.0→250.1 (−48) 443.1→293.0 (−43) 541.0→161.0 (−45) 521.2→135.0(−46) 537.0→151.0 (−50) 312.0→160.9 (−40)

229.0→187.0 (−30) 229.0→145.0 (−35) 307.0→250.1(−30) 291.0→247.1 (−30) 340.9→284.1 (−32) 339.0→171.9 (−28) 373.0→144.9 (−30) 525.0→219.0 (−64) 443.1→135.0 (−43) 541.0→289.1 (−43) 521.2→142.9(−51) 537.0→203.2 (−50) 312.0→255.0 (−40)

7.2 81.8 61.0 39.4 87.0 25.8 18.5 9.0 99.5 49.3 29.8 41.1 71.3

Rt retention time, D.P. declustering potential (V), C.E. collision energy (eV), IS internal standard a

The ion ratio is calculated from mean values obtained from the matrix-matched calibration curves

Food Anal. Methods

methods was studied for their successful separation. The selection of mobile phase can affect the retention time, peak shape, and ionization efficiency of the target compound, thus affecting the sensitivity. In this study, the gradient elution with the mobile phase of 5 mM ammonium acetate aqueous solution (A) and methanol or acetonitrile (B) was investigated and compared. It can be seen that the similar sensitivity and peak shape of the 12 rodenticides were achieved when the mobile phase (B) used was methanol or acetonitrile, as shown in Fig. 1. In order to adapt to the QuEChERS method, we use the 5 mM ammonium acetate aqueous solution (A) and acetonitrile (B) as the mobile phase in this study. To optimize the mass spectrometric parameters of each anticoagulant rodenticide, the single standard solution (100 ng/ml) was injected into the ESI ion source continuously in a flow rate of 7 μl/min with a flow injection pump. Obvious deprotonated molecules [M-H]− at m/z 307.0, m/z 521.0, m/z 537.0, m/z 340.9, m/z 291.0, m/z 525.0, m/z 373.0, m/z 443.1, m/z 339.0, m/z 229.0, m/z 229.0, m/z 541.0, and m/z 312.0 for warfarin, brodifacoum, difethialone, coumachlor, coumatetralyl, bromadiolone, chlorophacinone, difenacoum, diphacinone, pindone, valone, flocoumafen, and warfarinD5, respectively, were observed in the Q1 full scan using negative ion mode. The flow injection of standard solutions (7 μl/ min) was also used in the product ion scan and multi-reaction monitor, and at the same time, the declustering potential and collision energies of different target compounds were optimized. The experiment selected higher abundance of parent/ daughter ions for quantitative analysis to ensure that each peak has at least 12 collection points. The optimized retention time, parent ions, and daughter ions as well as used declustering potential, collision energies, and ion ratio are listed in Table 1. The extracted ion (quantitative) chromatograms of blank cabbage spiked with the 12 anticoagulant rodenticide-

Fig. 1 Chromatograms of 12 rodenticides in blank cabbage using gradient elution with different mobile phases of a 5 mM ammonium acetate aqueous solution–methanol and b 5 mM ammonium acetate aqueous solution–acetonitrile. The rodenticide concentration was 20 ng/ml

mixed standard solutions at 20 ng/ml are shown in Fig. 2. In order to simultaneously achieve the requirements of quantitative and qualitative analysis, a MRM-IDA-EPI scanning and library search mode was used in this study. The retention time, peak height, peak area, and ion ratio were obtained in MRM; a two-stage mass spectrogram was obtained by EPI scan at the same time. Library search can be performed with this twostage mass spectrogram, and it is very useful for confirming the structure of the compound. The MRM chromatogram and the results of search library for the doubtful positive sample are shown in Figs. 3 and 4, respectively. Optimization of the Sample Preparation Method Sample pretreatment about the detection of anticoagulant rodenticides in food samples includes extraction and purification of the analytes. In order to minimize matrix effects and improve chromatographic separation, a modified QuEChERS sample preparation method was studied in this work to improve the effect of extraction and purification. The modified QuEChERS sample preparation method was developed by combining the experience of analyzing rodenticides in blood (Dong et al. 2015), soil (Hernández et al. 2013), and water (Chen et al. 2014) with our published QuEChERS method for the analysis of pesticides in edible fungi (Cao et al. 2016). The advantage of the new method is that it can save time and reduce solvent consumption compared with the conventional SPE method. Because of the differences between the structures and polarities, how to extract the anticoagulant rodenticides to organic solvents was the most important issue in multi-residue methods. The selection of the extracting solvent in sample pretreatment process with a proper polarity to match the analyte was beneficial to improve recovery. The extraction

Food Anal. Methods Fig. 2 The extracted ion (quantification) chromatogram of blank cabbage spiked with the 12 rodenticide-mixed standard solution at 20 ng/ml. 1 valone, 2 pindone, 3 warfarin, 4 coumatetralyl, 5 coumachlor, 6 diphacinone, 7 chlorophacinone, 8 bromadiolone, 9 difenacoum, 10 flocoumafen, 11 brodifacoum, 12 difethialone

solvent usually requires a high dissolving ability for anticoagulant rodenticides and good permeability into the sample matrix. In this study, various organic solvents, e.g., acetone, ethyl acetate, methanol, and acetonitrile, were employed to extract the target analytes from the food samples. The pigments and other impurities in the sample were easy to be extracted from acetone and ethyl acetate, and the matrix effect was increased. When the methanol or acetonitrile was used as an extraction solvent, high extraction efficiency for the 12 compounds was obtained. In order to adapt to the QuEChERS method, we chose the acetonitrile as the extraction solution, and at the same time, the protein in the sample matrix was precipitated. The matrix of the food can interfere with the analytes, resulting in enhancement or suppression of chromatographic peaks and ambiguity of identification. In order to minimize matrix effects, rodenticide extraction effectiveness was studied at different sorbents to improve the purification effect, since the use of single sorbent could not completely remove the co-extractants and eliminate the interference peaks present in the chromatograms. In this study, the recovery of the method was evaluated by comparison of several different sorbents in the samples of different matrices. For the general colorless and colored food samples, we select PSA (100 mg) + MgSO4 (300 mg) and PSA (100 mg) + MgSO4 (300 mg) + GCB (10 mg) as the scavenger, respectively. For the samples with a high fat content (cooked meat products), a step of degreasing was desired. In this work, we use Agilent’s enhanced lipid

removal products (EMR) as the degreaser. The lipid removal efficiency of the EMR is better than that of the C18. EMR can remove most of the lipid impurities in the matrix, so it can greatly reduce the interference of the matrix and improve the recovery of analytes. Matrix Effect Matrix effects are a serious problem in multi-residue analysis. They can severely compromise qualitative and quantitative analysis of the target compounds at trace levels as well as method reproducibility, especially when electrospray ionization is used. In this study, matrix effects were evaluated by comparing the slopes of calibration curves which were prepared in pure solvent (acetonitrile), in purified blank matrix, and in unpurified blank matrix, respectively. We used cabbage, strawberry, vinegar, and sausage as blank matrix to study the matrix effect. In the four matrices studied, signal enhancement was the effect observed in all unpurified blank matrix. Of the four commodities evaluated, cabbage was the one with less matrix effect and sausage was the one with the strongest matrix effect. But for all the 12 kinds of rodenticides, slight matrix enhancement effects occurred in all purified blank matrix. The considerable differences were observed in terms of calibration slopes, using warfarin (Fig. 5a) in cabbage matrix and bromadiolone (Fig. 5b) in sausage matrix as representative examples. In order to obtain

1.4x104

Fig. 3 The MRM chromatogram of the doubtful positive sample

307.0/161.0[-27] 307.0/250.1[-30]

Intensity/cps

1.2x104 1.0x104

Warfarin 0.26 ng/ml

8.0x103 6.0x103 4.0x103 2.0x103 0.0 2

3

Time/min

4

Food Anal. Methods

Fig. 4 Results of index search from library for the doubtful positive sample

more reliable results, a matrix-matched internal standard calibration curve was used for quantification in this work. Additionally, due to the complexity of the sample, matrix components interfered with the analyte ions monitored. Confirmation of peak identity in the samples was also limited by the presence of matrix interferences in some particular cases. In order to confirm the presence of residues in the real samples, the ratio between identification (I) and quantification (Q) has to be in a tolerance level. Because there is no guidance document on analytical quality control and validation procedures for the rodenticide analysis in food, we refer to the SANCO/12571 of the European Commission (Peters et al. 2007). SANCO guide establishes that the difference between I/Q value obtained from a standard solution and I/ Q value obtained from a real sample has to have a maximum tolerance (relative) of ±30% and the retention time of the analyte in the extract should correspond to that of the calibration standard (may need to be matrix-matched) with a tolerance of ±0.2 min. However, when matrix interferences coeluted with analytes, the I/Q ratio could not be properly measured. In this method, using the function of linear ion trap of the QTrap, the MRM information and EPI secondary structure information combined with the library search can be obtained simultaneously. This can help to confirm the structure of the compound and reduce false positives.

Method Validation Calibration curves were constructed by plotting the ratio of the peak area, divided by the peak area of the internal standard, against the analyte concentration. The method developed was validated for different types of food samples, and the different types of food samples free of anticoagulant rodenticides were used for the preparation of a blank matrix. Figure 2 shows total ion chromatogram and extracted ion chromatogram of blank cabbage spiked with the 12 anticoagulant rodenticidemixed standard solutions at 20 ng/ml. The calibration curve was obtained by analyzing blank cabbage spiked with the anticoagulant rodenticides in the content of 0.1, 1, 5, 10, 50, and 100 ng/ml. The six standards were prepared in the initial mobile phase by adding known quantities of standard mixture solution (1 μg/ml) and 100 μl of the internal standard (warfarin-D5, 1 μg/ml). The calibration curves showed good linearity for all 12 anticoagulant rodenticides over the concentration range 0.1–100 ng/ml under the optimized MS-MS conditions. The results shown in Table 2 suggest that good linearity with correlation coefficients (R2) higher than 0.99 was achieved for 12 kinds of anticoagulant rodenticides in the concentration range from 0.1 to 100 ng/ml. The limits of detection (LOD) and the limits of quantification (LOQ) for all compounds were established by extraction

Food Anal. Methods

A 1.2x107

Warfarin

y=10713x-77205 R2=0.9998 y=99171X-66607 R2=0.9998

calibration curves in unpurified blank matrix calibration curves in purified blank matrix 9.0x10

calibration curves in pure solvent (MeCN)

6

Area

Fig. 5 Comparison of the calibration curves obtained in pure solvent (MeCN), in purified blank matrix, and in unpurified blank matrix for a warfarin and b bromadiolone

y=96321x-57487 R2=0.9999 6.0x106

3.0x106

0.0 0

20

40

60

80

100

Concentration(ng/ml)

B 2.5x106

Bromadiolone calibration curves in unpurified blank matrix calibration curves in purified blank matrix

2.0x106

y=21556x-13945 R2=0.9981

calibration curves in pure solvent (MeCN)

y=14623x-2409 R2=0.9976

Area

1.5x10

6

y=12746x-2476 R2=0.9982

1.0x106

5.0x105

0.0 0

20

40

60

80

100

Concentration(ng/ml)

of blank food samples spiked with decreasing concentrations of the analytes. The LOD was determined as the sample concentration that produced a peak with a height three times the level of the baseline noise. The data in Table 2 show that the LODs were in the range of 0.01–0.05 μg/kg. The LOQ was

Table 2 Linearity parameters, correlation coefficients, limit of detection (LOD), and limit of quantification (LOQ) obtained from blank soy sauce spiked with 12 kinds of anticoagulant rodenticides

Peak

Target compound

number

calculated as the sample concentration that produced a peak with a height 10 times the ratio of signal to noise. Furthermore, a signal-to-noise ratio for analytes above 10 was required and the criteria for accuracy ±20% and for precision within 20% (US Department of Health and Human

Correlation coefficient (R2)

LOD (S/N = 3)

LOQ (S/N = 10)

(ng/ml)

Linear range

(μg/kg)

(μg/kg)

1 2 3 4 5 6

Valone Pindone Warfarin Coumatetralyl Coumachlor Diphacinone

0.10–100 0.10–100 0.10–100 0.10–100 0.10–100 0.10–100

0.9985 0.9946 0.9998 0.9994 0.9996 0.9992

0.02 0.02 0.01 0.01 0.02 0.02

0.04 0.04 0.02 0.02 0.02 0.04

7 8 9 10 11 12

Chlorophacinone Bromadiolone Difenacoum Flocoumafen Brodifacoum Difethialone

0.10–100 0.10–100 0.10–100 0.10–100 0.10–100 0.10–100

0.9991 0.9976 0.9929 0.9931 0.9986 0.9958

0.01 0.03 0.02 0.02 0.03 0.05

0.02 0.06 0.04 0.04 0.06 0.10

Food Anal. Methods

Services 2001; European Commission 2013) were demonstrated. The LOQs ranged from 0.02 to 0.10 μg/kg for 12 anticoagulant rodenticides. We validated analytical results by establishing the precision and recovery of the analysis on the different types of food samples. Precision was represented by an estimate of the variability of measurements and the reproducibility of the test method. We tested the recovery of each anticoagulant rodenticide at the three spiked levels (1 LOQ, 5 LOQ, and 10 LOQ) in the blank soy sauce and flour matrix, respectively, to assess the accuracy of the method. Every spiking level was assessed in six repetitions. The precision of the method was described as the value of relative standard deviation (RSD), and the recovery of the assay was calculated from the percentage of the mean calculated concentration to the nominal spiking value. SANCO guide establishes that a quantitative method should be demonstrated as being capable of providing mean recoveries within the range of 70–120% and RSDs lower than 20% (European Commission 2013). Better recoveries were obtained in this study with the average recoveries of all anticoagulant rodenticides ranged from 79.5 to 113.2%. The RSDs were between 1.8 and 13.2%. The recovery data and RSD values obtained are shown in Table 3. Application to Real Samples Our laboratory is responsible for the collection and detection of samples in the foodborne emergent public health events in our city. Half of the poisoning cases processed in our laboratory were involved in anticoagulant rodenticides in recent years. Some of them were caused by ingestion of

contaminated food, some by suicide, and some by homicide. In a poisoning case of suspected anticoagulant rodenticides, the validated method was applied to detect the surplus food, spices, flour, salts, soy sauce, and vinegar which were collected from the scene of the accident. 0.26 μg/kg warfarin was detected in a stir-fried shredded potato, and the MRM chromatogram and the results of search library for the doubtful positive sample are shown in Figs. 3 and 4, respectively.

Conclusions A multi-residue analysis method based on modified QuEChERS preparation method for the simultaneous screening and detection of 12 anticoagulant rodenticides in food samples has been demonstrated. In this work, the tandem mass spectrometry library was established for the 12 anticoagulant rodenticides, and using the MRM-IDA-EPI scan and library search mode, the qualitative and quantitative analysis can be carried out at the same time. The liquid-liquid extraction using acetonitrile was optimized. With PSA, EMR, and GCB as modified QuEChERS adsorbents to clean up the extracts of different matrices is efficient and good extraction recoveries are achieved, most exceeding 79.5%. This developed method is convenient for the promotion and application, and it can improve the emergency detection capabilities of dealing with the rodenticide poisoning cases for all levels of the testing organizations. It was validated according to international guidelines and has been successfully applied to a series of poisoning cases.

Table 3 Average recoveries and precision (n = 6) of the 12 anticoagulant rodenticides spiked in the soy sauce and flour samples at 1 LOQ, 5 LOQ, and 10 LOQ analyte concentrations, respectively Peak number

Target compound

Average recoveries % ± RSD% (n = 6) Spiked in soy sauce

Spiked in flour

1 LOQ

5 LOQ

10 LOQ

1 LOQ

5 LOQ

10 LOQ

1 2 3 4 5 6

Valone Pindone Warfarin Coumatetralyl Coumachlor Diphacinone

80.2 ± 13.2 79.8 ± 7.8 92.6 ± 6.9 89.5 ± 8.9 86.7 ± 6.8 94.2 ± 8.2

95.1 ± 8.2 95.6 ± 6.3 98.8 ± 4.3 98.4 ± 5.6 99.1 ± 5.9 103.9 ± 6.1

98.9 ± 6.2 97.8 ± 6.5 103.2 ± 4.9 103.2 ± 2.9 96.5 ± 4.2 113.2 ± 4.6

85.4 ± 10.5 83.1 ± 8.4 95.6 ± 7.6 92.1 ± 6.5 85.9 ± 7.2 93.8 ± 7.5

99.1 ± 7.1 97.8 ± 5.4 102.1 ± 3.1 105.2 ± 4.3 98.2 ± 5.0 104.6 ± 5.3

97.2 ± 4.4 104.2 ± 2.4 100.2 ± 1.8 102.9 ± 3.1 99.4 ± 2.5 108.1 ± 3.4

7 8 9 10 11 12

Chlorophacinone Bromadiolone Difenacoum Flocoumafen Brodifacoum Difethialone

81.4 ± 12.3 79.5 ± 10.2 87.1 ± 7.6 82.7 ± 7.5 89.1 ± 8.4 84.5 ± 7.5

98.1 ± 8.0 87.6 ± 6.8 100.4 ± 5.2 95.6 ± 6.0 97.1 ± 6.2 98.6 ± 5.8

102.2 ± 6.9 95.6 ± 5.7 106.9 ± 4.7 101.7 ± 4.2 99.2 ± 3.8 101.2 ± 4.7

82.3 ± 10.2 82.1 ± 9.4 91.2 ± 8.9 87.4 ± 8.4 92.3 ± 7.2 88.1 ± 6.9

103 ± 7.9 94.5 ± 7.2 103.2 ± 6.1 99.4 ± 5.7 102.2 ± 5.1 99.4 ± 4.9

100.8 ± 5.5 98.7 ± 5.7 101.8 ± 3.7 104.1 ± 2.8 99.4 ± 2.9 105.2 ± 3.1

Food Anal. Methods Compliance with Ethical Standards Funding This work was supported by medicine and health science and technology development plan of Shandong Province [2015WSA01012]. Conflict of Interest Cao Xiaoli has no conflict of interest. Yang Xiaoqian has no conflict of interest. Liu Zhong has no conflict of interest. Jia Haitao has no conflict of interest. Liu Suhua has no conflict of interest. Liu Lanzheng has no conflict of interest. Qingfen Meng has no conflict of interest. Ethical Approval This article does not contain any studies with human participants or animals performed by any of the authors. Informed Consent Not applicable.

References Berny P, Buronfosse T, Lorgue G (1995) Anticoagulant poisoning in animals: a simple new high-performance thin-layer chromatographic (HPTLC) method for the simultaneous detection of eight anticoagulant rodenticides in liver samples. J Anal Toxicol 19:576–580 Berny P, Caloni F, Croubels S, Sachana M, Vandenbroucke V, Davanzo F, Guitart R (2010) Animal poisoning in Europe. Part 2: companion animals. Vet J 183(3):255–259 Cai MQ, Chen XH, OuYang XK, Jin MC (2009a) Rapid identification and determination of the rodenticide valone in serum by highperformance liquid chromatography–tandem mass spectrometry. J Anal Toxicol 33:104–108 Cai MQ, Dong XY, Chen XH, Jin MC (2009b) An assay for identification and determination of toxic rodenticide valone in serum by ion chromatography–electrospray ionization tandem mass spectrometry with ion trap detector. Talanta 78:242–247 Cao X, Liu S, Yang X, Liu Z, Liu L (2016) A modified QuEChERS sample preparation method for simultaneous determination of 62 pesticide residues in edible fungi using gas chromatography–triple quadrupole mass spectrometry. Food Anal Methods 9:263–274 Chalermchaikit T, Felice LJ, Murphy MJ (1993) Simultaneous determination of eight anticoagulant rodenticides in blood serum and liver. J Anal Toxicol 17:56–61 Chen M l, Zhu G f, Zhou L x, Min J q, Chen X h, Jin M c (2014) Analysis of trace bromadiolone and brodifacoum in environmental water samples by ionic liquid ultrasound-assisted dispersive liquid–liquid microextraction and LC-MS/MS. Anal Methods 6:5879–5885 Chua JD, Friedenberg WR (1998) Superwarfarin poisoning. Arch Intern Med 158:1929–1932 Dong X, Liang S, Sun H (2015) Determination of seven anticoagulant rodenticides in human serum by ultra-performance liquid chromatography-mass spectrometry. Anal Methods 7:1884–1889 Duffield P, Duffield AM, Kennedy M, Birkett DJ, Wade DN (1979) Warfarin and warfarin-alcohol levels in anti-coagulated patients. Aust NZ J Med 9:534–537 European Commission (2013) Guidance document on analytical quality control and validation procedures for pesticide residues analysis in food and feed. SANCO/12571 Felice LJ, Chalermchaikit T, Murphy MJ (1991) Multicomponent determination of 4-hydroxycoumarin anticoagulant rodenticides in blood serum by liquid chromatography with fluorescence detection. J Anal Toxicol 15:126–129 Gómez-Canela C, Barata C, Lacorte S (2014a) Occurrence, elimination, and risk of anticoagulant rodenticides and drugs during wastewater treatment. Environ Sci Pollut Res 21(11):7194–7203

Gómez-Canela C, Vázquez-Chica A, Lacorte S (2014b) Comprehensive characterization of rodenticides in wastewater by liquid chromatography-tandem mass spectrometry. Anal Bioanal Chem 406:345–358 Grobosch T, Angelow B, Schonberg L, Lampe D (2006) Acute bromadilone intoxication. J Anal Toxicol 30:281–286 Guan F, Ishii A, Seno H, Watanabe-Suzuki K, Kumazawa T, Suzuki O, Suzuki W (1999b) Use of an ion-pairing reagent for highperformance liquid chromatography-atmospheric pressure chemical ionization mass spectrometry determination of anionic anticoagulant rodenticides in body fluids. J Chromatogr B 731:155–165 Guan F, Ishii A, Seno H, Watanabe K, Kumazawa T, Suzuki O (1999a) A method for simultaneous determination of five anticoagulant rodenticides in whole blood by high-performance liquid chromatography. J Pharm Biomed Anal 211(1):179–185 Hernández AM, Bernal J, Bernal JL, Martín MT, Caminero C, Nozal MJ (2013) Simultaneous determination of nine anticoagulant rodenticides in soil and water by LC–ESI-MS. J Sep Sci 36:2593–2601 Huang C, Yang J, Du Y, Miao L (2008) Measurement of free concentrations of highly protein-bound warfarin in plasma by ultra performance liquid chromatography–tandem mass spectrometry and its correlation with the international normalized ratio. Clin Chim Acta 393:85–89 Hunter K, Sharp EA, Newton A (1998) Determination of diastereoisomers of bromadiolone, an anticougulant rodenticide, in animal tissues by high-performance liquid chromatography. J Chromatogr A 435:83–95 Jin M, Chen X (2006) Rapid determination of three anticoagulant rodenticides in whole blood by liquid chromatography coupled with electrospray ionization mass spectrometry. Rapid Commun Mass Spectrom 20(18):2741–2746 Jin M, Chen X, Zhu Y (2007a) Determination of five 4-hydroxycoumarin rodenticides in animal liver tissues by ion chromatography with fluorescence detection. J Chromatogr A 1155:57–61 Jin MC, OuYang XK, Xu XM, Ren YP, Chen XH (2007b) Rapid determination of coumatetralyl in human serum by high-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry. Anal Lett 40(4):737–746 Jin MC, OuYang XK, Chen XH (2007c) High-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry for the determination of flocoumafen and brodifacoum in whole blood. J Appl Toxicol 27(1):18–24 Jin MC, Ren YP, Xu XM, Chen XH (2007d) Determination of bromadiolone in whole blood by high-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry. Forensic Sci Int 171:52–56 Jin MC, OuYang XK, Chen XH (2007e) High-performance liquid chromatography coupled with electrospray ionization tandem mass spectrometry for the determination of flucoumafen and brodifadoum in whole blood. J Appl Toxicol 27:18–24 Jin MC, Cai MQ, Chen XH (2009) Simultaneous measurement of indandione-type rodenticides in human serum by liquid chromatography-electrospray ionization-tandem mass spectrometry. J Anal Toxicol 33(6):294–300 Jones A (1996) HPLC determination of anticoagulant rodenticide residues in animal livers. Bull Environ Contam Toxicol 56(1):8–15 Kelly MJ, Chambers J, MacNicoll AD (1993) Simple and rapid method for the determination of the diastereomers of difenacoum in blood and liver using high-performance liquid chromatography with fluorescence detection. J Chromatogr B 620(1):105–112 Kieboom AJ, Rammel CG (1981) Determination of brodifacoum in animal tissues by HPLC. Bull Environ Contam Toxicol 26:674–678 Kuijpers EAP, den hartigh J, Savelkoul JF, de Wolff FA (1995) A method for the simultaneous identification and quantitation of five superwarfarin rodenticides in human serum. J Anal Toxicol 19: 557–562

Food Anal. Methods Maršálek P, Helena M, Veronika D, Vladimír V (2015) Simultaneous determination of ten anticoagulant rodenticides in tissues by column-switching UHPLC-ESI-MS/MS. Anal Bioanal Chem 407: 7849–7854 O’Bryan SM, Constable DJ (1991) Quantification of brodifacoum in plasma and liver tissue by HPLC. J Anal Toxicol 15:144–147 Olmos V, López CM (2007) Brodifacoum poisoning with toxicokinetic data. Clin Toxicol 45:487–489 Palazoglu MG, Tor ER, Holstege DM, Galey FD (1998) Multiresidue analysis of nine anticoagulant rodenticides in serum. J Agric Food Chem 46:4260–4266 Papin F, Clarot F, Vicomte C, Gaulier JM, Daubin C, Chapon F, Vazb E, Proust B (2007) Lethal paradoxical cerebral vein thrombosis due to suspicious anticoagulant rodenticide intoxication with chlorophacinone. Forensic Sci. Int. 166:85–90 Peters FT, Drummer OD, Musshoff F (2007) Validation of new methods. Forensic Sci Int 165:216–224 Rengel I, Friedrich A (1993) Detection of anticoagulant rodenticides (4hydroxycoumarins) by thin-layer chromatography and reversedphase high-performance liquid chromatography with fluorescence detection. Vet Res Commun 17(1993):421–427 Sánchez-Barbudo IS, Camarero PR, Mateo R (2012) Primary and secondary poisoning by anticoagulant rodenticides of non-target animals in Spain. Sci Total Environ 420:280–288 Spahr JE, Maul JS, Rodgers GM (2007) Superwarfarin poisoning: a report of two cases and review of the literature. Am J Hematol 82(7): 656–660

US Department of Health and Human Services. Food and Drug Administration (2001) Guidance for industry: bioanalytical method validation. CDER, Rockville, MD Vandenbroucke V, Desmet N, De Backer P, Croubels S (2008) Multiresidue analysis of eight anticoagulant rodenticides in animal plasma and liver using liquid chromatography combined with heated electrospray ionization tandem mass spectrometry. J Chromatogr B 869(1–2):101–110 Vindenes V, Karinen R, Hasvold I, Bernard J-P, Mørland JG, Christophersen AS (2008) Bromadiolone poisoning: LC–MS method and pharmacokinetic data. J Forensic Sci 53(4):993–996 Watson WA, Litovitz TL, Klein-Schwartz W, Rodgers GC, Youniss J, Reid N, Rouse WG, Rembert RS, D. borys T (2004) 2003 annual report of the American association of poison control centers toxic exposure surveillance system. Am J Emerg Med 22:335–404 Watt BE, Proudfoot AT, Bradberry SM, Vale JA (2005) Anticoagulant rodenticides. Toxicol Rev 24:259–269 Yan H, Xiang P, Zhu L, Shen M (2012) Determination of bromadiolone and brodifacoum in human blood using LC-ESI/MS/MS and its application in four superwarfarin poisoning cases. Forensic Sci. Int. 222:313–317 Zhu L, Yan H, Shen B, Shi Y, Shen M, Xiang P (2013) Determination of bromadiolone and brodifacoum in human hair by liquid chromatography/tandem mass spectrometry and its application to poisoning cases. Rapid Commun Mass Spectrom 27:513–520