This article was downloaded by: [Michigan State University] On: 07 March 2015, At: 12:28 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Food Additives & Contaminants: Part A Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tfac20
Estimation of the validation parameters for a fast analysis of herbicide residues by LC-MS/MS a
a
A. Santilio , S. Girolimetti & D. Attard Barbini
a
a
Department of Environmental and Primary Prevention, Pesticide Unit, National Institute of Health, Rome, Italy Accepted author version posted online: 05 Feb 2014.Published online: 07 Apr 2014.
Click for updates To cite this article: A. Santilio, S. Girolimetti & D. Attard Barbini (2014) Estimation of the validation parameters for a fast analysis of herbicide residues by LC-MS/MS, Food Additives & Contaminants: Part A, 31:5, 845-851, DOI: 10.1080/19440049.2014.891296 To link to this article: http://dx.doi.org/10.1080/19440049.2014.891296
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Food Additives & Contaminants: Part A, 2014 Vol. 31, No. 5, 845–851, http://dx.doi.org/10.1080/19440049.2014.891296
Estimation of the validation parameters for a fast analysis of herbicide residues by LC-MS/MS A. Santilio*, S. Girolimetti and D. Attard Barbini Department of Environmental and Primary Prevention, Pesticide Unit, National Institute of Health, Rome, Italy
Downloaded by [Michigan State University] at 12:28 07 March 2015
(Received 19 June 2013; accepted 31 January 2014) Due to strict regulatory requirements for pesticide residue analysis, only the results of residue analysis with acceptable quality should be reported. As a consequence proper validation of the measurement method is required. In this context, accuracy, precision, specificity, limit of determination (LOQ), matrix effect, linearity, uncertainty calculation and ruggedness become increasingly important. This paper reports a description of the validation parameters of a fast method for the determination of five phenoxy acid herbicides (2,4-D, MCPA, MCPP, haloxyfop and fluazifop) in food crops. The recoveries were performed in the concentration range from 0.05 to 0.5 mg kg–1 for apples, pears, carrots and celeriac with five replicates at each level. The mean recoveries ranged from 70% to 95% for all crops. The precision of the method expressed as a relative standard deviation (RSD%) was found to be in the range 3–14%. For all herbicides, the linearity response of the detector was tested by correlation coefficients (r2 > 0.99) in the concentration range from 0.05 to 0.5 mg kg–1. The LOQ was determined as the lowest spiked level meeting the requirement of accuracy (70–120%) and precision (RSD% < 20% according to European Union guidelines. The uncertainty and robustness were calculated. On the basis of the results, the method can be considered fast, simple and robust and is suitable to be applied to the analysis of studied herbicides in routine testing laboratories. Keywords: phenoxy acid herbicides; validation; pesticide residue; LC-MS/MS
Introduction In pesticide analysis undertaken to monitor residue levels against legal MRLs much effort is put in to achieve an analysis of the appropriate quality. Therefore, quantitative analytical methods always need to be fully validated after their development. In recent years analytical methodologies (Anastassiades et al. 2003; Payá et al. 2007) were developed to improve and support the work of official laboratories in order to have fast and simple analytical methods in compliance with the requirements of European Commission document SANCO/12495/2011. For official laboratories it is very important to have analytical methods of high quality. Therefore, analytical methods should be fast, robust and simple in order to be applied after appropriate validation following quality criteria in routine laboratories (Thompson et al. 2002; Paya et al. 2007; Gómez-Ramos et al. 2013; Ho et al. 2013; Hou et al. 2013). In pesticide analysis intensive method validation is required in order to meet the strict regulations set by the regulatory authorities. According to European Commission document SANCO/12495/2011 a check for accuracy, precision, linearity, specificity, limit of determination (LOQ) and/or robustness is required. This document is also consulted with respect to audit and accreditation of official laboratories for the pesticide residue analysis, according to ISO/IEC/17025. In this context, *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
accuracy, precision, linearity, LOQ determination, uncertainty and robustness testing are gaining interest and they are becoming increasingly important. Several authors have described the validation parameters of analytical methods (Hernández et al. 2006; Leandro et al. 2007; Pizzutti et al. 2007; Kovalczuk et al. 2008; Mol et al. 2012; Hayward et al. 2013; Kittlaus et al. 2013) performed by techniques such as GC-MS/MS and LC-MS/MS. Some authors have focused attention on specific parameters such as uncertainty calculation (Yenisoy-Karakas 2006; Sanyal et al. 2011; Stefanelli et al. 2012) and robustness (Mastovska et al. 2004; Kirchner et al. 2005; Dejaegher & Vander Heyden 2007). In the last few years many laboratories have become involved in the calculation of uncertainty of analytical results instead of SDs because the uncertainty gives an overall measure of the quality of the data. The expanded uncertainty provides an interval within which the value of the analyte is believed to lie with a high level of confidence (Stefanelli et al. 2012). Initially, robustness testing was performed to indicate important factors that could affect the results (reproducibility estimates) of an inter-laboratory study. Therefore, such a test was performed at the end of the method validation. Since a method that is considered not to be sufficiently robust should not be adopted but redeveloped and revalidated, this results in an increase in development time and costs. Nowadays, robustness is verified early in the lifetime
Downloaded by [Michigan State University] at 12:28 07 March 2015
846
A. Santilio et al.
of a method, i.e. at the end of the method development or at the beginning of the validation procedure. The literature described some definitions of ruggedness. According to the United States Pharmacopeia (2006) and International Conference on Harmonisation of Technical Requirements for the Registration of Pharmaceuticals for Human Use (ICH) guidelines the terms of robustness are “The robustness of an analytical method is a measure of its capacity to remain unaffected by small, but deliberate variations in method parameters and provides an indication of its reliability during normal usage” (2005). In addition, according to the ICH guidelines, one consequence of the evaluation of robustness should be that a series of system suitability parameters is established to ensure that the validity of an analytical procedure is maintained whenever it is used. According to this definition, Youden and Steiner (1975) adopted the term “ruggedness test” for an experimental set-up that involves deliberately altering method parameters in order to detect those factors with a large influence (non-rugged factors). The purpose of the work reported here was to validate an analytical method developed by Santilio et al. (2009) according to the requirements of European Commission document SANCO/12495/2011. The Italian National Reference Laboratory for pesticide residues by single methods (NRL-SRM) validated the method, which can be considered useful for the official laboratories for the monitoring purposes. The phenoxy acid herbicides studied cannot be analysed by multi-residue methods due to their chemical properties, and as a consequence they were analysed by a specific method that can be considered a single method, in the framework of the scope of NRL-SRM. In particular, attention was focused on the five herbicides (2,4-D, MCPA, MCPP, haloxyfop and fluazifop) because they are included in the European Union monitoring programmes.
Materials and methods Chemicals and reagents The pesticides studied were 2,4-D acid, MCPA, mecoprop/MCPP, haloxyfop acid and fluazifop acid. The reference standards of these pesticides were purchased from ChemService (West Chester, PA, USA), which were from 98% to 99.9% certified purity. Individual stock standard solutions were prepared at 1 mg ml–1 in acetonitrile and stored in darkness at 4°C. The mixture working solutions were prepared at concentrations of 100 and 10 μg ml–1 and stored at 4°C. Acetonitrile (HPLC grade) was obtained from Carlo Erba (Milan, Italy). Water purified with Elix/Milli-Q water purification system (Millipore, Bedford, MA, USA) was used. Orthophosphoric acid (85%, pure for analysis) was provided by VWR International (West Chester, PA, USA).
Acetic acid (pure for analyses) was purchased from Carlo Erba. Anhydrous magnesium sulfate (97% pure) was supplied from Across Organics (Geel, Belgium). Sodium chloride (98%) was obtained from Merck (Darmstadt, Germany) and purified from interfering substances by heating for 3 h at 600°C and then for 12 h at 120°C in a muffle furnace and an oven, respectively. Anotop 25 LC (0.2 μm pore size) filters were purchased from Whatman (Maidstone, UK). Sample About 1 kg each of carrot, celeriac, apple and pear samples were obtained from a local market. The samples were frozen at –20°C in the laboratory. Before analysis the samples were chopped and divided into subsamples of 10 g each. Each sample was frozen at –20°C until subsequent analysis. Experimental procedure The analytical method used to analyse the herbicides was published previously (Santilio et al. 2009). In this paper, the method was applied to determine residues of phenoxy acid herbicides in vegetable samples (apples, pears, carrots and celeriac) by using a QuEChERS modified method. The method consisted of extraction of the analytes with 10 ml of acetonitrile followed by vortex mixing for 1 min, followed by the addition of 4 g of anhydrous magnesium sulphate, 1 g of sodium chloride and 100 μl of orthophosphoric acid. Due to chemical properties of the compounds, the sample was acidified at pH 2.5 with orthophosphoric acid to allow all compounds to be analysed by LC. Afterwards the homogenate was centrifuged at 3000 rpm for 5 min. The acetonitrile phase was passed through a 0.2 μm anotop filter and collected in a 15 ml tube; the final extract was filled to 10 ml with acetonitrile. The analytes were chromatographically resolved by LCMS/MS using a reversed-phase C18 column (Aquasil C18). Chromatographic and mass spectrometry conditions The LC-MS/MS system comprised a Varian ProStar HPLC system (Varian, Walnut Creek, CA, USA) and a Varian 1200L triple quadrupole LC-MS/MS. The LC was equipped with a Varian 410 autosampler and a Valco injection valve with a loop of 100 μl. The injection mode was partial loop fill. The injection volume was 5 μl, and to avoid carryover the autosampler was flushed with methanol before sample injection. The chromatographic separation was performed on an Aquasil C18 column (dimension 150 × 2.1 mm; with a 5 μm pore size) (Thermo Fisher Scientific, Milan, Italy). Acetonitrile and water acidified with acetic acid (0.1%) were used as the mobile phase at a flow rate of
Food Additives & Contaminants: Part A
847
Table 1. Instrumental conditions. Liquid chromatography conditions Stationary phase Mobile phase Linear gradient: Flow rate Detector conditions Electrospray ionisation interface (ESI) Ionisation voltage Capillary voltage Nebulised gas (synthetic air) Drying gas (heater synthetic)
Downloaded by [Michigan State University] at 12:28 07 March 2015
MRM transitions (Collision energy, eV)a
Aquasil C18 column (150 mm × 2.1 mm; 5 μm) CH3CN/H2O acidified acetic acid (0.1%) pH 2.5 10% CH3CN raise 100% CH3CN within 20 min and held for 7 min at 100% CH3CN after 20% CH3CN (3 min) 0.3 ml min–1 Negative mode 1400 V 40 V 30 psi 250°C at 30 psi Compounds 2,4-D Fluazifop Haloxyfop MCPA MCPP
Transitions 219 219 326 326 326 362 360 201 199 213 213
→ → → → → → → → → → →
125 161 108 226 254 290 288 143 141 105 141
Collision energy (eV) 36 16 20 20 20 44 22
Note: aQuantification ion transitions are shown in bold.
0.3 ml min–1. ESI operating in negative-ion mode was used for all analytes. Data analyses were assured by software Varian MS workstation v.6.8. The LC-MS/MS conditions are given in the Table 1.
Method validation testing For accreditation purposes, the method was validated according to European Commission Document SANCO/ 12495/2011 and the following parameters were determined: accuracy/precision, linearity, specificity, LOQ, matrix effect, uncertainty and robustness.
Results and discussion Accuracy and precision The accuracy of the method was tested by recovery experiments with apple and carrots and extraction was performed by acidification with ortho-phosphoric acid to ensure that all forms of phenoxy herbicides could be converted in the acidic form suitable for LC-MS/MS analysis. Five recovery replicates for apples and carrots were performed in the range of 0.05–0.5 mg kg–1 spiked levels. According to the SANCO/12495/2011 document the recoveries have to be conducted at legal level MRL and
10× MRL. For all investigated compounds the MRL for apples and carrots is 0.05 mg kg–1 with the exception of fluazifop for which the MRLs are 0.2 and 0.3 mg kg–1 for apple and carrots, respectively. For fluazifop the recoveries were tested at 0.05 mg kg–1 considered as the LOQ and at 0.5 mg kg–1 which covers the legal limits. The levels of fortification covered the lower and the higher legal residue levels according to European Union Regulation Nos 396/2005, 149/2008 and 839/2008. On the basis of European Commission document SANCO/12495/2011, the precision of the measurement of a compound was determined by RSD of replicate recoveries, obtained using the same method on the same sample in a single laboratory in a short time, during which differences in the materials and equipment used and/or the analyst involved will not occur. As described in European Commission document SANCO/12495/2011, the criterion of acceptability of the precision is RSD ≤ 20%. The samples were tested before fortification to check for the absence of the compounds of interest. To verify the applicability of the method other recovery experiments were performed with pear and celeriac at 0.05, 0.1 and 0.5 mg kg–1. In this case the recoveries were performed only at the legal limits. For all analysed crops, mean recoveries data, SD and RSD% are given in Table 2. The recoveries ranged from
848
A. Santilio et al.
MCount
response of the detector and it was not used to quantify. Linearity was evaluated by calculation of a five-point linear plot (0.025, 0.05, 0.1, 0.2 and 0.5 mg kg–1) with three replicates, based on linear regression and a squared correlation coefficient. Linearity was demonstrated by correlation coefficients for each compound that were ≥0.999 for each herbicide.
2
3 4 3 1
2
Specificity
Downloaded by [Michigan State University] at 12:28 07 March 2015
1
5
0 10
11
12
13
14 min
Figure 1. Total ion chromatogram (TIC) of a pear sample spiked at the 0.05 mg kg–1 level and a blank sample of pear: (1) 2,4-D; (2) MCPA; (3) MCPP; (4) Fluazifop; and (5) Haloxyfop.
70% to 95% for all matrices and for all spiked levels. The RSD was between 3% and 15%. For both recoveries and RSD% good agreement with the parameters defined in European Commission document SANCO/12495/2011 was obtained. The criteria for acceptability, according to European Commission document SANCO/12495/2011, are for recoveries of 70–120% and for RSDs to be ≤ 20%. Figure 1 shows total ion chromatograms for a pear sample spiked at 0.05 mg kg–1 with the herbicides being studies and a blank pear sample.
Linearity Calibration curves for each compound were obtained from standard solutions in CH3CN in the concentration range from 0.025 to 0.5 mg kg–1. For each compound the calibration curves were used only to verify the linearity
According to European Commission document SANCO/ 12495/2011, the specificity of the method was evaluated in terms of the ability of the detector to provide signals that effectively identify the compounds. LC-MS/MS can be considered a highly specific system. In this sense, the identification of the compounds was performed by comparison of the retention times of the compounds in the sample and the retention times of the compounds in the matrix-matched reference solution and by two ion transitions for each compound. The quantitative determination was performed by the quantification ions and using a matrix-matched reference solution corresponding to the fortification levels. The ion transitions for qualitative and quantitative analyses are shown in Table 1. The target precursor and product ions were selected based on the recommendations from the European Union Reference Laboratory for Residues of Pesticides data pool. Usually, the protonated or deprotonated molecule was used for the precursor ion in order to prevent interference of transition from matrix. Product ions greater than m/z 100 were selected because of they are more diagnostic. Analyte specific MS parameters were optimised by infusion of standard solution for each pesticide.
Limit of determination The LOQ was calculated as the minimum concentration that can be quantified with acceptable accuracy and precision. According to European Commission document SANCO/12495/2011, for accuracy the range of acceptability is 70–120% and precision is RSD% ≤ 20%. LOQ was calculated as 0.05 mg kg–1, which is the lowest
Table 2. Mean recoveries, standard deviation (SD) and relative standard deviation (RSD%) at each fortification level. Apple (n = 5) Compounds 2,4-D Fluazifop Haloxyfop MCPA MCPP
0.05 mg kg–1 77 81 83 79 79
± ± ± ± ±
3; 5; 4; 3; 5;
4 6 5 4 7
Pear (n = 5)
0.5 mg kg–1 70 74 73 72 73
± ± ± ± ±
5; 4; 5; 4; 4;
7 5 7 5 6
Note: Values shown are mean recoveries ± SD; RSD%.
0.05 mg kg–1 87 87 89 89 86
± ± ± ± ±
9; 10 7; 8 12; 14 8; 9 4; 5
Carrots (n = 5)
0.1 mg kg–1 90 83 91 85 95
± ± ± ± ±
6; 4; 5; 4; 5;
7 5 5 5 5
0.05 mg kg–1 81 73 74 71 72
± ± ± ± ±
11; 14 7; 10 7; 10 7; 9 8; 11
Celeriac (n = 5)
0.5 mg kg–1 82 74 77 70 70
± ± ± ± ±
12; 15 5; 7 4; 6 3; 4 5; 7
0.5 mg kg–1 71 82 77 75 77
± ± ± ± ±
3; 4; 3; 2; 3;
5 5 4 3 4
Food Additives & Contaminants: Part A fortification level of the validation method with acceptable accuracy and precision as shown in Table 2.
Table 3. Results of the European proficiency tests on phenoxy acid herbicides, 2008–2010. Compounds/year
Downloaded by [Michigan State University] at 12:28 07 March 2015
Matrix effects The influence of one or more undetected components in the sample on the measurement of the analyte concentration was determined. The response of the LC-MS/MS system may be affected by the presence of co-extractives form the sample (matrix). These matrix effects derive from various physical and chemical processes and may be difficult or impossible to eliminate because they may differ from one matrix to another and also according to the concentration of matrix. The presence or absence of a matrix effect can be demonstrated by comparing the response produced from the analyte in a simple solvent solution with that obtained from the same quantity of analyte in the presence of the sample extract. The matrix effect was studied for all combination substances/matrices at 0.05 and 0.5 mg kg–1 levels according to the following formula: %ME ¼
PeakArea post extraction spike 100 PeakArea Standard
The %ME was calculated for apple and carrots at two different spike levels, and the values were below 100% indicating that the matrix effect is showed with an ionisation suppression for all compounds except for haloxyfop in carrots at 0.05 mg kg–1 with an ME% value of 130% and for MCPA in carrots at 0.5 mg kg–1 with an ME% value of 109%. To compensate the matrix effect matrixmatched standard solutions were used to determine the concentration of compounds for each spiked levels. Uncertainty The measure of uncertainty is a quantitative indicator of the confidence in the analytical data and describes the range around an experiment result within which the true value can be expected to lie within a defined probability. The concentration of the herbicides in the sample, expressed in mg kg–1, is calculated by Equation (1): C ¼ Ac=Ast Cst VfinðmLÞ=PðgÞ
(1)
where Ac is the area of the sample; Ast is the area of the reference standard; Cst is the concentration of the reference standard; Vfin is the final volume of the sample (10 ml); and P(g) is the weight of the sample. Uncertainty was calculated according to the Ishikawa diagram in which we considered the ‘cause’ and ‘effects’ as described by Yenisoy-Karakas (2006). The ‘cause’ comprises the main parameters controlling the results and the ‘effects’ are the result of the analysis. The
849
MCPA/2008 Fluazifop/2008 Fluazifop/2010
Reference value
Results
z-score
0.124 0.084 0.262
0.101 0.066 0.203
−0.7 −0.8 −0.9
parameters that control the results were: weight, volume, dilution and repeatability of the instrument. The recovery contribution was not included in the uncertainty estimation as the recoveries were in the range of 70–120% as recommended by European Commission document SANCO/ 12495/2011 and no correction of the recoveries is needed. As a consequence, because the result is not corrected by recovery, the contribution of recovery to uncertainty estimation is not applied. The contributions to the expanded uncertainty for the investigated crops at two fortification levels ranged between 15% and 41%. The values of relative expanded uncertainty for all analysed matrices are well below the 50% as described by European Commission document SANCO/12495/2011. Moreover, as described by Christensen et al. (2003), participation in the proficiency tests (PTs) can be considered a good approach to update the uncertainty estimation because the PT results provide the contribution of interlaboratory bias. Table 3 shows the results of participation in the European Proficiency Test during 2009–11 on apples and carrots. The z-score parameter was used to evaluate the performance of the laboratory and good results were obtained by applying the method validated. The z-score parameter was calculated according to the European Proficiency Test procedure and the acceptability criteria is:
jzj 2 Acceptable Robustness Robustness was calculated according to the Youden approach (Youden & Steiner 1975). The approach is a fractional factorial design. The basic idea is not to study one alteration at a time but to introduce several variations at once. The interactions between the different factors cannot be detected. The study was carried out by selecting four factors from the method that may influence the measurement results. Such factors may include the analyst, matrix, type of solvent and rate of centrifugation. These factors should be modified in an order of magnitude that matches the deviations usually encountered among laboratories.
850
A. Santilio et al.
Downloaded by [Michigan State University] at 12:28 07 March 2015
When a factor significantly influences the result, it is necessary to perform further investigations to decide the acceptability of the factor. At the end these factors should be indicated in the method. According to the Youden approach some modifications were performed simultaneously and the following factors were modified: type of solvent, rate of centrifuge, operator and matrix. For each test two factors were modified simultaneously and each test was replicated five times and the mean recoveries, SD and variation coefficient were calculated. The average of the results obtained with the parameters in the modified conditions was subtracted from the results obtained in the nominal conditions. To establish if the difference were significantly different from zero, the significantly test was applied using the following formula: 1
1
t ¼ ½ðnÞ =2 0 =Δi==ð2Þ =2 0 s
residues according to European Union Regulation 396/ 2005 and its amendments. The monitoring programme gives information on the substances and samples that need to be checked. For this work, the combination of substances/crops is performed on the basis of both the European Union monitoring programmes and the request from the European Reference Laboratory for single methods during the participation to the PTs. The results obtained in this work indicate that the method was satisfactorily validated and can be applied by official laboratories in the routine analyses for the determination of 2,4-D, MCPA, MCPP, haloxyfop and fluazifop on pomace and vegetables root samples.
References (2)
where n is the number of the test for each parameter; /Δi/ is the absolute value of the difference; and s is the SD of the method calculated as the average of the recoveries obtained in the validation study at two fortification levels. The comparison between the ‘t’ experimental with the ‘t’ theoretic at 95% of probability with seven freedom grades give information on the robustness of the method. The experimental ‘t’ values for each compound were below the theoretic value of 2.36 and for this it can be considered the method robust.
Internal quality criteria To be sure about the quality of results when the method is applied to routine analyses, various internal criteria were established. Firstly, the blank matrix is processed every time in each set of experiments and the extract has to be checked to evaluate the absence of the compounds of interest. The blank is accepted when the signal is less than 30% against to the LOQ. The extraction efficiency is tested by recoveries at 0.05 mg kg–1 (LOQ) efficiency. The recoveries at the LOQ level are considered acceptable when the mean recoveries range from 70% to 120% for all compounds. Finally, the sample is tested twice and the criteria of acceptability based on the relative percentage difference between two values are applied. If the relative difference percentage is less than 57%, the value can be considered acceptable. The value of 57% is calculated considering that the relative standard deviation (CV%) of the repeatability is 20%.
Conclusion To ensure food safety for consumers, the European Union submits monitoring programmes for pesticide residues to test compliance against the legal levels for pesticide
Anastassiades M, Lehotay SJ, Stajnbaher D, Schenck FJ. 2003. Fast and easy multiresidue method employing acetonitrile extraction/partitioning and “dispersive solid-phase extraction” for the determination of pesticide residue in produce. J AOAC Int. 86:412–431. Christensen HB, Poulsen ME, Pedersen M. 2003. Estimation of the uncertainty in a multiresidue method for the determination of pesticide residues in fruit and vegetables. Food Addit Contam. 20:764–775. Dejaegher B, Vander Heyden Y. 2007. Ruggedness and robustness testing. J Chromatogr A. 1158:138–157. European Council. 2005. Regulation (EC) No 396/2005 of the European Parliament and of the Council of 23 February 2005 on maximum residue levels of pesticides in or on food and feed of plant and animal origin and amending Council Directive 91/414/EEC. Off J Eur Union L70:1–16. European Council. 2008. Regulation (EC) No 149/2008 of 29 January 2008 amending Regulation (EC) No 396/2005 of the European Parliament and of the Council by establishing Annex II, III and IV setting maximum residue levels for products covered by Annex I thereto. Off J Eur Union L58:1–398. European Council. 2008. Regulation (EC) No 839/2008 of 30 August 2008 amending Regulation (EC) No 396/2005 of the European Parliament and of the Council as regards Annex II, III and IV on maximum residue levels of pesticides in or on certain products. Off J Eur Union L234:1–216. European Commission SANCO. 2011. Method validation and quality control procedures for pesticide residues analysis in food and feed, 2011. Document SANCO/12495/2011. [Internet]. [cited 2014 Jan 24]. Available from: http://www. eurl-pesticides.eu/library/docs/allcrl/AqcGuidance_Sanco_ 2011_12495.pdf Gómez-Ramos MM, Ferrer C, Malato O, Agüera A, FernándezAlba AR. 2013. Liquid chromatography-high-resolution mass spectrometry for pesticide residue analysis in fruit and vegetables: screening and quantitative studies. J Chromatogr A. 1287:24–37. Guidelines prepared within the International Conference on Harmonization of Technical Requirements for the Registration of pharmaceuticals and Human Use (ICH), Validation of Analytical Procedures: text and Methodology Q2(R1) November 2005. [Internet]. [cited 2014 Jan 24]. Available from: http://www.ich.org/products/guidelines/ quality/article/quality-guidelines.html
Downloaded by [Michigan State University] at 12:28 07 March 2015
Food Additives & Contaminants: Part A Hayward DG, Wong JW, Shi F, Zhang K, Lee NS, Dibenedetto AL, Hengel MJ. 2013. Multiresidue pesticide analysis of botanic dietary supplements using salt-out acetonitrile extraction, solid-phase extraction, clean up column, and gas chromatography-triple quadrupole mass spectrometry. Anal Chem. May 7. 85:686–693. Hernández F, Pozo OJ, Sancho JV, Bijlsma L, Barreda M, Pitarch E. 2006. Multiresidue liquid chromatography tandem mass spectrometry determination of 52 non gas chromatographyamenable pesticides and metabolites in different food commodities. J Chromatogr A. 1109:242–252. Ho YM, Tsoi YK, Leung KS. 2013. Highly sensitive and selective organophosphate screening in twelve commodities of fruits, vegetables and herbal medicines by dispersive liquid-liquid microextraction. Anal Chim Acta. May. 775:58–66. Hou X, Han M, Dai X, Yang X, Yi S. 2013. A multi-residue method for the determination of 124 pesticides in rice by modified QuEChERS extraction and gas chromatography-tan dem mass spectrometry. Food Chem. Jun. 1. 138:1198–1205. ISO/IEC/17025 General requirements for the competence of testing and calibration laboratories. September 2005 Kirchner M, Matisová E, Otrekal R, Hercegová A, de Zeeuw J. 2005. Search on ruggedness of fast gas chromatography-mass spectrometry in pesticide residues analysis. J Chromatogr A. 1084:63–70. Kittlaus S, Schimanke J, Kempe G, Speer K. 2013. Development and validation of an efficient automated method for the analysis of 300 pesticides in foods using two-dimensional liquid chromatography-tandem mass spectrometry. J Chromatogr A. 29:1283–1298. Kovalczuk T, Lacina O, Jech M, Poustka J, Hajšlová J. 2008. Novel approach to fast determination of multiple pesticide residues using ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Food Addit Contam: Part A. April 25:444–457. Leandro CC, Hancock P, Fussell RJ, Keely BJ. 2007. Ultraperformance liquid chromatography for the determination of pesticide residues in foods by tandem quadrupole mass spectrometry with polarity switching. J Chromatogr A. 1144:161–169. Mastovska K, Hajslova J, Lehotay SJ. 2004. Ruggedness and other performance characteristics of low pressure gas chromatography-mass spectrometry for the fast analysis of
851
multiple pesticide residues in food crops. J Chromatogr. A. 1054:335–349. Mol HGJ, Zomer P, de Koning M. 2012. Qualitative aspects and validation of a screening method for pesticides in vegetables and fruits based on liquid chromatography coupled to full scan high resolution (Orbitrap) mass spectrometry. Anal Bioanal Chem. July 403:2891–2908. Payá P, Anastassiades M, Mack D, Sigalova I, Tasdelen B, Oliva J, Barba A. 2007. Analysis of pesticide residues using the Quick Easy Cheap Effective Rugged and Safe (QuEChERS) pesticide multiresidue method in combination with gas and liquid chromatography and tandem mass spectrometric detection. Anal Bioanal Chem. 389:1697–1714. Pizzutti IR, de Kok A, Zanella R, Adaime MB, Hiemstra M, Wickert C, Prestes OD. 2007. Method validation for the analysis of 169 pesticides in soya grain, without cleanup, by liquid chromatography-tandem mass spectrometry using positive and negative electrospray ionization. J Chromatogr A. 1142:123–136. Santilio A, Stefanelli P, Dommarco R. 2009. Fast determination of phenoxy acid herbicides in carrots and apples using liquid chromatography coupled triple quadrupole mass spectrometry. J Environ Sci. Health, Part B. 44:584–590. Sanyal D, Rani A, Alam S, Gupta R. 2011. Development, validation and uncertainty measurement of multi residue analysis of organochlorine and organophosphorus pesticides using pressurized liquid extraction and dispersive-SPE techniques. Environ Monitor Assess. 182:97–113. Stefanelli P, Barbini DA, Girolimetti S, Dommarco R. 2012. Estimation of measurement uncertainty associated to the determination of pesticide residues: a case study. J Environ Sci Health, Part B. 47:804–813. Thompson M, Ellison SLR, Wood R. 2002. Harmonized guidelines for single-laboratory validation of methods of analysis (IUPAC Technical Report). Pure Appl Chem. 74:835–855. United States Pharmacopoeia. 2006. National formulary. 24 ed. Rockville (MD): United States Pharmacopoeial Convention. Yenisoy-Karakaş S. 2006. Validation and uncertainty assessment of rapid extraction and cleanup methods for the determination of 16 organochlorine pesticide residues in vegetables. Anal Chim Acta. 571:298–307. Youden WJ, Steiner EH. 1975. Statistical manual of the association of official analytical chemists. The association of official analytical chemists. Washington (DC): AOAC; p. 33, 70, 82.
Food Additives & Contaminants: Part A Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tfac20
Estimation of the validation parameters for a fast analysis of herbicide residues by LC-MS/MS a
a
A. Santilio , S. Girolimetti & D. Attard Barbini
a
a
Department of Environmental and Primary Prevention, Pesticide Unit, National Institute of Health, Rome, Italy Accepted author version posted online: 05 Feb 2014.Published online: 07 Apr 2014.
Click for updates To cite this article: A. Santilio, S. Girolimetti & D. Attard Barbini (2014) Estimation of the validation parameters for a fast analysis of herbicide residues by LC-MS/MS, Food Additives & Contaminants: Part A, 31:5, 845-851, DOI: 10.1080/19440049.2014.891296 To link to this article: http://dx.doi.org/10.1080/19440049.2014.891296
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http:// www.tandfonline.com/page/terms-and-conditions
Food Additives & Contaminants: Part A, 2014 Vol. 31, No. 5, 845–851, http://dx.doi.org/10.1080/19440049.2014.891296
Estimation of the validation parameters for a fast analysis of herbicide residues by LC-MS/MS A. Santilio*, S. Girolimetti and D. Attard Barbini Department of Environmental and Primary Prevention, Pesticide Unit, National Institute of Health, Rome, Italy
Downloaded by [Michigan State University] at 12:28 07 March 2015
(Received 19 June 2013; accepted 31 January 2014) Due to strict regulatory requirements for pesticide residue analysis, only the results of residue analysis with acceptable quality should be reported. As a consequence proper validation of the measurement method is required. In this context, accuracy, precision, specificity, limit of determination (LOQ), matrix effect, linearity, uncertainty calculation and ruggedness become increasingly important. This paper reports a description of the validation parameters of a fast method for the determination of five phenoxy acid herbicides (2,4-D, MCPA, MCPP, haloxyfop and fluazifop) in food crops. The recoveries were performed in the concentration range from 0.05 to 0.5 mg kg–1 for apples, pears, carrots and celeriac with five replicates at each level. The mean recoveries ranged from 70% to 95% for all crops. The precision of the method expressed as a relative standard deviation (RSD%) was found to be in the range 3–14%. For all herbicides, the linearity response of the detector was tested by correlation coefficients (r2 > 0.99) in the concentration range from 0.05 to 0.5 mg kg–1. The LOQ was determined as the lowest spiked level meeting the requirement of accuracy (70–120%) and precision (RSD% < 20% according to European Union guidelines. The uncertainty and robustness were calculated. On the basis of the results, the method can be considered fast, simple and robust and is suitable to be applied to the analysis of studied herbicides in routine testing laboratories. Keywords: phenoxy acid herbicides; validation; pesticide residue; LC-MS/MS
Introduction In pesticide analysis undertaken to monitor residue levels against legal MRLs much effort is put in to achieve an analysis of the appropriate quality. Therefore, quantitative analytical methods always need to be fully validated after their development. In recent years analytical methodologies (Anastassiades et al. 2003; Payá et al. 2007) were developed to improve and support the work of official laboratories in order to have fast and simple analytical methods in compliance with the requirements of European Commission document SANCO/12495/2011. For official laboratories it is very important to have analytical methods of high quality. Therefore, analytical methods should be fast, robust and simple in order to be applied after appropriate validation following quality criteria in routine laboratories (Thompson et al. 2002; Paya et al. 2007; Gómez-Ramos et al. 2013; Ho et al. 2013; Hou et al. 2013). In pesticide analysis intensive method validation is required in order to meet the strict regulations set by the regulatory authorities. According to European Commission document SANCO/12495/2011 a check for accuracy, precision, linearity, specificity, limit of determination (LOQ) and/or robustness is required. This document is also consulted with respect to audit and accreditation of official laboratories for the pesticide residue analysis, according to ISO/IEC/17025. In this context, *Corresponding author. Email: [email protected] © 2014 Taylor & Francis
accuracy, precision, linearity, LOQ determination, uncertainty and robustness testing are gaining interest and they are becoming increasingly important. Several authors have described the validation parameters of analytical methods (Hernández et al. 2006; Leandro et al. 2007; Pizzutti et al. 2007; Kovalczuk et al. 2008; Mol et al. 2012; Hayward et al. 2013; Kittlaus et al. 2013) performed by techniques such as GC-MS/MS and LC-MS/MS. Some authors have focused attention on specific parameters such as uncertainty calculation (Yenisoy-Karakas 2006; Sanyal et al. 2011; Stefanelli et al. 2012) and robustness (Mastovska et al. 2004; Kirchner et al. 2005; Dejaegher & Vander Heyden 2007). In the last few years many laboratories have become involved in the calculation of uncertainty of analytical results instead of SDs because the uncertainty gives an overall measure of the quality of the data. The expanded uncertainty provides an interval within which the value of the analyte is believed to lie with a high level of confidence (Stefanelli et al. 2012). Initially, robustness testing was performed to indicate important factors that could affect the results (reproducibility estimates) of an inter-laboratory study. Therefore, such a test was performed at the end of the method validation. Since a method that is considered not to be sufficiently robust should not be adopted but redeveloped and revalidated, this results in an increase in development time and costs. Nowadays, robustness is verified early in the lifetime
Downloaded by [Michigan State University] at 12:28 07 March 2015
846
A. Santilio et al.
of a method, i.e. at the end of the method development or at the beginning of the validation procedure. The literature described some definitions of ruggedness. According to the United States Pharmacopeia (2006) and International Conference on Harmonisation of Technical Requirements for the Registration of Pharmaceuticals for Human Use (ICH) guidelines the terms of robustness are “The robustness of an analytical method is a measure of its capacity to remain unaffected by small, but deliberate variations in method parameters and provides an indication of its reliability during normal usage” (2005). In addition, according to the ICH guidelines, one consequence of the evaluation of robustness should be that a series of system suitability parameters is established to ensure that the validity of an analytical procedure is maintained whenever it is used. According to this definition, Youden and Steiner (1975) adopted the term “ruggedness test” for an experimental set-up that involves deliberately altering method parameters in order to detect those factors with a large influence (non-rugged factors). The purpose of the work reported here was to validate an analytical method developed by Santilio et al. (2009) according to the requirements of European Commission document SANCO/12495/2011. The Italian National Reference Laboratory for pesticide residues by single methods (NRL-SRM) validated the method, which can be considered useful for the official laboratories for the monitoring purposes. The phenoxy acid herbicides studied cannot be analysed by multi-residue methods due to their chemical properties, and as a consequence they were analysed by a specific method that can be considered a single method, in the framework of the scope of NRL-SRM. In particular, attention was focused on the five herbicides (2,4-D, MCPA, MCPP, haloxyfop and fluazifop) because they are included in the European Union monitoring programmes.
Materials and methods Chemicals and reagents The pesticides studied were 2,4-D acid, MCPA, mecoprop/MCPP, haloxyfop acid and fluazifop acid. The reference standards of these pesticides were purchased from ChemService (West Chester, PA, USA), which were from 98% to 99.9% certified purity. Individual stock standard solutions were prepared at 1 mg ml–1 in acetonitrile and stored in darkness at 4°C. The mixture working solutions were prepared at concentrations of 100 and 10 μg ml–1 and stored at 4°C. Acetonitrile (HPLC grade) was obtained from Carlo Erba (Milan, Italy). Water purified with Elix/Milli-Q water purification system (Millipore, Bedford, MA, USA) was used. Orthophosphoric acid (85%, pure for analysis) was provided by VWR International (West Chester, PA, USA).
Acetic acid (pure for analyses) was purchased from Carlo Erba. Anhydrous magnesium sulfate (97% pure) was supplied from Across Organics (Geel, Belgium). Sodium chloride (98%) was obtained from Merck (Darmstadt, Germany) and purified from interfering substances by heating for 3 h at 600°C and then for 12 h at 120°C in a muffle furnace and an oven, respectively. Anotop 25 LC (0.2 μm pore size) filters were purchased from Whatman (Maidstone, UK). Sample About 1 kg each of carrot, celeriac, apple and pear samples were obtained from a local market. The samples were frozen at –20°C in the laboratory. Before analysis the samples were chopped and divided into subsamples of 10 g each. Each sample was frozen at –20°C until subsequent analysis. Experimental procedure The analytical method used to analyse the herbicides was published previously (Santilio et al. 2009). In this paper, the method was applied to determine residues of phenoxy acid herbicides in vegetable samples (apples, pears, carrots and celeriac) by using a QuEChERS modified method. The method consisted of extraction of the analytes with 10 ml of acetonitrile followed by vortex mixing for 1 min, followed by the addition of 4 g of anhydrous magnesium sulphate, 1 g of sodium chloride and 100 μl of orthophosphoric acid. Due to chemical properties of the compounds, the sample was acidified at pH 2.5 with orthophosphoric acid to allow all compounds to be analysed by LC. Afterwards the homogenate was centrifuged at 3000 rpm for 5 min. The acetonitrile phase was passed through a 0.2 μm anotop filter and collected in a 15 ml tube; the final extract was filled to 10 ml with acetonitrile. The analytes were chromatographically resolved by LCMS/MS using a reversed-phase C18 column (Aquasil C18). Chromatographic and mass spectrometry conditions The LC-MS/MS system comprised a Varian ProStar HPLC system (Varian, Walnut Creek, CA, USA) and a Varian 1200L triple quadrupole LC-MS/MS. The LC was equipped with a Varian 410 autosampler and a Valco injection valve with a loop of 100 μl. The injection mode was partial loop fill. The injection volume was 5 μl, and to avoid carryover the autosampler was flushed with methanol before sample injection. The chromatographic separation was performed on an Aquasil C18 column (dimension 150 × 2.1 mm; with a 5 μm pore size) (Thermo Fisher Scientific, Milan, Italy). Acetonitrile and water acidified with acetic acid (0.1%) were used as the mobile phase at a flow rate of
Food Additives & Contaminants: Part A
847
Table 1. Instrumental conditions. Liquid chromatography conditions Stationary phase Mobile phase Linear gradient: Flow rate Detector conditions Electrospray ionisation interface (ESI) Ionisation voltage Capillary voltage Nebulised gas (synthetic air) Drying gas (heater synthetic)
Downloaded by [Michigan State University] at 12:28 07 March 2015
MRM transitions (Collision energy, eV)a
Aquasil C18 column (150 mm × 2.1 mm; 5 μm) CH3CN/H2O acidified acetic acid (0.1%) pH 2.5 10% CH3CN raise 100% CH3CN within 20 min and held for 7 min at 100% CH3CN after 20% CH3CN (3 min) 0.3 ml min–1 Negative mode 1400 V 40 V 30 psi 250°C at 30 psi Compounds 2,4-D Fluazifop Haloxyfop MCPA MCPP
Transitions 219 219 326 326 326 362 360 201 199 213 213
→ → → → → → → → → → →
125 161 108 226 254 290 288 143 141 105 141
Collision energy (eV) 36 16 20 20 20 44 22
Note: aQuantification ion transitions are shown in bold.
0.3 ml min–1. ESI operating in negative-ion mode was used for all analytes. Data analyses were assured by software Varian MS workstation v.6.8. The LC-MS/MS conditions are given in the Table 1.
Method validation testing For accreditation purposes, the method was validated according to European Commission Document SANCO/ 12495/2011 and the following parameters were determined: accuracy/precision, linearity, specificity, LOQ, matrix effect, uncertainty and robustness.
Results and discussion Accuracy and precision The accuracy of the method was tested by recovery experiments with apple and carrots and extraction was performed by acidification with ortho-phosphoric acid to ensure that all forms of phenoxy herbicides could be converted in the acidic form suitable for LC-MS/MS analysis. Five recovery replicates for apples and carrots were performed in the range of 0.05–0.5 mg kg–1 spiked levels. According to the SANCO/12495/2011 document the recoveries have to be conducted at legal level MRL and
10× MRL. For all investigated compounds the MRL for apples and carrots is 0.05 mg kg–1 with the exception of fluazifop for which the MRLs are 0.2 and 0.3 mg kg–1 for apple and carrots, respectively. For fluazifop the recoveries were tested at 0.05 mg kg–1 considered as the LOQ and at 0.5 mg kg–1 which covers the legal limits. The levels of fortification covered the lower and the higher legal residue levels according to European Union Regulation Nos 396/2005, 149/2008 and 839/2008. On the basis of European Commission document SANCO/12495/2011, the precision of the measurement of a compound was determined by RSD of replicate recoveries, obtained using the same method on the same sample in a single laboratory in a short time, during which differences in the materials and equipment used and/or the analyst involved will not occur. As described in European Commission document SANCO/12495/2011, the criterion of acceptability of the precision is RSD ≤ 20%. The samples were tested before fortification to check for the absence of the compounds of interest. To verify the applicability of the method other recovery experiments were performed with pear and celeriac at 0.05, 0.1 and 0.5 mg kg–1. In this case the recoveries were performed only at the legal limits. For all analysed crops, mean recoveries data, SD and RSD% are given in Table 2. The recoveries ranged from
848
A. Santilio et al.
MCount
response of the detector and it was not used to quantify. Linearity was evaluated by calculation of a five-point linear plot (0.025, 0.05, 0.1, 0.2 and 0.5 mg kg–1) with three replicates, based on linear regression and a squared correlation coefficient. Linearity was demonstrated by correlation coefficients for each compound that were ≥0.999 for each herbicide.
2
3 4 3 1
2
Specificity
Downloaded by [Michigan State University] at 12:28 07 March 2015
1
5
0 10
11
12
13
14 min
Figure 1. Total ion chromatogram (TIC) of a pear sample spiked at the 0.05 mg kg–1 level and a blank sample of pear: (1) 2,4-D; (2) MCPA; (3) MCPP; (4) Fluazifop; and (5) Haloxyfop.
70% to 95% for all matrices and for all spiked levels. The RSD was between 3% and 15%. For both recoveries and RSD% good agreement with the parameters defined in European Commission document SANCO/12495/2011 was obtained. The criteria for acceptability, according to European Commission document SANCO/12495/2011, are for recoveries of 70–120% and for RSDs to be ≤ 20%. Figure 1 shows total ion chromatograms for a pear sample spiked at 0.05 mg kg–1 with the herbicides being studies and a blank pear sample.
Linearity Calibration curves for each compound were obtained from standard solutions in CH3CN in the concentration range from 0.025 to 0.5 mg kg–1. For each compound the calibration curves were used only to verify the linearity
According to European Commission document SANCO/ 12495/2011, the specificity of the method was evaluated in terms of the ability of the detector to provide signals that effectively identify the compounds. LC-MS/MS can be considered a highly specific system. In this sense, the identification of the compounds was performed by comparison of the retention times of the compounds in the sample and the retention times of the compounds in the matrix-matched reference solution and by two ion transitions for each compound. The quantitative determination was performed by the quantification ions and using a matrix-matched reference solution corresponding to the fortification levels. The ion transitions for qualitative and quantitative analyses are shown in Table 1. The target precursor and product ions were selected based on the recommendations from the European Union Reference Laboratory for Residues of Pesticides data pool. Usually, the protonated or deprotonated molecule was used for the precursor ion in order to prevent interference of transition from matrix. Product ions greater than m/z 100 were selected because of they are more diagnostic. Analyte specific MS parameters were optimised by infusion of standard solution for each pesticide.
Limit of determination The LOQ was calculated as the minimum concentration that can be quantified with acceptable accuracy and precision. According to European Commission document SANCO/12495/2011, for accuracy the range of acceptability is 70–120% and precision is RSD% ≤ 20%. LOQ was calculated as 0.05 mg kg–1, which is the lowest
Table 2. Mean recoveries, standard deviation (SD) and relative standard deviation (RSD%) at each fortification level. Apple (n = 5) Compounds 2,4-D Fluazifop Haloxyfop MCPA MCPP
0.05 mg kg–1 77 81 83 79 79
± ± ± ± ±
3; 5; 4; 3; 5;
4 6 5 4 7
Pear (n = 5)
0.5 mg kg–1 70 74 73 72 73
± ± ± ± ±
5; 4; 5; 4; 4;
7 5 7 5 6
Note: Values shown are mean recoveries ± SD; RSD%.
0.05 mg kg–1 87 87 89 89 86
± ± ± ± ±
9; 10 7; 8 12; 14 8; 9 4; 5
Carrots (n = 5)
0.1 mg kg–1 90 83 91 85 95
± ± ± ± ±
6; 4; 5; 4; 5;
7 5 5 5 5
0.05 mg kg–1 81 73 74 71 72
± ± ± ± ±
11; 14 7; 10 7; 10 7; 9 8; 11
Celeriac (n = 5)
0.5 mg kg–1 82 74 77 70 70
± ± ± ± ±
12; 15 5; 7 4; 6 3; 4 5; 7
0.5 mg kg–1 71 82 77 75 77
± ± ± ± ±
3; 4; 3; 2; 3;
5 5 4 3 4
Food Additives & Contaminants: Part A fortification level of the validation method with acceptable accuracy and precision as shown in Table 2.
Table 3. Results of the European proficiency tests on phenoxy acid herbicides, 2008–2010. Compounds/year
Downloaded by [Michigan State University] at 12:28 07 March 2015
Matrix effects The influence of one or more undetected components in the sample on the measurement of the analyte concentration was determined. The response of the LC-MS/MS system may be affected by the presence of co-extractives form the sample (matrix). These matrix effects derive from various physical and chemical processes and may be difficult or impossible to eliminate because they may differ from one matrix to another and also according to the concentration of matrix. The presence or absence of a matrix effect can be demonstrated by comparing the response produced from the analyte in a simple solvent solution with that obtained from the same quantity of analyte in the presence of the sample extract. The matrix effect was studied for all combination substances/matrices at 0.05 and 0.5 mg kg–1 levels according to the following formula: %ME ¼
PeakArea post extraction spike 100 PeakArea Standard
The %ME was calculated for apple and carrots at two different spike levels, and the values were below 100% indicating that the matrix effect is showed with an ionisation suppression for all compounds except for haloxyfop in carrots at 0.05 mg kg–1 with an ME% value of 130% and for MCPA in carrots at 0.5 mg kg–1 with an ME% value of 109%. To compensate the matrix effect matrixmatched standard solutions were used to determine the concentration of compounds for each spiked levels. Uncertainty The measure of uncertainty is a quantitative indicator of the confidence in the analytical data and describes the range around an experiment result within which the true value can be expected to lie within a defined probability. The concentration of the herbicides in the sample, expressed in mg kg–1, is calculated by Equation (1): C ¼ Ac=Ast Cst VfinðmLÞ=PðgÞ
(1)
where Ac is the area of the sample; Ast is the area of the reference standard; Cst is the concentration of the reference standard; Vfin is the final volume of the sample (10 ml); and P(g) is the weight of the sample. Uncertainty was calculated according to the Ishikawa diagram in which we considered the ‘cause’ and ‘effects’ as described by Yenisoy-Karakas (2006). The ‘cause’ comprises the main parameters controlling the results and the ‘effects’ are the result of the analysis. The
849
MCPA/2008 Fluazifop/2008 Fluazifop/2010
Reference value
Results
z-score
0.124 0.084 0.262
0.101 0.066 0.203
−0.7 −0.8 −0.9
parameters that control the results were: weight, volume, dilution and repeatability of the instrument. The recovery contribution was not included in the uncertainty estimation as the recoveries were in the range of 70–120% as recommended by European Commission document SANCO/ 12495/2011 and no correction of the recoveries is needed. As a consequence, because the result is not corrected by recovery, the contribution of recovery to uncertainty estimation is not applied. The contributions to the expanded uncertainty for the investigated crops at two fortification levels ranged between 15% and 41%. The values of relative expanded uncertainty for all analysed matrices are well below the 50% as described by European Commission document SANCO/12495/2011. Moreover, as described by Christensen et al. (2003), participation in the proficiency tests (PTs) can be considered a good approach to update the uncertainty estimation because the PT results provide the contribution of interlaboratory bias. Table 3 shows the results of participation in the European Proficiency Test during 2009–11 on apples and carrots. The z-score parameter was used to evaluate the performance of the laboratory and good results were obtained by applying the method validated. The z-score parameter was calculated according to the European Proficiency Test procedure and the acceptability criteria is:
jzj 2 Acceptable Robustness Robustness was calculated according to the Youden approach (Youden & Steiner 1975). The approach is a fractional factorial design. The basic idea is not to study one alteration at a time but to introduce several variations at once. The interactions between the different factors cannot be detected. The study was carried out by selecting four factors from the method that may influence the measurement results. Such factors may include the analyst, matrix, type of solvent and rate of centrifugation. These factors should be modified in an order of magnitude that matches the deviations usually encountered among laboratories.
850
A. Santilio et al.
Downloaded by [Michigan State University] at 12:28 07 March 2015
When a factor significantly influences the result, it is necessary to perform further investigations to decide the acceptability of the factor. At the end these factors should be indicated in the method. According to the Youden approach some modifications were performed simultaneously and the following factors were modified: type of solvent, rate of centrifuge, operator and matrix. For each test two factors were modified simultaneously and each test was replicated five times and the mean recoveries, SD and variation coefficient were calculated. The average of the results obtained with the parameters in the modified conditions was subtracted from the results obtained in the nominal conditions. To establish if the difference were significantly different from zero, the significantly test was applied using the following formula: 1
1
t ¼ ½ðnÞ =2 0 =Δi==ð2Þ =2 0 s
residues according to European Union Regulation 396/ 2005 and its amendments. The monitoring programme gives information on the substances and samples that need to be checked. For this work, the combination of substances/crops is performed on the basis of both the European Union monitoring programmes and the request from the European Reference Laboratory for single methods during the participation to the PTs. The results obtained in this work indicate that the method was satisfactorily validated and can be applied by official laboratories in the routine analyses for the determination of 2,4-D, MCPA, MCPP, haloxyfop and fluazifop on pomace and vegetables root samples.
References (2)
where n is the number of the test for each parameter; /Δi/ is the absolute value of the difference; and s is the SD of the method calculated as the average of the recoveries obtained in the validation study at two fortification levels. The comparison between the ‘t’ experimental with the ‘t’ theoretic at 95% of probability with seven freedom grades give information on the robustness of the method. The experimental ‘t’ values for each compound were below the theoretic value of 2.36 and for this it can be considered the method robust.
Internal quality criteria To be sure about the quality of results when the method is applied to routine analyses, various internal criteria were established. Firstly, the blank matrix is processed every time in each set of experiments and the extract has to be checked to evaluate the absence of the compounds of interest. The blank is accepted when the signal is less than 30% against to the LOQ. The extraction efficiency is tested by recoveries at 0.05 mg kg–1 (LOQ) efficiency. The recoveries at the LOQ level are considered acceptable when the mean recoveries range from 70% to 120% for all compounds. Finally, the sample is tested twice and the criteria of acceptability based on the relative percentage difference between two values are applied. If the relative difference percentage is less than 57%, the value can be considered acceptable. The value of 57% is calculated considering that the relative standard deviation (CV%) of the repeatability is 20%.
Conclusion To ensure food safety for consumers, the European Union submits monitoring programmes for pesticide residues to test compliance against the legal levels for pesticide
Anastassiades M, Lehotay SJ, Stajnbaher D, Schenck FJ. 2003. Fast and easy multiresidue method employing acetonitrile extraction/partitioning and “dispersive solid-phase extraction” for the determination of pesticide residue in produce. J AOAC Int. 86:412–431. Christensen HB, Poulsen ME, Pedersen M. 2003. Estimation of the uncertainty in a multiresidue method for the determination of pesticide residues in fruit and vegetables. Food Addit Contam. 20:764–775. Dejaegher B, Vander Heyden Y. 2007. Ruggedness and robustness testing. J Chromatogr A. 1158:138–157. European Council. 2005. Regulation (EC) No 396/2005 of the European Parliament and of the Council of 23 February 2005 on maximum residue levels of pesticides in or on food and feed of plant and animal origin and amending Council Directive 91/414/EEC. Off J Eur Union L70:1–16. European Council. 2008. Regulation (EC) No 149/2008 of 29 January 2008 amending Regulation (EC) No 396/2005 of the European Parliament and of the Council by establishing Annex II, III and IV setting maximum residue levels for products covered by Annex I thereto. Off J Eur Union L58:1–398. European Council. 2008. Regulation (EC) No 839/2008 of 30 August 2008 amending Regulation (EC) No 396/2005 of the European Parliament and of the Council as regards Annex II, III and IV on maximum residue levels of pesticides in or on certain products. Off J Eur Union L234:1–216. European Commission SANCO. 2011. Method validation and quality control procedures for pesticide residues analysis in food and feed, 2011. Document SANCO/12495/2011. [Internet]. [cited 2014 Jan 24]. Available from: http://www. eurl-pesticides.eu/library/docs/allcrl/AqcGuidance_Sanco_ 2011_12495.pdf Gómez-Ramos MM, Ferrer C, Malato O, Agüera A, FernándezAlba AR. 2013. Liquid chromatography-high-resolution mass spectrometry for pesticide residue analysis in fruit and vegetables: screening and quantitative studies. J Chromatogr A. 1287:24–37. Guidelines prepared within the International Conference on Harmonization of Technical Requirements for the Registration of pharmaceuticals and Human Use (ICH), Validation of Analytical Procedures: text and Methodology Q2(R1) November 2005. [Internet]. [cited 2014 Jan 24]. Available from: http://www.ich.org/products/guidelines/ quality/article/quality-guidelines.html
Downloaded by [Michigan State University] at 12:28 07 March 2015
Food Additives & Contaminants: Part A Hayward DG, Wong JW, Shi F, Zhang K, Lee NS, Dibenedetto AL, Hengel MJ. 2013. Multiresidue pesticide analysis of botanic dietary supplements using salt-out acetonitrile extraction, solid-phase extraction, clean up column, and gas chromatography-triple quadrupole mass spectrometry. Anal Chem. May 7. 85:686–693. Hernández F, Pozo OJ, Sancho JV, Bijlsma L, Barreda M, Pitarch E. 2006. Multiresidue liquid chromatography tandem mass spectrometry determination of 52 non gas chromatographyamenable pesticides and metabolites in different food commodities. J Chromatogr A. 1109:242–252. Ho YM, Tsoi YK, Leung KS. 2013. Highly sensitive and selective organophosphate screening in twelve commodities of fruits, vegetables and herbal medicines by dispersive liquid-liquid microextraction. Anal Chim Acta. May. 775:58–66. Hou X, Han M, Dai X, Yang X, Yi S. 2013. A multi-residue method for the determination of 124 pesticides in rice by modified QuEChERS extraction and gas chromatography-tan dem mass spectrometry. Food Chem. Jun. 1. 138:1198–1205. ISO/IEC/17025 General requirements for the competence of testing and calibration laboratories. September 2005 Kirchner M, Matisová E, Otrekal R, Hercegová A, de Zeeuw J. 2005. Search on ruggedness of fast gas chromatography-mass spectrometry in pesticide residues analysis. J Chromatogr A. 1084:63–70. Kittlaus S, Schimanke J, Kempe G, Speer K. 2013. Development and validation of an efficient automated method for the analysis of 300 pesticides in foods using two-dimensional liquid chromatography-tandem mass spectrometry. J Chromatogr A. 29:1283–1298. Kovalczuk T, Lacina O, Jech M, Poustka J, Hajšlová J. 2008. Novel approach to fast determination of multiple pesticide residues using ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Food Addit Contam: Part A. April 25:444–457. Leandro CC, Hancock P, Fussell RJ, Keely BJ. 2007. Ultraperformance liquid chromatography for the determination of pesticide residues in foods by tandem quadrupole mass spectrometry with polarity switching. J Chromatogr A. 1144:161–169. Mastovska K, Hajslova J, Lehotay SJ. 2004. Ruggedness and other performance characteristics of low pressure gas chromatography-mass spectrometry for the fast analysis of
851
multiple pesticide residues in food crops. J Chromatogr. A. 1054:335–349. Mol HGJ, Zomer P, de Koning M. 2012. Qualitative aspects and validation of a screening method for pesticides in vegetables and fruits based on liquid chromatography coupled to full scan high resolution (Orbitrap) mass spectrometry. Anal Bioanal Chem. July 403:2891–2908. Payá P, Anastassiades M, Mack D, Sigalova I, Tasdelen B, Oliva J, Barba A. 2007. Analysis of pesticide residues using the Quick Easy Cheap Effective Rugged and Safe (QuEChERS) pesticide multiresidue method in combination with gas and liquid chromatography and tandem mass spectrometric detection. Anal Bioanal Chem. 389:1697–1714. Pizzutti IR, de Kok A, Zanella R, Adaime MB, Hiemstra M, Wickert C, Prestes OD. 2007. Method validation for the analysis of 169 pesticides in soya grain, without cleanup, by liquid chromatography-tandem mass spectrometry using positive and negative electrospray ionization. J Chromatogr A. 1142:123–136. Santilio A, Stefanelli P, Dommarco R. 2009. Fast determination of phenoxy acid herbicides in carrots and apples using liquid chromatography coupled triple quadrupole mass spectrometry. J Environ Sci. Health, Part B. 44:584–590. Sanyal D, Rani A, Alam S, Gupta R. 2011. Development, validation and uncertainty measurement of multi residue analysis of organochlorine and organophosphorus pesticides using pressurized liquid extraction and dispersive-SPE techniques. Environ Monitor Assess. 182:97–113. Stefanelli P, Barbini DA, Girolimetti S, Dommarco R. 2012. Estimation of measurement uncertainty associated to the determination of pesticide residues: a case study. J Environ Sci Health, Part B. 47:804–813. Thompson M, Ellison SLR, Wood R. 2002. Harmonized guidelines for single-laboratory validation of methods of analysis (IUPAC Technical Report). Pure Appl Chem. 74:835–855. United States Pharmacopoeia. 2006. National formulary. 24 ed. Rockville (MD): United States Pharmacopoeial Convention. Yenisoy-Karakaş S. 2006. Validation and uncertainty assessment of rapid extraction and cleanup methods for the determination of 16 organochlorine pesticide residues in vegetables. Anal Chim Acta. 571:298–307. Youden WJ, Steiner EH. 1975. Statistical manual of the association of official analytical chemists. The association of official analytical chemists. Washington (DC): AOAC; p. 33, 70, 82.
