April 2012
Food Additives & Contaminants Editorial
Brazilian Ministry of Agriculture, Livestock and Food Supply (MAPA): strategies to tackle chemical food safety issues B. Le Bizec
481
Foreword
The National Agricultural Laboratories of Brazil and the control of residues and contaminants in food
Food Additives & Contaminants
Volume 29 Number 4
482
A. de Queiroz Mauricio and E.S. Lins
Articles
Evolution of a residue laboratory network and the management tools for monitoring its performance
490
E.S. Lins, E.S. Conceic° a¬ o and A. De Q. Mauricio
L. Jank, R.B. Ho¡, P.C.Tarouco, F. Barreto andT.M. Pizzolato
High-throughput multiclass screening method for antibiotic residue analysis in meat using liquid chromatography-tandem mass spectrometry: a novel minimum sample preparation procedure M.S. Bittencourt, M.T. Martins, F.G.S. de Albuquerque, F. Barreto and R. Ho¡ M.P. Almeida, C.P. Rezende, L.F. Souza and R.B. Brito
Occurrence of antimicrobial residues in Brazilian food animals in 2008 and 2009
C.K.V. Nonaka, A.M.G. Oliveira, C.R. Paiva, M.P. Almeida, C.P. Rezende, C.G.O. Moraes, B.G. Botelho, L.F. Souza and P.G. Dias
In-house validation of PremiÕ Test, a microbiological screening test with solvent extraction, for the detection of antimicrobial
508 517 526
Optimisation and validation of a quantitative and con®rmatory LC-MS method for multi-residue analyses of -lactam and tetracycline antibiotics in bovine muscle C.P. Rezende, M.P. Almeida, R.B. Brito, C.K. Nonaka and M.O. Leite
Determination and con®rmation of chloramphenicol in honey, ®sh and prawns by liquid chromatography±tandem mass spectrometry with minimum sample preparation: validation according to 2002/657/EC Directive F. Barreto, C. Ribeiro, R.B. Ho¡ andT. Dalla Costa
Simultaneous determination of chloramphenicol and ¯orfenicol in liquid milk, milk powder and bovine muscle by LC±MS/MS D.R. Rezende, N. Fleury Filho and G.L. Rocha
Producing a sulfamethazine quality control material under the framework of ISO/CD Guide 80
541
ISSN 1944–0049
Food Additives & Contaminants . . .
PART A: CHEMISTRY
ANALYSIS
CONTROL
EXPOSURE & RISK ASSESSMENT
Ministério da Agricultura, Pecuária e Abastecimento do Brasil – Laboratórios Nacionais Agropecuários: Methods of analysis for residue and contaminants in the food chain
550 559 571
A.L. Cunha, P.F. Silva, E.A. Souza, J.R.A.M. Ju¨ nior, F.A. Santos and E.A.Vargas
Bioactivity-based screening methods for antibiotics residues: a comparative study of commercial and in-house developed kits R. Ho¡, F. Ribarcki, I. Zancanaro, L. Castellano, C. Spier, F. Barreto and S.H. Fonseca
Optimisation and validation of a quantitative and con®rmatory method for residues of macrolide antibiotics and lincomycin in kidney by liquid chromatography coupled to mass spectrometry C.P. Rezende, L.F. Souza, M.P. Almeida, P.G. Dias, M.H. Diniz and J.C. Garcia
Validation of a rapid and sensitive routine method for determination of chloramphenicol in honey by LC±MS/MS T.Taka, M.C. Baras and Z.F. Chaudhry Bet
(
TFAC-29-4.indd 1
535
April 2012
April 2012
residues in poultry muscles
C.G. Magalha¬ es, C.R. De Paiva, B.G. Botelho, A.M.G. De Oliveira, L.F. De Souza, C.V. Nonaka, K.V. Santos, L.M. Farias and M.A.R. Carvalho
Number 4
Validation of a quantitative and con®rmatory method for residue analysis of aminoglycoside antibiotics in poultry, bovine, equine and swine kidney through liquid chromatography-tandem mass spectrometry
497
Volume 29
-lactam antibiotics residues analysis in bovine milk by LC-ESI-MS/MS: a simple and fast liquid±liquid extraction method
Volume 29 Number 4
577 587 596
Continued on inside back cover)
3/15/2012 6:53:51 PM
Food Additives and Contaminants Vol. 29, No. 4, April 2012, 481
EDITORIAL Brazilian Ministry of Agriculture, Livestock and Food Supply (MAPA): strategies to tackle chemical food safety issues This special issue of Food Additives and Contaminants is dedicated to Brazil and to the strategies implemented by the Ministry of Agriculture, Livestock and Food Supply (MAPA), to tackle chemical food safety issues. This issue features a selection of papers arising mainly from work conducted within laboratories belonging to MAPA. The papers deal with the determination of chemicals, such as heavy metals, polycyclic aromatic hydrocarbons, phytosanitary products, mycotoxins, veterinary drugs or dyes that are introduced into foods either as a result of their occurrence in the environment, natural infection by fungi, or other human activities. The agricultural sector in Brazil exhibits an impressive number of agricultural establishments responding directly to the favourable balance of the national trade results (a positive balance of US$60 billion in 2009). Facilitation of the trade implies a mandatory compliance with international rules. In particular, papers presented in this special issue focus on the European regulations involving compounds/ matrices of interest, maximum residue limits, criteria for sampling, performance and validation criteria for analytical methods as well as quality assurance of the results issued by the laboratories. This context has led to the development of rapid screening methods for
ISSN 1944–0049 print/ISSN 1944–0057 online ß 2012 Taylor & Francis http://dx.doi.org/10.1080/19440049.2012.662768 http://www.tandfonline.com
various analytes based on immunochemical techniques. Furthermore, highly sophisticated multianalyte methods based on liquid chromatography coupled with high-resolution and/or multipledimension mass spectrometry have been implemented to allow identification and simultaneous determination of a wide range of residues/contaminants. Thanks to the contributors, to whom we would like to express our great gratitude, 24 papers have been published in this special issue, which evidences a real attempt to work up Brazilian food control to the highest international standards. I am also very thankful to the Editorin-Chief of the journal, John Gilbert, and Managing Editor, Victoria Gardner, for their kind support, to the skilled scientists who have been involved in the assessment of the papers, and to Dr. Gaud DervillyPinel for her help in the management of the reviewing process. Bruno Le Bizec, Prof, Dr, HDR Guest Editor LABERCA, ONIRIS Nantes, France Email:
[email protected]
Food Additives and Contaminants Vol. 29, No. 4, April 2012, 482–489
FOREWORD The National Agricultural Laboratories of Brazil and the control of residues and contaminants in food A. de Queiroz Mauricio* and E.S. Lins Ministry of Agriculture, Livestock and Food Supply of Brazil, Esplanada dos Ministe´rios, 70043-900 Brası´lia-DF, Brazil (Received 26 November 2010; final version received 13 July 2011) The laboratory activity of the Ministry of Agriculture, Livestock and Food Supply in Brazil has a history that is richer than most people are aware of. The institutions that today are known as National Agricultural Laboratory – Lanagros – were once a smaller initiative that suffered ups and downs throughout the decades. The recognition that the Lanagros have today – as reference centres with open communication channels with some of the world’s greater laboratories in residue and contaminants in food analyses – is the fruit of several years of hard work, good ideas and a strong will never to let down society. Today the Lanagros act not only by performing analyses for the monitoring and investigation programmes, but also in the research and development of analytical methods, providing technical advice on the elaboration of guidelines and normatives, international negotiation and the evaluation of other laboratories. The Lanagros work in an ISO 17025 environment, and they are now being directed and prepared to be able to respond to outbreaks and crises related to the presence of residues and contaminants in food, with the readiness, quickness and reliability that an emergency requires. Investments are allocated strategically and have been giving concrete results, all to the benefit of consumers. Keywords: animal; vegetables; fruit; cereals; residues; veterinary environmental contaminants; regulations; quality assurance
The early days initiative Moving back in a timeline towards the origins of the laboratory system of MAPA, one can identify in the early days disease diagnostics, the quality control of raw materials and both animal and plant products laboratories, which were mostly located on the farms of the Ministry of Agriculture and where part of government official controls took place. The Ministry of Agriculture was created in 1860 – at that time it was named the Ministry of Agriculture, Industry, Commerce and Public Works. At the very beginning of the twentieth century, and within the wide range of responsibilities of the Ministry of Agriculture, some facilities were created in close relation with laboratory activities, such as the Superior School of Agriculture and Veterinary Medicine, Agricultural Inspection and Defense Service, Geological and Minerological Service, and the Chemistry Institute (Santos 2006). This was an important step towards improved science education in Brazil, especially the establishment of chemistry courses linked to the Ministry’s activities, as the Ministry of Education and Health was only subsequently created in 1930. In those days the Ministry of Agriculture contained several departments of applied chemistry throughout the country: Laboratories of Agricultural and Food
residues;
pesticide
residues;
Chemistry, of special studies on rubber in the Amazon region, of sugar analysis, laboratories attached to the Fisheries Stations, to the National Museum and the Botanical Garden. In parallel, municipalities were also establishing laboratories for foodstuff control. However, for reasons that would require further investigation, these remarkable facilities were closed one after the other (Santos 2006). In 1918, the Laboratory of Defense Inspection of butter, a working station responsible for the analysis of dairy products consumed in Brazil, was transformed into the Chemistry Institute. Among the responsibilities of this new institute, the legal mandate also included:
*Corresponding author. Email:
[email protected] This paper is kindly dedicated to the one I love. ISSN 1944–0049 print/ISSN 1944–0057 online ß 2012 Taylor & Francis http://dx.doi.org/10.1080/19440049.2011.620987 http://www.tandfonline.com
drug
. Research in specialties of agriculture, industry and livestock. . Tests and chemical studies for commercial, private, state and municipality government purposes. . Teaching of chemistry aimed at capacitybuilding for technicians. . Studies on fodder from a scientific perspective. . Inspection of butter and dairy products. . Inspection of fertilisers, insecticides and fungicides.
Food Additives and Contaminants
483
Table 1. Year of foundation of the laboratories that would later be the Lanagros. Old name LARA Pedro Leopoldo, MG LARA Campinas, SP LARA Porto Alegre, RS Labortatory of Animal Diseases and anti-rabies Vaccines, PE LARA Bele´m, PA Laboratory of Vaccines for Foot and Mouth Disease and Rabies
A Presidential Decree from the same year, 1918, stated that the Chemistry Institute would also be responsible for establishing assay methods for the food laboratories (Faria et al. 2010). What is also remarkable is that the Chemistry Institute courses are considered the first major official chemistry courses available in Brazil, but they were active only until 1926. One can see an important initiative from the bureaucrats of those days in establishing an analytical capability within the ministry, with a practical focus on innovation and development. Sadly, the courses have changed. Nevertheless, the ministry eventually realised that the inspection actions would not be carried out properly without laboratory scrutiny. Focused on inspection and defence demands, and because of a need for laboratory support linked to those actions, the Ministry of Agriculture created a number of official laboratories, which would later become the National Agricultural Laboratories – Lanagros. In 1976, a Presidential Decree created a Reference Laboratory Network; and in 1978 a Ministerial Ordinance transformed the existing structures into Reference Laboratories, each responsible for covering a region of Brazil (BRASIL 1978). The Regional Laboratories of Animal Reference (LARA) were established at different times (Table 1). They later merged with the Regional Laboratories of Plant Reference (LARV) to become one entity.
The agricultural laboratory network milestones Over more than 30 years the activity of these laboratories was considered a support of minor importance to the main agricultural inspection operation, which was regarded as an activity isolated from those inspection activities. However, along with an increase in the country’s production and retail volumes, as well as the constant upgrades observed in electronics, computers, nanotechnology, pharmacy, genetics, agricultural practices and, above all, analytical and instrumental chemistry, a clear move can be noticed towards the detection and identification of substances at a frequency and at concentrations unthinkable in the early days.
Year of foundation
Actual name
1983 1979 1950s 1947 1949 1948
Lanagro-MG Lanagro-SP Lanagro-RS Lanagro-PE Lanagro-PA Lanagro-GO
The significant worldwide expansion in food demand and commercialisation has posed new challenges to governments, precipitated by the current large scale of food production. On a global perspective, world trade increased by 17.6% in 2008 at US$12.6 trillion, US$853.2 billion of it solely related to food and agricultural products. In this scenario, the Brazilian export of such commodities began to expand in the worldwide market from 2003 onwards (BRASIL 2010). In 2002 Brazil had 4.6% of the agricultural world market; since then this figure has increased by 2.2 percentage points, achieving 6.8% of the total global market at present (BRASIL 2010). This increase is related to the fact that special efforts were directed towards the improvement of Brazil’s capacity regarding the production and export of goods. As an example, in 2009 Brazil exported US$54.8 billion in products originating from its agribusiness, the second highest value since 1997 (BRASIL 2010). Considering Brazil’s export basket, the participation of agricultural products in total Brazilian exports shifted from 23.9% in 2000 to 35.8% in 2009. This increase is directly related to the quantities of goods such as food produced and exported from 2000 to 2009, despite the oscillations in prices observed in the same period due to global crises. The amount of Brazilian agricultural products exported increased by 164.8% between 2000 and 2009 (BRASIL 2010), which explains the expansion of Brazilian participation in the global market. In a more detailed and particular analysis, the expansion of the quantities exported is linked to the increase of the exportable surplus which may be associated with the difference between the rate of population and production growth. As an example, while grain production in Brazil increased by 77.0% over the last decade, population growth was only 13.7%, which explains the generation of a surplus. The five main contributing areas to this scenario were soya complex, meat and meat products, sugar and alcohol complex, coffee and tobacco, jointly responsible for US$46.1 billion of the total US$54.8 billion exported, i.e. 84.0% of total Brazilian exports (BRASIL 2010). Together with recent improvements in technology, this remarkable increase in the volumes of food production and consumption demanded exceptional performance
484
A. de Queiroz Mauricio and E.S. Lins
Figure 1. Geographical distribution of the six Lanagros and their SLAV – Advanced Laboratory Units.
from the laboratories and principally required that they were completely capable of keeping up with this evolution, at the same time as regulations related to food safety were being tightened by governments all over the world. It also required a new integrated approach to governmental control, in which the laboratory activities were an inseparable part of further inspection operations, the laboratories acting directly on the verification and validation of the production and processing systems of agriculture by giving analytical insights, rather than by simply providing test reports. These pressures led the Brazilian government recently to carry out a reorganisation of MAPA’s laboratory system, with the objective of updating and improving the policies and analytical activities related to plant and animal defence and health. Thus, a national laboratory network was established, with the responsibility of performing studies and assays in order to assess the conformity of the agricultural inputs and food produced domestically, especially concerning official control of residues and contaminants. Nowadays that network comprises the six National Agricultural Laboratories – formally created in 2005 by Presidential Decree No 5.351 and designated in
2006 as the official laboratories by Presidential Decree No 5741 (BRASIL 2005, 2006), which are the official laboratories of the Ministry of Agriculture, Livestock and Food Supply, with the responsibility of serving as the national reference in laboratory activities regarding agricultural health and defence as legal mandate. Today the legal mandate of the Lanagros includes activities such as: . Official analyses. . Assays for inspection, monitoring and other legal purposes. . Laboratory audits. . Research and development of analytical methods. . Elaboration and review of legislation and technical guidelines. . Participation in international negotiations. The six Lanagros are located in the five geographical regions of Brazil (with two in the southeast region, the country’s main industrial and production site), and they therefore act as centres of production and diffusion of analytical knowledge and laboratory policies in each of those regions as depicted in Figure 1. The Lanagros were conceived as multidisciplinary centres and parts of
GO
PE
RS
Lanagro 1
3 1 0 2 0 0 4 0
1 1
0 0 0 1
Pharmacists Veterinarians
Agronomists
Biologists/engineers
Other areas
Undergrad/support Chemists
Pharmacists
Veterinarians
Agronomists
Biologists/engineers Other areas Undergrad/support Chemists
Number of permanent technicians
Chemists
Areas of expertise
Staff
0 0 0 1
0
1
0
4 0
0
1
0
1 0
3
Number of technicians under contract
Table 2. Summary of the analytical capability of Lanagros.
17
8
18
Number of associate researchers Main equipments in the operation
(continued )
AAS equipped with background correction capability and data-handling system, PerkinElmer Model AAnalyst 100; FIAS Flow injection analysis system working in the metal hydride mode, PerkinElmer Model FIAS 400; autosampler PerkinElmer Model AS-90 GTA graphite furnace atomic absorption equipped with Zeeman background correction Varian Model 240Z; autosample Varian Model 120; Quick Trace M-6100 Mercury Analyzer equipped with peristaltic pump; ASX-400 autosampler, system working in the metal hydride mode Microwave application for acid digestion, equipped with ramp to temperature as 1600 W, CEM, model MARS Xpress Varian HPLC System equipped with: ProStar 410TM HPLC AutoSampler with cooling option and standard sample tray: 84 1.5 ml vials with 3 10 ml vials; ProStar 363 fluorescence detector, with continuous Xenon lamp, wavelength range 200–731 nm for excitation and 200–900 nm for emission; ProStar 335TM Diode Array Detector, with wavelength range 190–950 nm, light source: D2 and quartz halogen; ProStar 240TM HPLC solvent delivery modules, ternary gradient pump; ProStar 210 HPLC solvent delivery modules, isocratic pump; ProStar 500TM column valve module, can accommodate up to six analytical or two semi-prep columns and two individually controlled heaters; control and data handling GalaxieTM Chromatography Data System, Galaxie Workstation
LC-MS/MS system, Applied Biosystems, API 5000, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI; LC-MS/MS system, Applied Biosystems, API 5000, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI; LC-MS/MS system, Waters, Quattro Micro, mass analyser triple quadrupole with ionisation source ESI/APCI HPLC system, Shimadzu, with fluorescence, UV-VIS and diode array detectors GC system, Thermo, Trace, with electron capture detector (ECD); hydride generation atomic absorption spectrometry (CV AAS), Analisty 200 with FIAS/100 (PerkinElmer, USA) Cold vapour atomic absorption spectrometry (CV AAS), Model FIMSÕ 400 Mercury Analysis System with AS91 Autosampler (PerkinElmer, USA) Graphite furnace atomic absorption spectrometry (GF AAS), Analisty 600 with AS 800 Autosampler (PerkinElmer, USA) Graphite furnace atomic absorption spectrometry (GF AAS), Model 4110ZL with AS 72 Autosampler (PerkinElmer, USA)
Analytical capabilities
Food Additives and Contaminants 485
MG
PA
SP
Lanagro
Staff
5 1 0 0 0 2
Pharmacists
Veterinarians
Agronomists
Biologists/engineers
Other areas
Undergrad/support
0 0 0 0 0 4
0 0 4 0
Biologists/engineers Other areas Undergrad/support Chemists
Veterinarians Agronomists Biologists/engineers Other areas Undergrad/support Chemists
2
Agronomists
2 1
0
Veterinarians
Chemists Pharmacists
0
Number of permanent technicians
Pharmacists
Areas of expertise
Table 2. Continued.
1 0 0 0 2 0
0 0
4
0
0
0
0
0
0 0 0 3
0
0
0
Number of technicians under contract
32
18
13
Number of associate researchers Main equipments in the operation
LC-MS/MS System, Applied Biosystems, API 5500, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI00 ; GFAA System, PerkinElmer, AA800; mercury analyser, Milestone, DMA-80 Mercury analyser (CV-AAS), CETAC, M-6100; HPLC-UV-fluorescence, Shimadzu Microwave, Anton-Paar, MV3000; microwave, Milestone, ultraclave UV-VIS, Varian, Cary 50 Conc; high shear mixer, Silverson, DX 60-2 unt Mill, Retsch, SK100
GF AAS and HG AAS systems, PerkinElmer, AAnalyst 800, atomic absorption spectrometer with graphite furnace and hydride generation techniques; GF AAS system, PerkinElmer, AAnalyst 600, atomic absorption spectrometer with graphite furnace technique SS TDA AAS system, Milestone, DMA-80, solid sampling thermal decomposition amalgamation atomic absorption spectrometer GC/MS-Shimadzu QP-2010, gas chromatograph with mass spectrometer detector; GC/ECD Thermo Finnigan, gas chromatograph with electron-capture detector GC-HRMS system with Kit autosampler model Al 3000, Mat 95XP, Thermo Finnigan, highresolution mass analyser LC-MS/MS system, Applied Biosystems, API 5000, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI00 LC-MS/MS system, Thermo Finnigan, Quantum Ultra, mass analyser triple quadrupole with ionisation source ESI LC-MS/MS system, Thermo Finnigan, TSQ Quantum Access, mass analyser triple quadrupole with ionisation source ESI
LC-MS/MS system, Applied Biosystems, API 5000, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI LC-MS/MS system, Waters, Quattro Premier, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI GC-MS system, Thermo Finnigan, mass analyser single quadrupole with ionisation source EI/CI Mercury analyser, Quick Trace M6100, atomic absorption spectrometer, equipped with autosampler ASX-400 Atomic absorption-graphite furnace-hydride generation – PerkinElmer-A600-FIAS100 Microwave sample preparation, ANTON PAAR, Multiwave 3000
Analytical capabilities
486 A. de Queiroz Mauricio and E.S. Lins
6
3
0
4
0
6
Pharmacists
Veterinarians
Agronomists
Biologists/engineers
Other areas
Undergrad/support
0
0
0
0
0
0
HRGC-MS/MS system, Agilent, 7000A model, mass analyser triple quadrupole with electronic impact and chemical ionisation sources HRGC/EI/CI; HRMS-HRGC system, high-resolution mass spectrometer system, Waters, AutoSpec Premier model, magnetic sector analyser with electronic impact, coupled with gas chromatograph Agilent, 6890N model with PTV injector, HRMS/HRGC/EI HRGC-MS system, Thermo Corporation, DSQ model coupled with gas chromatograph Thermo/Trace GC Ultra model, HRGC/MS/EI; HRGC-MS/MS system, Thermo Corporation, Polaris Q model coupled with gas chromatograph Thermo/Focus GC model, impact electronic and chemical ionisation sources, HRGC-MS/MS/EI/CI (ion trap); HRGC-MS/MS system, Thermo Corporation, Polaris Q model coupled with gas chromatograph Thermo/Focus GC model, impact electronic and chemical ionisation sources, HRGC-MS/MS/EI/CI (ion trap); DMA, Milestone, DMA-80, direct mercury analyser Gas chromatograph Thermo/Trace GC Ultra model with ECD and FID detectors, HRGC/ ECD/FID; LC-MS/MS system, Applied Biosystems, API 5500, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI; LC-MS/MS system, Waters, quadrupole time of flight (G1) – TOF; FIAS, PerkinElmer, FIAS-400, flow injection analysis system; microwave digestion system, Anton-Paar, Multiwave 3000; acid purification system, Milestone, DuoPUR ICP-MS, Varian, 820-MS, inductively coupled plasma mass spectroscopy; GF AAS, PerkinElmer, AA-600, graphite furnace atomic absorption spectrometry; FAAS, PerkinElmer, AA-400, flame atomic absorption spectrometry; FAAS, PerkinElmer, AA-100, flame atomic absorption spectrometry; Milli-Q Advantage, Millipore, A10 Element, water purification system unit; toxic organic sampler used for sampling of dioxins, PCBs, pesticides and PAHs in air. Amotox – Energe´tica Air Quality RRLC Agilent - MS/MS system, Applied Biosystems, API 5000, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI; RRLC Agilent - MS/MS system, Applied Biosystems, API 5000, mass analyser triple quadrupole with ionisation source ESI/ APCI/APPI; UFLC Shimadzu-MS/MS system, Applied Biosystems, API 4000 QTrap, mass analyser triple quadrupole with ionisation source ESI/APCI/APPI; UPLC AcquityMS/MS system, Waters, Quattro Premier XE, mass analyser triple quadrupole; HPLC-MS/ MS system, Waters, Quattro Premier XE,mass analyser triple quadrupole HPLC – Shimadzu, SCL-10VP, with fluorescence and UV detector and positive column derivatisation using Kobra cell; HPLC – Shimadzu, SCL – 10Avp, with fluorescence detector and iodine positive column derivatisation; HPLC – Shimadzu, SIL-HTC Prominence, with fluorescence detector and positive column derivatisation using Kobra cell; HPLC – Shimadzu, SBM – 20A Prominence, with fluorescence, UV and DAD detector and positive column derivatisation system HPLC, Marca: Shimadzu, RP 006.212, RI Lacqsa 511; LC-MS/MS system, Varian 1200L, Quadrupole; GC-MS/MS system, Agilent Technologies, 7890A model, mass analyser triple quadrupole with electronic impact and chemical ionisation sources HRGC/EI/CI; 03 Automated sample processor system (ASPECXl); Espectrophotometer, Shimadzu, UV-1601 PC
Food Additives and Contaminants 487
488
A. de Queiroz Mauricio and E.S. Lins
an unified system for the attention of agricultural health, which is coordinated by a central instance in Brası´ lia that brings various agents together involved in concerted actions. In direct communication with the other inspection bodies of MAPA and states, they act by spreading analytical capabilities in the areas where inspections are present, providing the required laboratorial services with agility and efficacy. Additionally, a research, development and innovation branch was introduced in the Lanagros with the objective of conducting exploratory studies and in anticipation of emerging food-borne risks in a proactive approach focused on the speed of response and emergency preparedness. This scenario demands that the Lanagros also expand their scopes. Therefore, there are several new methods, especially multi-residues, being developed and implemented in the analytical routine, and counting on the expertise of both old and new staff who develop research activities specifically on this matter in their workplans.
Today Today the Lanagros act in the following areas in the spectrum of action of the Ministry of Agriculture: . Residues and contaminants in food. . Agrochemicals including pesticides. . Physical chemistry of products of animal origin and water. . Physical chemistry classification of plant products. . Physical chemistry of feed. . Physical chemistry of beverages and vinegar. . Animal diagnoses. . Plant diagnoses. . Physical chemistry of fertilisers, soil corrective, substrates and correlates. . Inoculants and analogous compounds. . Microbiology of food and water. . Biotechnology and genetically modified organisms (GMOs). . Milk quality. . Genetic identification and animal semen. . Control of veterinary products. . Seeds and seedlings. The residues and contaminants area, which is a strategic front of the Lanagro Network, and on which the Agricultural Defense Secretariat relies substantially, is responsible for the execution of the Brazilian National Residue and Contaminants Control Plan (PNCRC). In this area the scope of the Lanagros is divided into various fronts: veterinary drug residues; organic contaminants and inorganic contaminants in products of animal origin; mycotoxins; and pesticide
residues in products of plant origin. The PNCRC comprises almost 20,000 samples collected throughout the year from producers in every state of the country. The Lanagros has state-of-the-art analytical chemistry equipment, with techniques based mainly on liquid and gas chromatography coupled with mass spectrometry (triple quadrupole and time-of-flight), and gas chromatography coupled with high-resolution mass spectrometry, atomic absorption spectroscopy, and inductively coupled plasma spectroscopy, besides microbiological screening methods. The laboratory staff are constantly seeking development and innovation in terms of analytical methodology, aiming to optimise the time of analysis, capacity, resources and quality. All the Lanagros’ analysis methods in the residues and contaminants areas follow rigorous validation and internal quality control so that the issued results are substantiated. An executive summary of the analytical capability of the Lanagros including the main equipment and staff available is shown in Table 2. There has been also a great effort in opening technical cooperation with reference centres around the world, as a valuable experience for the interchange of advances in techniques allied to the developments and innovation focused on continuous improvement and faster response to emerging risks in food production and consumption. Examples of these cooperations currently in place are projects such as ‘Project UE/Mercosur/SPS ALA/2005/17887’ aimed at cooperation for the harmonisation of veterinary, phytosanitary, and food safety standards and procedures; and ‘Project UE/Brasil ALA/2004/006-189’ aimed at the internationalisation of Brazilian enterprises, as well as bilateral cooperation for capacity-building with international reference centres such as CFIA in Canada, Wageningen-UR and RIKILT in the Netherlands (LNV-BOCI Project 2010), SARAF/ LABERCA in France, and the Japan International Cooperation Agency (JICA). Research initiatives such as the ‘MycoRed Project – IFA-Tulln Center for Analytical Chemistry, Austria’ and ‘Aflatoxins in Brazil Nuts and their Shells, FSA call RRD 31 – FERA, UK’ are also being carried out with the active participation of the Lanagros. In a concerted action by both MAPA and the Ministry of Science and Technology of Brazil (MCT), a project was launched for the allocation of specialists through scholarships for MSc and PhD degrees, directed towards the improvement of the residues and contaminants area at the Lanagros. The project is being conducted together with the The National Council for Scientific and Technological Development (CNPq), which is a Federal Agency linked to the MCT dedicated to the promotion of scientific and technological research and to capacitybuilding of human resources for research in the country (its history is directly linked to the scientific
Food Additives and Contaminants and technological development of Brazil). According to the rules of the CNPq, these scholarships are classified as ‘Scholarships for Technological Development and Innovative Extension’ which are destined by the CNPq for the use of specialists and scientists focused on: (1) the execution of applied research projects; (2) the execution of technological development projects; and (3) activities of innovative extension and transfer of technology. The specialists selected are mainly responsible for activities such as the development and validation of analytical methods for substances/matrices currently lacking; the implementation and optimisation of instrumental laboratories; the transfer of knowledge and technology to Lanagros’ staff; the solution of QA/QC issues focusing on accreditation; the production of reference materials; undertaking engineering and architecture technical projects; the development and use of statistical tools applied to laboratory work; and improvements to the procedures related to the management of residue and contaminants laboratories. Since 2008, 109 scholarships for MSc and PhD degrees were made available for the residue and contaminants areas at the Lanagros, drawing together national and international experts and scientists from areas such as analytical chemistry, engineering, pharmacology, veterinary and agronomy sciences, mathematics and statistics, biology, quality assurance, architecture, administration, etc. The ultimate purpose of this project is to establish the Lanagros as high-level science-oriented institutions able to produce advanced technological solutions and respected by the scientific community and trusted by society. Lanagro Network’s Vision is to be recognised as a world reference point in agriculture and livestock laboratory services, capable of providing quick responses of high quality and scientific excellence, and striving for innovation, rigour and efficiency in the delivery. For this, investments were prioritised into three main areas: capacity-building, infrastructure and quality management tools. In 2009/2010, a total budget of more than US$5.5 million was allocated to different projects according to strategic needs aimed at
489
actions divided into these three main areas (including routine costs). Particularly in the case of infrastructure, the budget was directed to diagnostic studies and surveys, as well as to minor refurbishments and the funding of technical engineering and architecture projects (resources for buildings and major refurbishments need to be foreseen for the next biennium). The Lanagros recognise that it is necessary to be able to respond quickly and efficiently to outbreaks and crises. Significant effort and resources have been directed to make the Lanagros able to foresee public health concerns, working together with other departments in MAPA, not only to protect Brazilian consumers, but also to continue developing trade. It is time to be one step ahead. As in any timeline, ‘This is not the end. It’s not even the end’s beginning. But it might be the beginning’s end . . . .’
References BRASIL, Ministe´rio da Agricultura. 1978. Portaria Ministerial no. 904, de 29 de Setembro de 1978, Dia´rio Oficial da Unia˜o. Brası´ lia (Brazil). BRASIL, Ministe´rio da Agricultura, Pecua´ria e Abastecimento. 2005. Decreto Presidencial no. 5351, de 21 de janeiro de 2005, Dia´rio Oficial da Unia˜o. Brası´ lia (Brazil). BRASIL, Ministe´rio da Agricultura, Pecua´ria e Abastecimento. 2006. Decreto Presidencial no. 5741, de 30 de marc¸o de 2006, Dia´rio Oficial da Unia˜o. Brası´ lia (Brazil). BRASIL, Ministe´rio da Agricultura, Pecua´ria e Abastecimento. 2010. Intercaˆmbio Comercial do Agronego´cio: principais mercados de destino. Brası´ lia (Brazil). Faria LR, Caˆmara BP, Fonseca MR. 2010. Instituto de Quı´ mica. In: Diciona´rio Histo´rico-Biogra´fico das Cieˆncias da Sau´de no Brasil (1832–1930); [cited 2010 Nov 25]. Available from: http://www.dichistoriasaude. coc.fiocruz.br/ Santos NP. 2006. Fac¸amos Quı´ micos – a ‘certida˜o de nascimento’ dos cursos de quı´ mica de nı´ vel superior no Brasil. Quı´ m Nova. 29(3):621–626.
Food Additives and Contaminants Vol. 29, No. 4, April 2012, 490–496
Evolution of a residue laboratory network and the management tools for monitoring its performance E.S. Lins*, E.S. Conceic¸a˜o and A. De Q. Mauricio Ministry of Agriculture, Livestock and Food Supply of Brazil, Esplanada dos Ministe´rios, Annex B, Room 433 – Zip code: 70043-900, Brası´lia-DF, Brazil (Received 26 November 2010; final version received 16 December 2011) Since 2005 the National Residue & Contaminants Control Plan (NRCCP) in Brazil has been considerably enhanced, increasing the number of samples, substances and species monitored, and also the analytical detection capability. The Brazilian laboratory network was forced to improve its quality standards in order to comply with the NRCP’s own evolution. Many aspects such as the limits of quantification (LOQs), the quality management systems within the laboratories and appropriate method validation are in continuous improvement, generating new scenarios and demands. Thus, efficient management mechanisms for monitoring network performance and its adherence to the established goals and guidelines are required. Performance indicators associated to computerised information systems arise as a powerful tool to monitor the laboratories’ activity, making use of different parameters to describe this activity on a day-to-day basis. One of these parameters is related to turnaround times, and this factor is highly affected by the way each laboratory organises its management system, as well as the regulatory requirements. In this paper a global view is presented of the turnaround times related to the type of analysis, laboratory, number of samples per year, type of matrix, country region and period of the year, all these data being collected from a computerised system called SISRES. This information gives a solid background to management measures aiming at the improvement of the service offered by the laboratory network. Keywords: regulations; AAS; chromatography – GC/MS; chromatography – LC/MS; heavy metals; mycotoxins; pesticide residues; veterinary drug residues; animal products – meat; vegetables; fish and fish products
Introduction A major part of the Brazilian National Residue & Contaminants Control Plan (NRCCP) is related to laboratory assays. In order to verify and monitor consumer exposure to residues and contaminants, samples are collected by official inspectors according to an annual sampling plan and sent for analyses in a laboratory network, coordinated by the Ministry of Agriculture (BRASIL 2008). Figures 1a and 1b show the size of the Brazilian NRCCP in terms of analytical programme, and the distribution among the laboratories that are part of the network, here referred as LANAGROs (official laboratories) and authorised laboratories (both private and public laboratories), in 2007 and 2008 respectively (Mauricio et al. 2009). Region Southeast holds the largest number of laboratories because it has concentrated in it a large number of the industrial establishments in the food chain (especially for swine and poultry) when compared with the other regions; and consumers; as well as
*Corresponding author. Email:
[email protected] ISSN 1944–0049 print/ISSN 1944–0057 online ß 2012 Taylor & Francis http://dx.doi.org/10.1080/19440049.2011.653988 http://www.tandfonline.com
high industrial levels for laboratorial resources and inputs. In order to cover better areas in the Brazilian territories, the Ministry of Agriculture has taken two important initiatives, which are: the provision of a second network composed by laboratories from universities and governmental research centres intended to merge within the actual network when technicaly prepared; and the development of collaborative centres focused on residues and contaminants. Both projects are held in close cooperation between the Ministry of Agriculture and the Ministry of Science and Technology, and the centres receive resources and financial support with the clear objective to develop and enhance the analysis of residues and contaminants in food. A correlation can be observed in Figures 2b, 3a and 3b, since the major part of the samples originate from the South and Southeast regions in which are located most of the establishments for swine and poultry,
Food Additives and Contaminants
491
Figure 1. Number of samples destined for each laboratory participating on the National Plan.
Figure 2. (A) Geographical distribution of the laboratories. (B) Number of samples per region.
which in turn represent more than 50% of the NRCP samples. The sampling programme is based on Codex Alimentarius guidelines, and is designed, monitored and statistically reviewed every year by the Coordination of Control of Residue and Contaminants within the Ministry of Agriculture. It can be seen from Figure 4 that the distribution of samples throughout the year was not uniform, with periods of overload of operational capacity of the laboratories, especially by the end of the year and the end of the plan. However, in order to avoid an overflow of samples coming into the laboratories in a concentrated period, the managers at the central level strengthened the controls of the weekly ‘draw and collect’ of samples by verifying the number of samples actually tested, the number of rejections, and the number of samples needed for the completion of the programme. In this way the routine was constantly assessed and eventual problems in the network capability were quickly identified and
alternate routes taken to keep up with the NRCP schedule.
Monitoring laboratory performance Between 2007 and 2009, some measures were taken to discipline the laboratories’ procedures specifically for NRCCP samples. One of those measures was the Procedural Manual for Laboratories, issued at the beginning of 2008, and which dictates exactly how a laboratory should routinely behave. The present work shows the efficacy of these measures, reflected in the improvement of the whole laboratory network performance. The performance of the laboratories was assessed with regards to their turnaround time, accreditation status and performance in PT programmes. Taking into account that the scopes hold reasonable differences, some variables were considered, such as the number of samples received yearly, the number of
492
E.S. Lins et al.
Figure 3. (A) Number of samples taken for each species. (B) Profile of the distribution of species sampled in each region.
Figure 4. Seasonal distribution of the amount of samples.
Food Additives and Contaminants methods running in the laboratory, and the availability of PT providers. All these data were gathered and assessed with a computerised system called SISRES. The system comprises the steps of the National Plan, from the selection of the sampling site, species, matrices and analytes to test for, to the reception of the samples by the laboratories and the report of results into a database. It links the inspectors in the field, the laboratories and managers at the headquarters, all with different privileges of access to specific information on their user profiles. The database allows the managers of the NRCCP to verify the number of samples collected, the number of samples discarded for different reasons, the turnaround times of the laboratories and the results of the analyses. A computerised system for the laboratory information management is under development; it will hold data from all laboratories integrated in the network, giving the network manager tools to assess laboratories performance in real time. Not only that, this system is being designed also to keep complete validation data for each laboratory. The turnaround time (TRT) comprises procedures for sample reception, the analysis per se, the update of SISRES with the launch of results and report of results on the certificate of analysis. Following the seasonal sample distribution, the turnaround times may vary proportionaly. It can be noted that a large concentration of samples arrived in the laboratories by the end of 2007, taking them more time for analysis. This tendency was observed in 2008, but to a lesser extent.
493
Taking into account all kinds of analyses of NRCCP, it was found that the global average decreased from 17.6 in 2007 to 12.3 in 2008. At the beginning of 2008, CGAL took some measures in order to coordinate better the laboratory network and thus improve the quality of the service provided to the inspection body. These measures comprised the issuing of a Procedural Manual containing guidelines from the reception, through the analysis and to the report of results, followed by a policy of more assessment on their compliance to normatives and deadlines observed by them. This led to a considerable improvement in reliability of results as well as generally lower turnaround times, taking into account different kinds of analysis performed in different laboratories. Figure 5 depicts the turnaround times per laboratory distributed in percentiles; Table 1 shows the mean turnaround times of those laboratories. One can observe a considerable improvement in the number of analyses carried out within the turnaround time recommended, i.e., 15 working days. The aim was to achieve 100% of analyses conducted in this period for the whole laboratory network, and mostly they were close to that goal, with the solely exception of one laboratory which is not part of the network since 2008, and one of the Lanagros, which in this case received a large number of samples originally destined for other laboratories (data were not considered in this case). Authorised laboratories 2 and 3 are also out of the network, so their data are not included.
Figure 5. Comparison between average turnaround times in 2007 and 2008 for all laboratories against the 15 working days recommended TRT.
494
E.S. Lins et al.
Table 1. Mean turnaround times (working days) of the laboratory network in 2007 and 2008.
Authorised Authorised Authorised Lanagro A Lanagro B Lanagro C Authorised Authorised Authorised Authorised
laboratory 1 laboratory 2 laboratory 5
laboratory laboratory laboratory laboratory
6 7 8 9
2007
2008
5 11 26 13 24 13 35 42 5 13
7 19 7 10 22 9 21 8 7 8
To date all the 31 analytical units in the Brazilian laboratory network that perform the analysis of residue and contaminants under the NRCCP are accredited to ISO 17025 by Inmetro, except for two units in Lanagro-PA and one in Lanagro-GO, which already have this quality system running and have been already evaluated primarily by Inmetro and are awaiting the issuing of the certificate. Table 2 shows only a few examples of participation in the PT rounds and the results achieved by the Brazilian network laboratories, either official or authorised. Acting as a reference for the network, the Lanagros have some duties and responsibilities for the performance of the network as a whole. Taking that into account, CGAL at the central level introduced a
Table 2. Examples of recent PT results of the National Agricultural Laboratory Network. Laboratory
Year
Lanagro-RS
2011
Provider CFIA – Canada
Assay Endectocides in swine liver
Four samples containing: Doramectin
Ivermectin
Lanagro-MG
2011
CFIA – Canada
Tetracyclines in Horse kidney
Four samples containing: Chlortetracycline
Oxytetracycline
Tetracycline
Lanagro-GO (assay not in PNCRC 2011 routine)
2011
CFIA – Canada
Tetracyclines in horse kidney
Four samples containing: Chlortetracycline
Oxytetracycline
Tetracycline
z-score Eight satisfactory z-score out of eight tests 0.91 –0.22 0.00 0.00 0.00 1.14 0.00 1.65 Twelve satisfactory z-score out of 12 tests –1.31 –1.55 –1.47 –0.24 –1.41 –1.40 0.00 0.00 –1.06 –1.05 –0.83 –0.55 Eleven satisfactory z-score out of 12 tests 1.57 2.22 1.32 0.60 1.10 1.75 0.00 0.00 1.67 1.57 1.24 1.00 (continued )
Food Additives and Contaminants
495
Table 2. Continued. Laboratory
Year
Lanagro-SP
2010
Provider CFIA – Canada
Assay DES and zeranol in swine liver
Four samples containing: DES
Zeranol
Lanagro-MG
2010
CFIA – Canada
Sulfonamides in swine muscle
Four samples containing: Sulfadiazine
Sulfadimethoxine
Sulfamethazine
project to create in the Lanagros the capacity to produce reference materials and to provide PT rounds nationally and internationally. The first great outcome of that project was a ring test offered to the whole network and also to the Inter-American Network of Food Analysis Laboratories (INFAL). In total, four assays were offered (aflatoxins in Brazil nut, avermectins in bovine muscle, sulfonamides in pig liver, and inorganic contaminants in pig kidney), and 61 laboratories from 19 countries took part in the exercise. In 2012, the first Workshop of Residue and Contaminants in Food will take place, in which the Lanagros specialists will give lectures and present to the authorised laboratories the method validation criteria, QA/QC tools and others aspects that must be considered for network performance and the achievement of analytical excellence.
Conclusions The present work demonstrates that monitoring tools based on computerised systems are extremely useful to provide guidance for management initiatives that should be taken in order to optimise the operation of the NRCCP. From these data it can be concluded that some managerial actions have to be considered in order to improve the working conditions of the laboratories
z-score Eight satisfactory z-score out of eight tests 0.00 0.00 –1.84 –1.72 –0.99 –0.71 0.00 0.00 Eight satisfactory z-score out of eight tests 0.35 0.00 0.36 0.00 0.00 0.27 0.00 0.83 1.57 0.00 0.45 0.00
and the efficacy of the NRCCP itself. Conclusions given and points that lead to follow-up directrices are as follows: . By increasing the number of laboratories and spreading them along the five regions, taking into account the kind of analysis demanded for those regions and their herds, as well as equalising the seasonal sampling, it may lead to a better profile in laboratory performance in terms of turnaround times and efficiency. The concentration of analyses in a few laboratories should be avoided. . By managing and equalising the sample collect through the months and weeks, the routine of the laboratories can be normalised and allowed to develop a preparedness for events by diminishing ‘bottlenecks’ in sample reception. . Sample reception should have automated procedures as far as possible, giving the laboratories the possibility to be prepared to absorb the differences in the numbers of samples throughout the seasons. . The capacity of the laboratories located in the lower regions (South and Southeast) should be improved in order to respond to regional production demand. . Screening methods should be placed whenever possible.
496
E.S. Lins et al. . National initiatives, such as PT provision, workshops and the improvement of the reference laboratories, should be encouraged. . Performance parameters for the network should be clearly defined and thoroughly observed to allow corrective actions. . Lastly, enforcement of the regulatory guidelines and timelines should be kept and assessed regularly in order to improve the service.
References BRASIL, Ministe´rio da Agricultura, Pecua´ria e Abastecimento, Secretaria de Defesa Agropecua´ria. 2008. Instruc¸a˜o Normativa No 10/2008, Dia´rio Oficial da Unia˜o. Mauricio AQ, Lins ES, Alvarenga MB. 2009. A national residue control plan from the analytical perspective. Anal Chim Acta. 637:333–336.
Food Additives and Contaminants Vol. 29, No. 4, April 2012, 497–507
b-lactam antibiotics residues analysis in bovine milk by LC-ESI-MS/MS: a simple and fast liquid–liquid extraction method L. Jankabc*, R.B. Hoffabc, P.C. Taroucoab, F. Barretoabd and T.M. Pizzolatoc a
Laborato´rio Nacional Agropecua´rio – Lanagro/RS, Porto Alegre, RS, Brazil; bMiniste´rio da Agricultura, Pecua´ria e Abastecimento, Brazil; cInstituto de Quı´mica, Universidade Federal do Rio Grande do Sul – UFRGS, Porto Alegre, RS, Brazil; d Programa de Po´s-graduac¸a˜o em Cieˆncias Farmaceˆuticas, Universidade Federal do Rio Grande do Sul – UFRGS, Porto Alegre, RS, Brazil (Received 24 November 2010; final version received 28 June 2011) This study presents the development and validation of a simple method for the detection and quantification of six -lactam antibiotics residues (ceftiofur, penicillin G, penicillin V, oxacillin, cloxacillin and dicloxacillin) in bovine milk using a fast liquid–liquid extraction (LLE) for sample preparation, followed by liquid chromatographyelectrospray-tandem mass spectrometry (LC-MS/MS). LLE consisted of the addition of acetonitrile to the sample, followed by addition of sodium chloride, centrifugation and direct injection of an aliquot into the LCMS/MS system. Separation was performed in a C18 column, using acetonitrile and water, both with 0.1% of formic acid, as mobile phase. Method validation was performed according to the criteria of Commission Decision 2002/657/EC. Limits of detection ranged from 0.4 (penicillin G and penicillin V) to 10.0 ng ml1 (ceftiofur), and linearity was achieved. The decision limit (CC), detection capability (CC), accuracy, inter- and intra-day repeatability of the method are reported. Keywords: chromatography – LC/MS; method validation; regulations; veterinary drug residues; veterinary drug residues – antibiotics; milk
Introduction With livestock development, milk from animal origin, mainly bovine milk, began to be produced for human consumption. Bovine milk is a rich source of important nutrients and is present in the human diet and derived processed foods. The number of dairy products derived from bovine milk has increased significantly in the last decades, which is 95% of the total dairy products (Michaelidou 2008), and they contribute substantially for the increase in demand. Dairy products contain a significant number of nutrients essential to growth and a healthy life. Industrialized products can increase this nutritive value by adding vitamins, minerals and other substances. Brazil is one of the biggest milk producers of the world; in 2009, its production was over 25 billion litres, an increase of around 2% in relation to 2008. It is estimated that milk’s production could potentially increase by 2.75% per year. This corresponds to an output of approximately 37 billion litres by the end of the projection, in 2019 (Ministe´rio da Agricultura, Pecua´ria e Abastecimento (MAPA) 2009). Milk agribusiness occupies a prominent position in the Brazilian economy, with great expectations for continually growing productivity for the next decade. *Corresponding author. Email:
[email protected] ISSN 1944–0049 print/ISSN 1944–0057 online ß 2012 Taylor & Francis http://dx.doi.org/10.1080/19440049.2011.604044 http://www.tandfonline.com
Milk quality is related to health, the management of animals and equipment during milking, the presence of microorganisms, drug residues and odours. Antibiotic residues present in milk are the result of the application of veterinary drugs, such as -lactams, for the prevention or treatment of diseases, especially infection of mammary gland and reproductive diseases. The presence of these substances in levels above the maximum residue level (MRL) renders milk unusable in dairy plants, as it makes the product unsuitable for use in industry and human consumption since there is no technological treatment that can inactivate these substances. The presence of -lactams antibiotics in milk may represent a risk to consumer health, such as allergic reactions and anaphylactic shock in sensitive individuals (Mendes et al. 2008), and their exposure may lead to an increase in the numbers of antibiotic resistant microorganisms. In the European Union, MRLs for this class of compounds in milk vary from 4.0 to 100.0 ng ml1 (Table 1). For the Brazilian National Residue Control Plan (NRCP) (Mauricio et al. 2009), no MRL was set for the -lactams in milk until 2011. Currently, similar values to European Union MRLs were adopted. -lactam antibiotics have several pharmaceutical
498
L. Jank et al.
dosage forms for veterinary use in Brazil, with more than 160 registered products in 2011, of which 74 products have benzylpenicillin, also known as penicillin G, as the active compound. Sixty-eight of these benzylpenicillin products are injectable formulations, four are intra-mammarian infusions and two are ointments. In other words, the majority are for dairy cattle treatment. The same situation is observed in the case of cloxacillin, which has only injectable products, and with ceftiofur, which is available in seven injectable forms and one intra-mammarian infusion. Pharmaceuticals preparations containing -lactams antibiotics for bovine are summarized in Table 1. The widespread occurrence of -lactam antibiotics in milk has been reported in several publications (Kress et al. 2007). Several methods have been published in the past years using LC-MS/MS and other techniques (Mastovska and Lightfield 2008; Bailo´n-Pe´rez et al. 2009; Ortelli et al. 2009). -lactam antibiotics have poor stability in standard solutions and some specific compounds such as amoxicillin suffer degradation in solutions in 1 week or even fewer days. Due to the high polar characteristic of this class of compounds, chromatographic separation can be more complex and laborious than other antibiotic groups commonly analysed in milk, such as quinolones or sulfonamides. Generally, reversed-phase chromatography is used, but polar phases columns have been successfully used, as hydrophilic interaction columns (HILIC). Riediker et al. (2004) studied the stability of five -lactams in milk samples, and developed an extraction method that consists of a liquid–liquid extraction (LLE) with n-hexane followed by solid-phase extraction (SPE). A fast method was developed by Kantiani et al. (2009) to determine ten analytes in bovine milk (six penicillins and four cephalosporins), based on on-line solidphase extraction-liquid chromatography/electrospraytandem mass spectrometry (SPE-LC/ESI-MS-MS). Gaugain-Juhel et al. (2009) developed a method of screening 58 antibiotics, including penicillins, with two very short LLE, using acetonitrile for penicillins, cephalosporins, macrolides and sulfonamides and 5% trichloroacetic acid (TCA) solution for tetracyclines, quinolones, aminoglycosides and lincomycin. Another LLE method for detection of six penicillin residues, followed by an evaporation clean-up step, was carried by Feng et al. (2009). In order to prevent the degradation of penicillins throughout extraction, a derivatization reaction was proposed by van Holthoon et al. (2010). Their method used the precipitation of milk proteins with acids and clean-up of the supernatant in solid-phase extraction cartridge. Summarizing, -lactams analysis in milk are considered very complex mainly due to the low stability of this class of compound (Bittencourt 2003). Satisfactory results were obtained using extraction and clean-up techniques in tandem (i.e. LLE-SPE, SPE-SPE)
Table 1. European Community MRLs for -lactams antibiotics residues in bovine milk and the number of pharmaceutical dosage forms for veterinary use in milkproducing cattle in Brazil.
Compound
MRLa (ng ml1)
Pharmaceuticalsb
4 4 4 30 30 30 50 50 100 100
24 11 74 – 13 – 6 – 13 8
Amoxicillin Ampicillin Penicillin G Oxacillin Cloxacillin Dicloxacillin Cefoperazone Cefazolin Cefalexin Ceftiofur
Notes: aEuropean Commission (1990). Data were obtained from the Sindan website, 2010 (National Union of Industry Products for Animal Health). Available from: http://www.cpvs.org.br/. b
(Turnipseed et al. 2008; Kantiani et al. 2010). However, these techniques require a more complex and more expansive analysis. For this reason, the present work aims to develop a fast and simple method for quantitative and confirmatory analysis of -lactams in milk samples. Extraction consists of a simple LLE protocol, using a small volume of sample and solvent, and a very fast chromatographic method. Chemical structure of -lactam antibiotics included in this work are shown in Figure 1.
Materials and methods Chemicals and reagents Ceftiofur (CFT), penicillin G (PNG), penicillin V (PNV), oxacillin (OXA), cloxacillin (CLX) and dicloxacillin (DCX) standards were obtained from SigmaAldrich Logistik (Scnelldorf, Germany) of >95% certified purity. Stock standard solutions were prepared by dissolving all compounds individually in 0.5% of polypropileneglycol 3000 in acetate buffer (pH 4.5), at a concentration of 0.5–3.75 mg ml1, to make easier dissolutions to the work concentration pool. PNG and PNV stock solutions were diluted 1000-fold, to obtain 0.5 mg ml1. OXA, CLX and DCX were also diluted 1000-fold to obtain 3.75 mg ml1; to CFT, a dilution factor of 1:800 was applied to obtain a final concentration of 12.5 mg ml1. Dilutions of stock solutions to prepare a pool were made with ultrapure water. Acetonitrile HPLC grade (ACN) and ammonium acetate were purchased from Merck (Darmstadt, Germany); formic acid (FA) was from J. T. Baker (Phillipsburg, NJ, USA). Deionized ultra-pure water ( 241a 524 > 210 335 > 176 335 > 160 351 > 114 351 > 160 402 > 160 402 > 243 436 > 160 436 > 277 467 > 114 467 > 160
126 126 96 96 86 86 96 96 101 101 106 106
25 79 17 17 45 19 19 19 19 19 47 19
16 10 20 18 16 22 16 26 16 30 24 16
Compound
Typical retention time (min) 7.30 7.57 7.70 7.82 7.95 8.10
Notes: aBold transitions are used for quantitative analysis. DP, declustering potential; CE, collision energy; CXP, collision cell exit potential.
Optimised extraction procedure consisted of subsequently adding four aliquots of 1.0 ml of ACN to a volume of 2.0 ml of milk sample, mixing in a vortex for approximately 10 s approximately between each addition. After this step, the sample was mixed in a headover-head shaker for 15 min, then 1.0 g of sodium chloride was added, and 15 min mixing more in the head-over-head shaker was carried out. Samples were then centrifuged for 5 min, at approximately 3000 g under refrigeration (5 C). Aliquots of supernatant were transferred to HPLC vials and submitted to LC-MS/MS analysis. A volume of 10 ml of extract was injected in the analytical system.
Stability of stock solutions A stability study to evaluate stock solutions was proposed. It consisted of a comparison of stock solutions prepared in water and in acetate buffer with different concentrations of polypropyleneglycol 3000 (PPG), a polymer that can provide some difficulty to -lactam degradation, probably because of pseudomicelle formation. PPG concentrations tested were 0.25%, 0.5%, 1.0%, 1.5% and 2.0%. Analytes’ concentrations were 1.0 mg ml1 (CFT), 0.5 mg ml1 (PNG and PNV) and 3.75 mg ml1 (OXA, CLX and DCX). Solutions were diluted 1000-fold before every injection. Solutions were injected on the day of preparation, in the following 2 weeks and after that every 15 days.
Matrix effect The assessment of matrix effect was performed through the preparation and analysis of three calibration curves. Curve type I, called ‘curve in solvent’, was prepared by diluting the standard solution in the
mobile phase directly into the vial, giving a range of concentrations of 1–200 ng ml1, according to the MRL for each substance. Curve type II, or ‘recovery curve’, was prepared by spiking blank samples with the desired amounts of -lactams, which were extracted and analysed as conventional samples. Curve type III, or ‘tissue standard curve’, was prepared by adding a -lactam standard solution in extracts of blank samples after extraction. All curves were made in the same range of concentrations.
Validation procedure Method validation was carried out following European Commission Decision 2002/657/EC (European Commission 2002) requirements for veterinary drug residue methods. Specificity, selectivity and stability were evaluated. Blank milk samples were spiked with -lactams at concentrations corresponding to 0.5, 1.0 and 1.5 MRL, in order to investigate parameters such as linearity, repeatability and reproducibility, as well as decision limit (CC) and detection capability (CC). For validation procedures, batches were composed by 21 spiked samples (seven for each concentration level: 0.5, 1.0 and 1.5 MRL), a calibration curve (0, 0.25, 0.50, 1.0, 1.5 and 2.0 MRL), and three ‘tissue standards’ (i.e. extracts of blank samples to which an amount of standard solution was added to obtain a concentration at MRL). This procedure was repeated three times on three different days.
Results and discussion Several experiments were performed for method optimisation, such as different chromatographic columns,
Food Additives and Contaminants different extraction methods and chromatographic conditions employed.
Sample extraction Despite the fact that the majority of authors in the literature adopted solid-phase extraction (SPE) to prepare milk samples, we chose to develop a method through liquid–liquid extraction (LLE) to reduce costs and, mainly, analysis time, making it simpler and easier, and preferentially utilising a small amount of solvent and sample. Although numerous solvents can be used to promote proteins precipitation in milk, as trichloroacetic acid, methanol or ethanol acidified, we chose ACN for solvent extraction to avoid the acid degradation of -lactams. Five extraction procedures were evaluated. Volumes of ACN addition, sample volume, saltingout effect, lipids removal and concentration by evaporation were investigated and optimised. . In extraction procedure 1 (EP1), 5 ml of milk were extracted with 10 ml of ACN. Extraction solvent was added in four aliquots of 2.5 ml and sample was vortexed between each addition for approximately 10 s. Tubes were mixed for 20 min in a mechanical shaker, followed by centrifugation for 5 min at 5000 g under refrigeration (5 C). Supernatant was evaporated, reconstituted in 1 ml of ACN:H2O (1:1) and injected in LC-ESI-MS/MS system. . For extraction procedure 2 (EP2), 2 ml of milk were extracted with 4 ml of ACN. Extraction solvent was added in four aliquots of 1.0 ml and sample was vortexed between each addition for approximately 10 s. Tubes were mixed per 20 min in a mechanical shaker and then 1.0 g of sodium chloride was added. Tubes were mixed for more 20 min followed by centrifugation for 5 min at 5000 g under refrigeration (5 C). An aliquot of supernatant (1 ml) was directly injected in LC-ESI-MS/MS system. . Extraction procedure 3 (EP3) was similar to EP2, but here just 2 ml of ACN were used (4 0.5 ml) without the addition of sodium chloride. In this protocol milk proteins do not precipitate and the resulting samples showed cloudy aspect. A small portion of ACN in a ratio of 1:1 with milk was insufficiently able to promote protein removal. . In extractions procedure 4 (EP4), using 2 ml of milk and 5 ml of ACN (2.0 þ 3.0 ml), 5 ml of chloroform were added after shaking step. The tubes were then shaken for 20 min more and centrifuged as in EP1 and EP2. The aqueous portion was removed, the
501
supernatant was evaporated, it was reconstituted in 1 ml of ACN:H2O (1:1) and injected into the LC-ESI-MS/MS system. . For extraction procedure 5 (EP5), 2 ml of milk and 5 ml of ACN (2.0 þ 3.0 ml) were used following the same procedures as for EP2. In all tests, solvent (ACN) was added gradually to improve protein precipitation, and for the same reason samples were vortexed between additions. With the exception of EP 3, where no protein precipitation was observed, all others proved to be adequate, showing satisfactory recovery (60%). EP1, indeed, had a fivefold concentration factor and EP4 twofold. Although EP2 has no concentration factor, this protocol gave the best response. Furthermore, EP2 does not require an evaporation step, which provides the shortest analysis time of all experiments. For EP4, chloroform was tested to remove lipids and waterinsoluble sample components, but no advantage was perceived over EP1, EP2 or EP5. Sodium chloride was added in this method to saturate the aqueous phase and force organic compounds to migrate to the organic phase. After centrifugation, the organic phase was very clear, and an aliquot was injected directly into the LCMS/MS system, without a filtration step. EP2 demonstrates itself to be very effective, even with the advantage of using a small sample volume. Turnipseed et al. (2008) developed a method for several veterinary drugs residues in milk, including -lactams. Extraction was first performed with ACN (1 ml). Two additional steps of clean-up were proposed using SPE (OasisÕ HLB 3 ml 60 mg) followed by ultrafiltration trough a Microcon YM-30 centrifugal filter device (Millipore). Becker et al. (2004) published another similar method for the analysis of 15 -lactams in milk and kidney. For milk, specifically, the extraction protocol was very similar to the present approach using protein precipitation with CAN followed by a salting-out procedure. But final clean-up was also performed with Oasis SPE cartridges. SPE was also applied by Stolker et al. (2008) to analyse more than 100 veterinary drugs residues (including six -lactams) in milk. After ACN addition, supernatant was applied to a Strata-X SPE column. Our method has a lower scope, but the extraction procedure is complete without SPE or additional steps. As the overall aim was to develop a simple, fast and cheap extraction protocol, the present results were considered to be an achievement. The method can be easily applied for several samples in routine laboratories.
Stability of stock solutions The fast degradation of -lactams is well-established. Our tests to determine the expiry period for stock solutions showed a loss of analytes after 7 days when
502
L. Jank et al.
the solvent is pure water. In the presence of acetate buffer and PPG, degradation was only detectable in the second week. However, in concentrations more elevated with PPG (above 1.0%), some compounds (OXA, DCX and CLX) showed a decrease up to 60%. After 45 days, a rate of 60–90% of degradation was observed for all compounds in all solvents compositions. Moreover, when stock solutions were stored in aliquots (i.e. microcentrifuge tubes of 2 ml) under temperatures below 10 C, avoiding the exposing stock solutions to several cycles of frozen and refrozen at the moment of make dilutions, the stability was maintained for more longer periods. It was established in this test that stock solutions with 0.5% of PPG, stored in aliquots of 1–2 ml, showed the best stability, and can be stored for 6 months at a temperature under 10 C. However, work solutions should always be prepared at the moment of analysis to avoid the degradation of compounds in lower concentrations.
LC-MS/MS All compounds investigated showed an adequate response in positive ionisation mode (ESIþ), whereas PNG, OXA and DCX were satisfactorily ionised in both positive- and negative-ion modes. To proceed with the simultaneous analysis of all six compounds, a positive-ESI mode was chosen for all analytes. Experimental results showed formic acid 0.1% was the most suitable additive to enhance peak resolution and sensitivity. Acetic acid (0.1%) and ammonium formiate (5 mM) were also evaluated. As a stationary phase, two columns were tested: XTerraÕ C18 (100 2.1 mm, 3.5 mm) and SinergyÕ C18 (150 4.6 mm, 4.0 mm) which generated the best result. In X-Terra column, no effective retention was observed, with elution of analytes in the first 1 min of chromatographic analysis. Sinergy columns gave higher retention, which avoided co-elution of analytes with matrix polar co-extractives in the beginning of analysis, which generally cause enhancement of matrix effects under analytes, as signal suppression. Chromatographic separation of -lactams is a problem since due to the high polarity of these compounds they generally elute in the first minutes of a run, together with the dead volume of the columns. This is especially true for reverse-phase columns. Optimisation proceeded using a polar column, which promotes a more efficient analyte–column interaction. Indeed, the retention time window for all group of analytes is very small, which can lead to co-elution. Gradient mode was optimised to avoid possible co-elutions. The optimised chromatographic method has a total time of 12 min, which reliable for routine analysis since more samples can be analysed in a short period.
Chromatograms for all analytes with two monitored transitions are shown in Figure 2.
Validation procedure The validation procedure was carried out according to European Union Commission Decision 2002/657/EC requirements (European Commission 2002). Parameters considered more significant are described below.
Determination of limit of detection (LOD) and limit of quantification (LOQ) Considering that the mathematical approach to LOD and LOQ determination using the deviation of blank samples resulted in improbably low values, these parameters were established using data from spiked samples. To carry out the experimental determination of the lowest concentration detectable as required by guidelines for implementation of the European Union Decision (LOD and LOQ), calibration curves with lower concentrations than those used in previous tests (0.10 and 0.25 MRL) were analysed. The lowest spiked points were correctly identified and quantified. Based on these experimental data, LOD and LOQ were defined as 5% and 10%, respectively, of the MRL for each compound. LOD and LOQ values are presented in Table 3 as correlation coefficients, ranging from 0.9676 to 0.9992, which matches the internal criteria of our laboratory, which requires r2 > 0.95 for matrixmatched calibration curves.
Repeatability and reproducibility Repeatability and reproducibility data are summarized in Table 4. All values were satisfactory, considering the level calculated by the Horwitz equation. The coefficient of variation of repeatability (CVr) is acceptable when less than two-thirds the Horwitz CV. The CV for within-laboratory reproducibility (CVwIR) was performed by combined data obtained by three different analysts. CVwIR must be not higher than the Horwitz CV.
Accuracy Accuracy was determined using a comparison between the calculated concentration obtained by the matrixmatched calibration curve and the analyte amount added to the sample in the spiking procedure. The average accuracies obtained in three batches are listed in Table 5. In routine analysis, the accuracy determination for ‘tissue standard’ samples-type fortified in the value of the MRL was accompanied in each batch either for validation studies and analyses of routine or
Food Additives and Contaminants
503
Figure 2. LLE-LC-MS/MS chromatograms for compounds at the MRL level concentration spiked milk sample (1 ¼ CFT, 2 ¼ PNG, 3 ¼ PNV; 4 ¼ OXA, 5 ¼ CLX and 6 ¼ DCX).
504
L. Jank et al.
Table 3. LOD and LOQ for -lactam antibiotics in milk.
LOD (ng ml1) LOQ (ng ml1) r2
CFT
PNG
PNV
OXA
CLX
DCX
10.0 25.0 0.9718–0.9956
0.4 1.0 0.9898–0.9982
0.4 1.0 0.9676–0.9980
3.0 7.5 0.9894–0.9992
3.0 7.5 0.9852–0.9986
3.0 7.5 0.9718–0.9956
Table 4. Repeatability and reproducibility data. Fortification levels (n ¼ 6 for each batch; three batches for each level) Analyte
Parameters
CFT
Average CVr CVwIR Average CVr CVwIR Average CVr CVwIR Average CVr CVwIR Average CVr CVwIR Average CVr CVwIR
PNG PNV OXA CLX DCX
0.5 MRL 49.5 6.3 6.9 2.0 5.8 6.3 2.0 4.5 7.4 13.7 7.3 10.9 15.4 6.4 6.3 15.2 8.7 9.1
1.0 MRL
52.0 6.5
51.3 8.1
2.1 2.2
2.2 7.9
2.3 6.6
2.1 4.9
16.5 5.3
16.9 5.2
16.8 3.7
16.0 5.5
17.5 2.6
15.1 5.7
100.5 13.8 9.5 4.2 14.0 8.9 4.2 10.7 9.9 28.1 10.2 12.4 29.9 6.4 10.1 31.0 8.5 11.0
1.5 MRL
101.9 4.0
109.9 7.3
4.3 5.9
4.1 5.5
4.6 6.4
3.9 4.2
36.2 6.5
31.9 2.1
36.2 5.8
30.8 1.1
37.0 5.9
30.4 2.9
129.2 6.3 11.1 5.4 4.7 10.2 5.6 10.1 14.2 36.6 10.2 19.9 38.0 11.2 19.8 38.5 11.3 21.8
156.5 7.8
155.6 7.0
6.6 7.5
6.0 4.5
7.3 7.9
5.8 5.5
57.8 7.4
48.0 2.6
58.4 7.2
45.1 3.4
60.0 10.7
43.9 2.1
Note: CVr, coefficient of variation (%) of the repeatability; CVwIR, coefficient of variation (%) of the within-laboratory reproducibility according to European Commission (2002). Table 5. Limit of decision (CC), detection capability (CC), accuracy and recovery data.
CFT PNG PNV OXA CLX DCX
MRL (mg l1)
CC (mg l1)
CC (mg l1)
Accuracy average (%)a
Recovery (%)
100.0 4.0 4.0 30.0 30.0 30.0
120.4 4.7 4.7 36.5 35.6 36.3
147.9 5.7 6.1 53.7 52.8 56.6
104.1 107.0 105.0 106.9 107.8 109.9
41.9 54.2 63.7 73.8 79.3 81.3
Note: aAccuracies for the determination of samples spiked at the MRL (n ¼ 7): Accuracy (%) ¼ (calculated concentration/spike concentration) 100
for the construction of control charts for statistical tracking of the process, providing data for future estimation of the uncertainty of measurement. Tissuestandard samples were composed by extracts of blank samples in which a standard solution to obtain an MRL concentration in the final volume was added.
Recovery and matrix effect Area values found for the TS curve were more intense than those observed in the solvent curve. This fact can
be explained if one considers that the matrix in the study, milk, provides an increment to ionisation when compared only with the mobile phase. In some cases, such as oxacillin, the increment to ionisation is not so intense; on the other hand, ceftiofur presents considerable increases in the intensity of points on the TS curve in relation to the curve in solvent. Based on these data, recovery was calculated by considering that values found in the TS curve represent the intensity of the analyte considering the matrix effect, but without losses from the extraction process (thus, 100%)
Food Additives and Contaminants
505
Figure 3. Matrix effect evaluation.
Table 6. ANOVAs for the comparison of raw and UHTtreated milk.
Analyte
Found value
Table value
Is there a significant difference?
CFT PNG PNV OXA CLX DCX
0.70 4.02 21.74 667.81 1318.25 41.33
12.22 12.22 12.22 12.22 12.22 12.22
No No Yes Yes Yes Yes
compared with the curve of the recovered which represents the signal observed after sample treatment. Recovery values calculated for each compound are shown in Table 5. Results for each compound are shown in Figure 3. Full validation was carried out with UHT milk. To evaluate if raw milk could present some matrix effect, three calibration curves, with eight points, were
prepared for each matrix (UHT and raw milk). For some compounds there was no significant difference between each kind of milk; however, for others the difference was very significant, presenting a signal twice that of the same level of concentration of one type of milk over another. When there was this difference, the UHT milk signal was larger than that from the raw milk, indicating that the raw milk had increased the ion suppression in comparison with milk treated industrially. An ANOVA test was performed for UHT and raw milk, as shown in Table 6; plots for each compound are presented in Figure 4. For this reason, it is mandatory to make calibration curves in blank samples with the same kind of milk (raw or UHT, for instance) that will be analysed.
Application to real samples The present method was used to analyse 84 raw milk samples, collected in several regions of Brazil. Just one non-compliant sample was detected, containing 7.9 mg l1 of penicillin G.
506
L. Jank et al.
Figure 4. Matrix effect according to milk treatment: RAW, raw milk; UHT, UHT-treated milk.
Conclusion Almost all currently used methodologies for the analysis of -lactams antibiotics residues use SPE as clean-up procedure. The development of an LLE method as sample pretreatment is of great value since it becomes cheaper by avoiding the use of SPE cartridges, and faster, which are very important factors on deciding what method should be adopted for routine analysis. The method reported here has high sensitivity and gives satisfactory results for the identification and quantification of six -lactams antibiotics in bovine milk. The LODs are well below the MRLs set by the European Union for all compounds. The present method is currently included in the Brazilian National Residues Control Plan. Acknowledgements The authors would like to thank the National Council for Scientific and Technological Development
(CNPq) for the fellowships provided to L. Jank and P. Tarouco.
References Bailo´n-Pe´rez MI, Garcı´ a-Campan˜a AM, del Olmo-Iruela M, Ga´miz-Gracia L, Cruces-Blanco C. 2009. Trace determination of 10 -lactam antibiotics in environmental and food samples by capillary liquid chromatography. J Chromatogr A. 47:8355–8361. Becker M, Zittlau E, Petz M. 2004. Residue analysis of 15 penicillins and cephalosporins in bovine muscle, kidney and muscle by liquid chromatography-tandem mass spectrometry. Analyt Chim Acta. 520:19–32. Bittencourt MS. 2003. Cefixima: validac¸a˜so de me´todos analı´ ticos e estudo preliminar da estabilidade [master thesis]. Porto Alegre: Universidade Federal do Rio Grande do Sul. European Commission. 1990. Council Regulation (EEC) No. 2377/90 of 26 June 1990: laying down a Community procedure for the establishment of maximum residue limits
Food Additives and Contaminants of veterinary medicinal products in foodstuffs of animal origin. Off J Eur Comm. L224:1–8. European Commission. 2002. Commission Decision 2002/ 657/EC of 12 August 2002: implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Off J Eur Comm. L221:8–36. Feng Q, Zheng WJ, Chen GL, Qiu SL. 2009. Determination of six penicillins residues in milk by LC-MS/MS. Chin J Antibiot. 36:248–351. Gaugain-Juhel M, Dele´pine B, Gautier S, Fourmond MP, Gaudin V, Hurtaud-Pessel D, Verdon E, Sanders P. 2009. Validation of a liquid chromatography-tandem mass spectrometry screening method to monitor 58 antibiotics in milk: a qualitative approach. Food Addit Contam. 26:1459–1471. Kantiani L, Farre´ M, Freixedas JMG, Barcelo´ D. 2010. Development and validation of a pressurised liquid extraction liquid chromatography-electrospray-tandem mass spectrometry method for -lactams and sulfonamides in animal feed. J Chromatogr A. 1217:4247–4254. Kantiani L, Farre´ M, Sibum M, Postigo C, Alda ML, Barcelo´ D. 2009. Fully automated analysis of -lactams in bovine milk by online solid phase extraction-liquid chromatography-electrospray-tandem mass spectrometry. Analyt Chem. 81:4285–4295. Kress C, Seidler C, Kerp B, Schneider E, Usleber E. 2007. Experiences with an identification and quantification program for inhibitor-positive milk samples. Analyt Chim Acta. 586(1–2):275–279. Mastovska K, Lightfield AR. 2008. Streamlining methodology for the multiresidue analysis of -lactam antibiotics in bovine kidney using liquid chromatography-tandem mass spectrometry. J Chromatogr A. 1202(2):118–123. Mauricio AQ, Lins ES, Alvarenga MB. 2009. A national residue control plan from the analytical perspective – the Brazilian case. Analyt Chim Acta. 637(1–2):333–336.
507
Mendes CG, Sakamoto SM, Silva JBA, Leite AI. 2008. Pesquisa de Resı´ duos de Beta-Lactaˆmicos no leite cru comercializado clandestinamente no municı´ pio de Mossoro´, RN, utilizando o Delvotest SP. Arquivos do Instituto de Biologia. 75:95–98. Michaelidou AM. 2008. Factors influencing nutritional and health profile of milk and milk products. Small Ruminant Res. 79:42–50. Ministe´rio da Agricultura, Pecua´ria e Abastecimento (MAPA). 2009. Projec¸o˜es dos Agronego´cios – Brasil 2008/09 a 2018/19. Brası´ lia: MAPA. p. 30–31. Ortelli D, Cognard E, Jan P, Edder P. 2009. Comprehensive fast multiresidue screening of 150 veterinary drugs in milk by ultra-performance liquid chromatography coupled to time of flight mass spectrometry. J Chromatogr B. 23:2363–2374. Riediker S, Rytz A, Stadler R. 2004. Cold-temperature stability of five -lactams antibiotics in bovine milk and milk extracts prepared for liquid chromatography-electrospray ionization tandem mass spectrometry analysis. J Chromatogr A. 1054:359–363. Stolker AAM, Rutgers P, Oosterink E, Lasaroms JJP, Peters RBJ, van Rhijn JA, Nielen MWF. 2008. Comprehensive screening and quantification of veterinary drugs in milk using UPLC-ToF-MS. Analyt Bioanalyt Chem. 391:2309–2322. Turnipseed SB, Andersen WC, Karbiwnyk CM, Madson MR, Miller KE. 2008. Multi-class, multi-residue liquid chromatography/tandem mass spectrometry and confirmation methods for drug residues in milk. Rapid Comm Mass Spectrom. 22:1467–1480. van Holthoon F, Mulder PPJ, van Bennekom EO, Heskamp H, Zuidema T, van Rhijin HA. 2010. Quantitative analysis of penicillins in porcine tissues, milk and animal feed using derivatisation with piperidine and stable isotope dilution liquid chromatography tandem mass spectrometry. Analyt Bioanalyt Chem. 396: 3027–3040.
Food Additives and Contaminants Vol. 29, No. 4, April 2012, 508–516
High-throughput multiclass screening method for antibiotic residue analysis in meat using liquid chromatography-tandem mass spectrometry: a novel minimum sample preparation procedure M.S. Bittencourta*, M.T. Martinsab, F.G.S. de Albuquerquea, F. Barretoab and R. Hoffac a Ministe´rio da Agricultura, Pecua´ria e Abastecimento, Laborato´rio Nacional Agropecua´rio – LANAGRO/RS, Porto Alegre, RS, Brazil; bFaculdade de Farma´cia, Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brazil; cInstituto de Quı´mica, Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brazil
(Received 23 November 2010; final version received 13 July 2011) A multiresidue and multiclass method based on liquid chromatography-tandem mass spectrometry for the determination of antibacterials was developed and validated for screening purposes. This method can be applied to commonly used drugs in veterinary medicine such as tetracyclines, quinolones and sulfonamides. Sample preparation consists in cell disruption with sand (previously purified and washed with EDTA 100 mM) followed by protein precipitation with acidified acetonitrile. Validation was conducted in accordance to European Union requirements (2002/657/EC) for qualitative methods covering detection capability (CC ), selectivity, specificity and stability. The method enabled the detection of 21 different drugs and had a false-compliant rate of 55% ( error) at between 25% and 50% of the maximum residue levels established by legal authorities. The methodology was successfully applied to incurred poultry samples. Keywords: chromatography – LC/MS; extraction; in-house validation; method validation; veterinary drug residues – antibiotics; veterinary drug residues – fluoroquinolones; veterinary drug residues – sulphonamides; veterinary drug residues – tetracycline; meat; animal products – meat
Introduction Antibacterials are substances frequently used for the prevention and treatment of diseases in cattle and poultry management. Tetracyclines, quinolones and sulfonamides are the most commonly used antibacterial groups for these purposes and may leave residues in edible tissues that could be associated with public health problems (Stolker et al. 2007; Gaugain-Juhel et al. 2009). In many countries, governmental authorities have established monitoring programmes to determine antibacterials in foods, as well as the highest allowable residue levels. Regarding residues of veterinary drugs in foodstuffs of animal origin, maximum residue limits (MRLs) were set by the Codex Alimentarius and/or regional or local authorities and monitoring plans were set up for ensuring MRLs and prevent the presence of prohibited substances in food. These MRLs are generally in range of 25–300 mg kg1 for most common antibacterials; however, they go up to 400 mg kg1 for flumequine in chicken muscle, for example (Boscher et al. 2010). An important tool to monitor closely and ensure this compliance in Brazil is the National Residue Control Plan (NRCP) (Mauricio et al. 2009). For this purpose, several analytical methods were applied. Microbiological assays have been most commonly used to analyse such residues; the *Corresponding author. Email:
[email protected] ISSN 1944–0049 print/ISSN 1944–0057 online ß 2012 Taylor & Francis http://dx.doi.org/10.1080/19440049.2011.606228 http://www.tandfonline.com
advantages of these methods are the possibility of detecting a wide number of compounds simply and at a low cost. However, in some cases these methods are not sensitive enough (false-negative results) and are not really specific (false-positive results) (Gaugain-Juhel et al. 2009). Liquid chromatography (LC) coupled with mass spectrometry (MS) or tandem mass spectrometry (MS/MS) has become the most powerful technique for the determination of antibacterials in food matrices. LC-MS/MS, in particular LC triple-quadrupole (QqQ) MS/MS, has today become the technique of choice in antibacterial residue analysis. European Commission Decision 2002/657/EC states that methods based only on chromatographic analysis without the use of molecular spectrometric detection are not suitable for use as confirmatory methods (Bogialli and Di Corcia 2009). Over the last decade LC-MS/MS has become an essential technique for food analysis (Chico et al. 2008). Several papers have been published in recent years dealing with this issue. Most reported multiresidue methods target a few closely related compounds, usually belonging to a single drug class (Stolker et al. 2007). The cost-effectiveness of analytical procedures is becoming an important issue for all laboratories involved in the residue analysis of contaminants in food. An alternative to improve
Food Additives and Contaminants cost-effectiveness is to maximize analyte numbers that may be determined by a single portion of test material (Bogialli and Di Corcia 2009). Granelli and Banzell (2007) developed a screening method for detecting 19 antibacterials from five different classes in muscle and kidney. The scope of this work was extended to quantification and confirmation for the same compounds in muscle (Granelli et al. 2009). A multiclass antimicrobial determination using pressurized liquid extraction and LC-MS/MS for 31 drugs in beef was developed, but not including sulfamethoxazole, sulfamerazine, sulfamethazine, sulfachlorpyridazine, sarafloxacin, difloxacin, oxolinic acid and nalidixic acid which are considered very relevant for the NRCP (Carretero et al. 2008). Chico et al. (2008) reported the validation of a multiresidue method for 39 antibacterials in poultry meat and applied this method to several animal species using water and methanol as extraction solvents and ultra-high-pressure-liquid chromatography coupled with MS/MS. Another method for the determination of antibacterials in poultry meat based on QuEChERS methodology was applied to a large number of substances (Stubbings and Bigwood 2009). Martos et al. (2010) developed a method using liquid– liquid extraction (LLE) and LC-MS/MS for the determination of antibacterials. Recently, a multiclass method for detecting and quantifying veterinary drug residues in feedingstuffs was developed (Boscher et al. 2010). The aim of the present study was to develop and validate a simple, fast and inexpensive multiresidue screening method for the determination of 21 antibacterials in meat (cattle and poultry) using LC-ESIMS/MS in positive-ion mode. These drugs are included in NRCP for meat matrices. Validation was conducted for screening purposes based on European Commission Directive 2002/657/EC with measurements of detection capability (CC ), stability, specificity and applicability. The proposed method presented adequate compound separation, a simple extraction procedure, and a detection capability (CC ) between 25% and 50% of maximum residues level established by legal authorities, having a false-compliant rate of 55% ( -error).
Materials and methods Materials and reagents Analytical standards of sulfadimethoxine (SDMX), sulfaquinoxaline (SQX), sulfadiazine (SDZ), sulfachlorpyridazine (SCP), sulfathiazole (STZ), sulfapyridine (SPY), sulfamerazine (SMR), sulfamethoxazole (SMA), sulfamethazine (SMZ), chlortetracycline (CTC), tetracycline (TC), oxytetracycline (OTC), doxicycline (DOX), oxolinic acid (OXO), nalidixic acid (NALID), flumequine (FLU), ciprofloxacin
509
(CIPRO), difloxacin (DIFLO), enrofloxacin (ENRO), norfloxacin (NOR), and sarafloxacin (SARA) were purchased from Riedel-de-Haen (Buchs, Switzerland) or by Sigma-Aldrich (St. Louis, MO, USA) as a powder. Stock standard solutions of each compound were prepared by dissolving 10 mg of analytical standard in 10 ml of appropriate solvent (methanol for tetracyclines, sulfonamides and quinolones; methanol with some drops (approximately two) of NaOH 1 M for fluorquinolones). Aliquots of each stock solution were diluted to obtain final concentrations of 10 and 1 mg ml1 and were stored at 20 C. Formic acid was obtained from J.T. Baker (Phillipsburg, NJ, USA); methanol and acetonitrile (HPLC grade) were purchased from Merck (Darmstadt, Germany). All water used was ultra-pure deionized water produced by a Milli-Q apparatus (Millipore, Bedford, MA, USA). Di-sodium ethylenediamine tetracetate (Na2EDTA) was obtained from Sigma. Sand was purchased from Merck or was home treated by purification of sea sand (USP 1995), and washed with EDTA 100 mM. Sea sand (40–200 mesh size) was purified by calcination at 500 C for 2 h and washed three times with hydrochloric acid:water (1:2), using elution by gravity. The sand was then dried at 100 C for 4 h with periodic mixing and stored until analysis. At the day of analysis, portions of sand were treated with EDTA 100 mM, in a proportion of 1:2 (w/v). EDTA-sand was just gently dried at room temperature for 12–24 h to avoid decreasing in the metal chelating action (Bogialli et al. 2006). Blank samples were obtained from previously analysed samples and obtained in local markets.
LC-MS/MS instrumentation LC-MS/MS measurements were carried out using a Waters Alliance 2795 system coupled to a Quattro Micro triple quadrupole mass spectrometer from Micromass (Waters) with an electrospray source. Separation was achieved on a Symmetry C18 LC column (75 4.6 mm; 3.5 mm particle diameter) from Waters. A Phenomenex C18 (4.0 3.0 mm) was used as a guard column. The flow rate used was 400 ml min1 and the column temperature was set at 20 C. A gradient elution programme was used with solvent A (aqueous solution 0.1% formic acid) and solvent B (acetonitrile with 0.1% formic acid) as follows: 98% A (5 min), 98–80% A (2 min), 80% A (3 min), 80–50% A (1 min), 50% A (4 min), 50–98% A (2 min), kept at 98% A for 17 min returning to the initial composition, and held for 3 min to equilibrate the column. Mass analysis conditions optimisation were achieved on infusion injection at a flow rate of 10 ml min1. Each standard solution was prepared
510
M.S. Bittencourt et al.
separately in methanol with formic acid 0.1% at 1 mg ml1. Source block temperature was set at 120 C in positive-ion mode with a capillary voltage of 3.0 kV. Nitrogen gas was used as a desolvation agent and nebuliser gas (N2) at flow rates of 400 and 50 l h1, respectively. Argon was used as the collision gas. Detection was operated in multiple reaction-monitoring (MRM) mode. Instrument control and data processing were carried out by means of Masslynx 4.1 software purchased from Micromass.
Meat samples Meat samples (muscle) of cattle and poultry were homogenised to a semi-solid consistency using a food processor (Oster, Sunbean Products, Inc., USA) and stored at 18 C until analysis. Samples were kept at room temperature until defrosted and a portion of 6 g was weighed into a 50 ml polypropylene centrifuge tube. Spiked samples were prepared by adding the proper amount (150 and 300 ml) of a working solution containing all analytes (1 mg ml1). Samples were stirred and allowed to stand for 10 min before extraction.
Methods Extraction procedure Extraction of antimicrobial agents from meat was performed through cell disruption by mixing 6 g of chopped and homogenised muscle sample with 4 g of sand (previously washed with EDTA) using a glass stick. To this mixture an aliquot of 250 ml of EDTA 100 mM was added and sample was homogenised in vortex for 30 s. Then, an aliquot of 600 ml of methanol was added (for spiked samples, just the necessary amount of methanol was added in order to complete 600 ml, considering methanol added in spiking solution). After that the mixture was vortexed for another 30 s and placed in an ultrasonic bath for 10 min at maximum power. Following, samples were centrifuged for 30 min at 3000 g (5 C). An aliquot of supernatant (800 ml) was transferred to a microtube containing 400 ml of acid formic 0.1% in acetonitrile, vortexed for 15 s and centrifuged for 30 min at 12 000 g (5 C). The supernatant was transferred to another microtube containing 600 ml of initial mobile phase (formic acid 0.1% in water:formic acid 0.1% in acetonitrile – 98:2) and centrifuged for 20 min at 12,000 g (5 C). The final supernatant was directly placed to HPLC vial and analysed by LC-MS/MS. Method validation An in-house validation procedure was conducted in accordance with European Commission Directive
2002/657/EC for screening purposes. According to these criteria one method can be validated and was used for screening purposes when there is a falsecompliant rate of 55% ( -error) at the level of interest. In the case of a suspected non-compliant result, it must be confirmed by a suitable confirmatory method (Ortelli et al. 2009). The proposed method in this work was mainly dedicated to screening and included parameters as follows: specificity, detection capability, applicability and stability. Selectivity/specificity Specificity was assessed by analysis of blank cattle and poultry muscles (n ¼ 20) for each matrix. Detection capability (CC ) Detection capability (CC ) was determined is the concentration at which the method can detect truly contaminated samples with a statistical certainty of 1 (false-compliant results were 5%). In a batch composed by 20 samples, this means a minimum of 19 samples with analyte detection and only one (5%) cannot be detected. In the case of substances with MLR, CC was determined by analysing meat samples spiked at 25% and 50% of MLR level. For each level, 20 samples were spiked and analysed in the LC-MS/ MS system. Stability It is well known that an inadequate or long-time storage for standard solutions may result in degradation products and, consequently, poor responses causing results deviations. Stability must be taken in account during the validation of residue methods. To evaluate the stability of standard solutions, dilutions of stock solutions were prepared with all analytes at 100 ng ml1 in water and acetonitrile (98:2) with 0.1% formic acid, and stored at 20 C (n ¼ 10) and at 4 C (n ¼ 10), and at room temperature in dark vials (n ¼ 10) and at room temperature in normal vials (n ¼ 10). Samples were analysed in the LC-MS/MS system weekly and measured values were compared with those of freshly prepared standard solutions. Results were summarised for 3 months of evaluation.
Results and discussion Sample preparation The aim was to develop a fast and simple screening method for meat samples (cattle and poultry) able to detect the most used veterinary antibacterials. Muscle is composed of fibres, connective tissues, adipose tissue, cartilage and bone (Aerts et al. 1995). Extraction must be adequate to avoid interference
Food Additives and Contaminants of these substances. The procedure described here uses the high specificity and sensitivity of LC-MS/MS to simplify sample preparation. Briefly, extraction was conducted in order to obtain a fast and environmentally friendly protocol. The use of solid disruptors together with centrifugation and low-volume organic solvent provided, respectively, tissue homogenisation, tissue ‘juice’ liberation and protein precipitation. For this last purpose, tests were conducted using acetonitrile and methanol as organic solvent. Despite the propensity of methanol to extract excessive matrix material, this solvent was chosen in the first step, because all analytes were extracted with equivalent recoveries using methanol (Anderson et al. 2005). Ethanol with acetic acid (3%) was tested providing the cleanest extracts compared with procedures using only methanol, but results for tetracyclines were poorer. As the major objective was to disrupt tissues using dispersion with sand to liberate intracellular and interstitial liquid, centrifugation was tested in order to separate liquid containing analytes from solid debris. Increasing the centrifugation time (from 20 to 30 min at 3000 g) was important in providing greater liquid separation. After generic sample preparation with methanol, acidified acetonitrile and methanol (formic acid 0.1%) were also used before centrifugation and the presence of acid provided better results, especially for fluorquinolones. Acetonitrile in this step gave the cleanest extract and showed fewer coextracted endogenous compounds in comparison with methanol. Final centrifugation in the presence of mobile phase gave additional extract clean-up. The chelation of tetracyclines with multivalent cations are an important consideration as there is a high propensity for forming these complexes. In biological matrices divalent ions can interfere with the extraction and disruption of these interactions, which is commonly achieved through addition of ethylenediaminetetraacetic acid (EDTA) (Kawata et al. 1996). EDTA addition is necessary in order to prevent chelates formation by tetracyclines. An ultrasonic bath was considered to improve cell disruption and 10 min was established as an adequate time. The great advantage of the proposed method is the use of sand as an external promoter of interstitial fluids, as a real sample may contain contaminants or drug residues in its interior. The use of EDTA-treated sand was previously reported for the determination of tetracyclines in animal tissues (Blasco et al. 2009). Comparison of two different sands (Merck or sea sand previously purified in our laboratory) demonstrated that both are interchangeable and give similar results. Diatomaceous earth and polyethylene glycol were also tested, but these techniques did not provide a good release of interstitial fluids, necessitating the use of cartridge and filtering material to remove the liquid, which increased analysis costs and the time involved.
511
Extraction with sand, which can be considered as a matrix solid-phase dispersion (MSPD) method, is not followed by elution or packaging in cartridges. As only centrifugation was used to obtain a liquid fraction enriched with analytes, this extraction protocol is not exhaustive. Thus, just a portion of total analyte mass was extracted. In order to determine the recovery, spiked samples were compared with ‘tissue standard’ samples, i.e. sample extracts spiked after completing the extraction process at a level equal to that expected considering no losses through analysis. Recoveries were in the range from 19% to 29%. To achieve higher recoveries, additional steps would need to be applied, but they would decrease the advantages of fast preparation and low cost. Considering this, we chose to validate this method as a qualitative method. In terms of sample preparation, the entire process can be performed for a batch of 25–30 samples in 3 h. Low-cost, low-solvent consumption and speed were factors which were prioritised in the development work. Several multiresidue methods published in recent years are more specific and/or more comprehensive in terms of the number of analytes. Bogialli et al. (2006) developed a MSPD method for tetracyclines analysis using sand as support but followed by a heated water extraction which required a specific apparatus. Yamada et al. (2010) published a multiresidue screening method for 130 veterinary drug residues in bovine, porcine and chicken muscle, but the extraction procedure required 50 ml of acetonitrile:methanol (95:5) and 30 ml of n-hexane per sample. In terms of low cost, the present method is comparable with the method of Chico et al. (2008) which used only 10 ml of methanol:water (70:30) as an extraction solvent. Other multiresidue methods require SPE or to split intermediary extract in aliquots to achieve satisfactory results (Stubbings and Bigwood 2009; Boscher et al. 2010).
Mass spectrometry To achieve maximum sensitivity, mass spectrometry parameters were optimised by direct infusion of standard solution with each analyte in methanol with formic acid (0.1%). A protonated molecular ion [M þ H]þ was selected as a precursor ion for all compounds, and the cone voltage was adjusted to its maximum signal at the first quadrupole of the mass spectrometer. Product ion spectra were recorded at different collision energies to find two most intense transitions for each analyte (Table 1). The ESI(þ) mode was chosen because of its sensitivity to all compounds studied in this work. Identification of individual antibacterials was based on chromatographic retention time, a characteristic qualifier ion and a confirmatory ion (Table 1). For sulfonamides,
512
M.S. Bittencourt et al.
Table 1. Mass spectrometry parameters. Compound (MW, g mol1) Sulfonamides Sulfadiazine (250.2) Sulfadimethoxine (310.2) Sulfapyridine (249.2) Sulfamethoxazole (253.3) Sulfaquinoxaline (299.2) Sulfathiazole (255.32) Sulfamerazine (264.31) Sulfamethazine (277.2) Sulfachlorpyridazine (284.7) Quinolones Sarafloxacin (385.3) Norfloxacin (319.3) Difloxacin (399.9) Ciprofloxacin (331.3) Enrofloxacin (359.3) Oxolinic Acid (261.2) Nalidixic Acid (232.2) Flumequine (261.2) Tetracyclines Doxicycline (444.4) Oxytetracycline (460.4) Tetracycline (444.4) Chlortetracycline (478.0)
CVa (V)
Transition 1 (CE2, eV)
Transition 2 (CE, eV)b
20 35 25 21 35 25 25 20 25
251.25 4 156.1 (15) 311.25 4 156.20 (20) 250.20 4 156.0 (15) 254.30 4 156.10 (20) 301.10 4 156.10 (15) 256.20 4 156.1 (15) 265.30 4 156.0 (15) 279.0 4 156.10 (20) 285.0 4 108 (25)
251.25 4 91.7 (25) 311.25 4 108.0 (25) 250.20 4 92.0 (25) 254.30 4 91.90 (25) 301.10 4 91.90 (30) 256.20 4 92.0 (28) 265.30 4 92.0 (30) 279.0 4 124.0 (20) 285.0 4 156 (15)
40 30 40 35 35 30 22 30
386.40 4 342.30 (20) 320.30 4 276.30 (17) 400.30 4 356.30 (20) 332.20 4 288.0 (17) 360.40 4 316.40 (20) 262.20 4 244.3 (18) 233.4 4 187.3 (26) 262.25 4 202.0 (30)
386.40 4 299.30 (28) 320.30 4 233.30 (20) 400.30 4 299.30 (28) 332.20 4 245.40 (23) 360.40 4 245.30 (25) 262.20 4 160.0 (25) 233.4 4 215.3 (15) 262.25 4 244.25 (20)
35 22 25 30
445.4 4 428.25 (20) 461.25 4 426.25 (20) 445.25 4 410.30 (20) 479.0 4 154.0 (30)
445.4 4 153.9 (30) 461.25 4 443.25 (12) 445.25 4 154.10 (30) 479.0 4 97.5 (40)
Notes: aCone voltage. b Collision energy.
two characteristic fragment ions at m/z 156 and m/z 92 were observed. The former corresponds to the common molecular fragment for all sulfonamides, p-sulfoaniline moiety and the latter corresponds to the loss of sulfonyl group from this structure (Chico et al. 2008). For sulfonamides, transition m/z 108 was also observed and corresponds to a loss of SO from psulfoaniline. Tetracyclines have a structure formed by an octahydrotetracene 2-carboxamide and conventional fragmentation in MS/MS shows a similar fragmentation pattern, where major ions usually obtained correspond to losses of NH3 and H2O or both. Oxytetracycline presented ammonia and water loss and a correspondent m/z transition 426 that was demonstrated to be the most intense in this present method. For chlortetracycline and tetracycline we could observe transition m/z 154 as the most intense (Diaz-Cruz and Barcelo 2005; Petrovic et al. 2005). The most frequent loss for flourquinolones was CO2. For nalidixic acid, oxolinic acid and flumequine (quinolones) the most intense fragments observed were m/z 187, 160 and 202, respectively. The total ion chromatogram (TIC) of 20 antibiotics spiked into a poultry sample at 50 ng ml1 is shown in Figure 1. Results of a typical MRM LC-MS/MS chromatogram of poultry muscle spiked with antimicrobials are illustrated in Figures 2(a) and 2(b). Owing to the diverse characteristics of drugs, especially
Figure 1. Total ion chromatogram of 21 antibacterials for poultry muscle spiked with 50 ng ml1.
polarity, various LC conditions were tested to achieve adequate separation. Methanol with formic acid was previously tested, but acetonitrile was best principally for tetracyclines, like most published methods. The gradient programme was based on a high proportion of aqueous phase in the first 5 min to extend the retention time for analytes to separate interferences from matrix products. In the second step in the gradient programme we used a greater portion of organic solvent (20% and subsequently 50%). However, simultaneous analysis of compounds from different groups with quite different physicochemical
Food Additives and Contaminants
513
Figure 2. (a, b) MRM of 21 antibacterials in poultry muscle at 50 ng ml1.
characteristics requires a compromise in the selection of experimental conditions, which in some cases are not the best conditions for all the analytes studied. These conditions provided elution of all analytes within 17 min. After this time, the mobile phase ratio of A:B was converted to an initial value (i.e., 98:2) to re-equilibrate the column. End-capping of the column on reversed-phase columns was preferred to improve
peak shapes, avoiding interactions with silanol groups, especially for tetracyclines (Anderson et al. 2005). Validation procedure Selectivity/specificity For the specificity study the absence of background peaks with a signal-to-noise ratio 43 at the retention
514
M.S. Bittencourt et al. Table 2. CC and MRL values for sulfonamides, tetracyclines and quinolones.
Analytes
Figure 3. Total ion chromatogram of cattle blank sample.
time of analytes showed that the method was not affected by interference from endogenous compounds. Two transitions choices for each analyte provide specificity for this method. No interference peaks were observed within the retention time window for all substances. A TIC chromatogram for cattle blank samples is shown in Figure 3. Detection capability (CC ) Detection capability (CC ) was established for each analyte as the level in which false-compliant results were 5%. Experimental data showed CC values between 25 and 50 ng g1; the results are given in Table 2, together with the MRLs valued adopted by the Brazilian NRCP. Detection capabilities were considered satisfactory, taking into account that a real sample with an analyte at the MRL will be correctly detected. Just for two compounds (SARA and NALID) were CC values not adequate for poultry samples analysis. Stability The assessment of stability guarantees the use of stock solutions at least for 2 months, when stored at 20 C and for 1 month at 4 C in dark glass bottles. Samples maintained at room temperature, but protected from light, are stable at least for 1 week, and this characteristic is important during analysis. Samples left in transparent vials at room temperature showed some degradation, especially for tetracyclines. A summary of the most relevant stability results is shown in Table 3. Results show that ENRO, TETRA and NALID had higher losses when comparing the storage of standard solutions for 3 months at 20 C. For the particular case of NALID precipitation was observed. For solutions maintained at room temperature in amber and clear vials OXI had the highest losses, with a 31% decrease in amber vials and 84% in clear vials.
Sulfadimethoxine Sulfaquinoxaline Sulfadiazine Sulfachlorpyridazine Sulfathiazole Sulfapyridine Sulfamerazine Sulfamethoxazole Sulfamethazine Doxicycline Chlortetracycline Tetracycline Oxytetracycline Oxolinic acid Nalidixic acid Flumequine Ciprofloxacin Difloxacin Enrofloxacin Norfloxacin Sarafloxacin
MRL (mg kg1)a
CC (mg kg1), cattle muscle
CC (mg kg1), poultry muscle
100 100 100 100 100 100 100 100 100 100 200 200 200 100b 20b 500 100 100 100 100 202
25 25 25 25 25 25 25 50 25 50 50 50 50 25 25 25 50 50 25 25 25
25 25 25 25 25 25 25 50 25 50 50 50 50 25 25 25 50 50 25 25 25
Notes: aMRL values were adopted by the Brazil National Residues Control Plan (BRASIL 2010). b Adopted only for poultry muscle.
Tetracyclines are well known to form 4-epimers, which results in double peaks in the chromatograms. Generally, MRLs for tetracyclines are given as the sum of the parent compound and 4-epimers plus metabolites. In the present method, mild extraction conditions do not lead to the formation of 4-epimers, probably because extracts are not exposed to high temperatures or extreme pHs, which are the usual conditions in several extraction methods. The stability of the extracts was not assessed, since the maximum storage period for these samples is stipulated in the NRCP that requires a maximum of 15 days for the receipt, analysis and reporting of results (Blasco et al. 2009; Hoff et al. 2009). Ruggedness and method applicability The applicability and robustness of the described multiresidue LC-ESI-MS/MS are defined as the susceptibility of an analytical method to changes in experimental conditions, either minor changes such as solvent and reagents from different batches or major changes such as operator and matrix (Cherlet et al. 2003). Different batches of solvents and reagents during this period of tests did not cause interferences in the present study. Different analysts have performed sample extraction without affecting the results.
Food Additives and Contaminants
515
Table 3. Stability data for some analytes (just analytes with significant alterations are shown). Signal loss (peak area, %) Analyte
RS1 versus RS2a
ART1 versus ART2b
CRT1 versus CRT2c
RS versus ART
RS versus CRT
ART versus CRT
ENRO TETRA NALID SMA SDZ SARA SCP OXI FLU SQX SMZ CIPRO STZ CLOR SDMX SMR SDX DOXI
11 22 39e n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
n.d.d n.d. n.d. 11 14 14 23 31 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d.
n.d. 11 11 13 12 n.d. n.d. 84 12 14 14 16 24 n.d. n.d. n.d. n.d. n.d.
n.d. 45 62 37 71 41 46 95 n.d. 39 32 22 34 44 15 35 16 23
n.d. 57 74 39 67 25 37 99 30 46 27 30 43 60 29 39 39 44
n.d. 22 32 n.d. n.d. n.d. 16 74 32 11 n.d. n.d. 13 29 17 n.d. 27 27
Notes: Solutions assigned as ‘1’ were freshly prepared and those assigned as ‘2’ were prepared 3 months later. a RS ¼ refrigerated standard solution (20 C). b ART ¼ amber vial at room temperature. c CRT ¼ clear vial at room temperature. d n.d., No difference; loss or gain lower than 10%. e Precipitation occurrence under storage.
The method was applied to samples from different producers, demonstrating its suitability. The method was successfully applied to the analysis of samples of poultry (n ¼ 26) and cattle muscle (n ¼ 21), including analysis for a considerable number of samples in 1 day (n ¼ 30). Two incurred poultry samples were included and quantified in specific methods for ENRO and SQX (Figures 4a and b). In addition, a proficiency test for sulfonamides in cattle muscle was performed including screening and quantitative methods (Progetto Trieste 2010, 2nd Round, Veterinary Drug Residues). Sulfadimethoxine was correctly detected by this present method and calculated as containing 59.0 mg kg1. The z-score obtained in this test was 0.35 (satisfactory values are from 2 to þ2). Moreover, the method was recently accredited under ISO 17025 by the National Institute of Metrology, Standardization and Industrial Quality (INMETRO) under CRL 0384 (accreditation certificate).
Conclusions An LC-MS/MS method for the screening of 20 veterinary drug residues in meat samples was developed and validated. Although several multiclass methods for veterinary drugs have been published in recent years, in many cases the sample preparation
Figure 4. MRM of incurred poultry samples with sulfaquinoxaline and enrofloxacin.
is complex and laborious. In this paper a novel MSPDlike method was developed, using sand and some millilitres of organic solvents, to provide a fast, very cheap and environmentally friendly protocol, which was successfully applied to naturally incurred samples. This method was also applied to official analysis, and screenings results obtained for proficiency material were in agreement with the results from quantitative
516
M.S. Bittencourt et al.
and confirmatory analysis. Recently, the present method was recommended for accreditation under ISO 17025 by the Brazilian Accreditation Body (INMETRO). The method is capable of satisfactory use for meat samples containing sulfonamides, tetracyclines and fluorquinolones, and is an important tool in the Brazilian National Residue Control Plan to monitoring antibacterial residues in cattle and poultry meat samples.
References Aerts MML, Hogenboon AC, Brinkman UAT. 1995. Analytical strategies for the screening of veterinary drugs and their residues in edible products. J Chromatogr B. 667:1–40. Anderson CR, Rupp HS, Wu WH. 2005. Complexities in tetracyclines analysis – chemistry, matrix extraction, cleanup, and liquid chromatography. J Chromatogr A. 1075:23–32. Blasco C, Di Corcia A, Pico Y. 2009. Determination of tetracyclines in multi-species animal tissues by pressurized liquid extraction and liquid chromatography-tandem mass spectrometry. Food Chem. 116:1005–1012. Bogialli S, Curini R, Di Corcia A, Lagana A, Rizzuti G. 2006. A rapid confirmatory method for analyzing tetracycline antibiotics in bovine, swine and poultry muscle tissues: matrix solid-phase dispersion with heated water as extractant followed by liquid chromatography-tandem mass spectrometry. J Agric Food Chem. 54:1564–1570. Bogialli S, Di Corcia A. 2009. Recent applications of liquid chromatography-mass spectrometry to residue analysis of antimicrobial in food of animal origin. Anal Bioanal Chem. 395:947–966. Boscher A, Guignard C, Pellet T, Hoffmann L, Bohn T. 2010. Development of a multi-class method for the quantification of veterinary drug residues in feedingstuffs by liquid chromatography-tandem mass spectrometry. J Chromatogr A. 1217:6394–6404. BRASIL. 2010. Ministe´rio da Agricultura, Pecua´ria e Abastecimento. Secretaria de Defesa Agropecua´ria. Instruc¸a˜o Normativa no. 08/2010. Carretero V, Blasco C, Pico´ Y. 2008. Multi-class determination of antimicrobials in meat by pressurized liquid extraction and liquid chromatography-tandem mass spectrometry. J Chromatogr A. 1209:162–173. Cherlet M, Schelkens M, Croubels S, Backer P. 2003. Quantitative multi-residue analysis of tetracyclines and their 4-epimers in pig tissues by high-performance liquid chromatography combined with positive-ion electrospray ionization mass spectrometry. Anal Chim Acta. 492:199–213. Chico J, Ru´bies A, Centrich F, Companyo´ R, Prat MD, Granados M. 2008. High-throughput multiclass method for antibiotic residue analysis by liquid chromatographytandem mass spectrometry. J Chromatogr A. 1213:189–199. Diaz-Cruz MS, Barcelo D. 2005. LC-MS2 trace analysis of antimicrobials in water, sediment and soil. Trends Anal Chem. 24:645–657.
European Commission. 2002. Commission Decision 2002/657/EC. Implementing Council Directive 96/23/EC concerning the performance of analytical methods and interpretation of results. Off J Eur Comm. L221:8–36. Gaugain-Juhel M, De´lepine B, Gautier S, Fourmond MP, Gaudin V, Hurtaud-Pessel D, Verdon E, Sanders P. 2009. Validation of a liquid chromatography-tandem mass spectrometry screening method to monitor 58 antibiotics in milk: a qualitative approach. Food Addit Contam. 26(11):1459–1471. Granelli K, Branzell C. 2007. Rapid multi-residue screening of antibiotics in muscle and kidney by liquid chromatography-electrospray ionization-tandem mass spectrometry. Anal Chim Acta. 586:289–295. Granelli K, Elgerud C, Lundstrom A, Ohlsson A, Sjoberg P. 2009. Rapid multi-residue analysis of antibiotics in muscle by liquid chromatography-tandem mass spectrometry. Anal Chim Acta. 637:87–91. Hoff RB, Barreto F, Kist TBL. 2009. Use of capillary electrophoresis with laser-induced fluorescence detection to screen and liquid chromatography-tandem mass spectrometry to confirm sulfonamide residues: validation according to European Union 2002/657/EC. J Chromatogr A. 1216:8254–8261. Kawata S, Sato K, Nishikawa Y, Iwama K. 1996. Liquidchromatographic determination of oxytetracycline in swine tissues. J. AOAC Int. 79(6):1463–1465. Martos PA, Jayasundara F, Dolbeer J, Jin W, Spilsbury L, Mitchell M, Varilla C, Shurmer B. 2010. Multiclass, multiresidue drug analysis, including aminoglycosides, in animal tissue using liquid chromatography coupled to tandem mass spectrometry. J Agric Food Chem. 58(10): 5932–5944. Mauricio AQ, Lins ES, Alvarenga MB. 2009. A national residue control plan from the analytical perspective – the Brazilian case. Anal Chim Acta. 637:333–336. Ortelli D, Cognard E, Jan P, Edder P. 2009. Comprehensive fast multiresidue screening of 150 veterinary drugs in milk by ultra-performance liquid chromatography coupled to time of flight mass spectrometry. J Chromatogr B. 877:2363–2374. Petrovic M, Hernando MD, Diaz-Cruz MS, Barcelo D. 2005. Liquid chromatography-tandem mass spectrometry for the analysis of pharmaceutical residues in environmental samples: a review. J Chromatogr A. 1067:1–14. Stolker AAM, Zuidema T, Nielen MWF. 2007. Residue analysis of veterinary drugs and growth-promoting agents. Trends Anal Chem. 26:967–979. Stubbings G, Bigwood T. 2009. The development and validation of a multiclass liquid chromatography tandem mass spectrometry (LC-MS/MS) procedure for the determination of veterinary drug residues in animal tissue using a QuEChERS (quick, easy, cheap, effective, rugged and safe) approach. Anal Chim Acta. 637:68–78. USP. 1995. USP 23 – The United States pharmacopeia. 23rd ed. Rockville (MD): USP. Yamada R, Kozono M, Ohmori T, Morimatsu F, Kitayama M. 2010. Simultaneous determination of residual veterinary drugs in bovine, porcine and chicken muscle using liquid chromatography coupled with electrospray ionization tandem mass spectrometry. Biosci Biotechnol Biochem. 70:54–65.
Food Additives and Contaminants Vol. 29, No. 4, April 2012, 517–525
Validation of a quantitative and confirmatory method for residue analysis of aminoglycoside antibiotics in poultry, bovine, equine and swine kidney through liquid chromatography-tandem mass spectrometry M.P. Almeida*, C.P. Rezende, L.F. Souza and R.B. Brito Ministry of Agriculture, Livestock and Food Supply (MAPA), Agricultural National Laboratory, Lanagro/MG, Brazil (Received 23 November 2010; final version received 31 August 2011) The use of aminoglycoside antibiotics in food animals is approved in Brazil. Accordingly, Brazilian food safety legislation sets maximum levels for these drugs in tissues from these animals in an effort to guarantee that food safety is not compromised. Aiming to monitor the levels of these drugs in tissues from food animals, the validation of a quantitative, confirmatory method for the detection of residues of 10 aminoglycosides antibiotics in poultry, swine, equine and bovine kidney, with extraction using a solid phase and detection and quantification by LC-MS/MS was performed. The procedure is an adaptation of the US Department of Agriculture, Food Safety and Inspection Service (USDA-FSIS) qualitative method, with the inclusion of additional clean-up and quantification at lower levels, which proved more efficient. Extraction was performed using a phosphate buffer containing trifluoroacetic acid followed by neutralization, purification on a cationic exchange SPE cartridge, with elution with methanol/acetic acid, evaporation, and dilution in ion-pair solvent. The method was validated according to the criteria and requirements of the European Commission Decision 2002/657/EC, showing selectivity with no matrix interference. Linearity was established for all analytes using the method of weighted minimum squares. CC and CC varied between 1036 and 12,293 mg kg1, and between 1073 and 14,588 mg kg1, respectively. The limits of quantification varied between 27 and 688 mg kg1. The values of recovery for all analytes in poultry kidney, fortified in the range of 500–1500 mg kg1, were higher than 90%, and the relative standard deviations were lower than 15%, except spectinomycin (21.8%). Uncertainty was estimated using a simplified methodology of ‘bottom-up’ and ‘top-down’ strategies. The results showed that this method is effective for the quantification and confirmation of aminoglycoside residues and could be used by the Brazilian programme of residue control. Keywords: animal products – meat; veterinary drug residues – antibiotics; chromatography – LC/MS
Introduction The aminoglycosides (AMGs) are antibiotics widely used in veterinary medicine due to their efficacy against Gram-negative bacilli and their positive synergism with other antibiotics in treating infections by Grampositive agents. These antibiotics have their use approved in Brazil. Their use in the country, however, is controlled as part of a veterinary drugs residue control programme for poultry, swine, equine and bovine (Brazil 1999) in an effort to keep the residues at levels considered safe for the consumers. The AMGs are stable at pH 6–8, highly soluble in water, have a cationic polar structure, are widely used in the treatment of respiratory and enteric bacterial infections, but should be used with criteria due to its nephrotoxic activity in humans (Kennedy et al. 1998; Bogialli et al. 2005; Oliveira et al. 2006). Figure 1 shows the chemical structures of the AMGs. Various
methods for the analysis of AMGs in biological samples are described in the literature, including indirect ultraviolet (UV) techniques, fluorescence and GC-MS/MS with derivatisation. The latter has the disadvantages of being time-consuming, with instability of the products and the generation of sub-products in the reaction. LC-MS/MS does not need the derivatisation step, provides more sensitivity and chromatographic efficiency, and has the possibility of confirmation (McGlinchey 2008; Zhu et al. 2008). The AMGs have a strong polar character, resulting in a poor retention in reversed-phase columns, low resolution and ion suppression due to co-elution of matrix components. To increase retention and separation one can use an ionic pair reagent (McLaughlin et al. 1994). Heptafluorobutyric acid (HFBA) is a common ionic pair reagent used in AMG analysis (Niessen 1998; Li et al. 2009).
*Corresponding author. Email:
[email protected] ISSN 1944–0049 print/ISSN 1944–0057 online ß 2012 Taylor & Francis http://dx.doi.org/10.1080/19440049.2011.623681 http://www.tandfonline.com
518
M.P. Almeida et al.
Figure 1. Molecular structures of the aminoglycosides.
The objective of this study was the validation of a quantitative and confirmatory multi-residue sampling method for the analysis of spectinomycin, hygromycin, streptomycin, dihydrostreptomycin, amikacyn, kanamycin, apramycin, tobramycin, gentamicin and neomycin in samples of swine kidney with solid-phase extraction (SPE) and detection/quantification using LC-MS/MS with the aim to comply with The National Control Plan of Residues and Contaminants (PNCRC) of Brazil. The method used as reference was obtained from the US Food Safety and Inspection Service
(FSIS-USDA). The original qualitative method was changed into a quantitative method and its validation was performed according to the European Commission 2002/657/EC (2002) criteria. Validation was subsequently extended to samples of poultry, equine, bovine kidney through the evaluation of inter-species matrix effects and comparison of linearity and CC and CC values obtained for each species. The uncertainty of the method was estimated from linearity and precision data using a simplified methodology composition the strategies of ‘bottom-up’ and ‘top-down’.
519
Food Additives and Contaminants Materials and methods
Table 1. Gradient of mobile phase HPLC.
Chemicals and reagents Reference spectinomycin, hygromycin, dihydrostreptomycin, amikacyn, kanamycin, apramycin, tobramycin, gentamicin and neomycin standards with a minimum purity of 68% were obtained from SigmaAldrich (St Louis, MO, USA). A reference, streptomycin standard with purity of 99% was obtained from Riedel (Brazil). The following reagents were used: heptafluorobutyric acid (HFBA) (Fluka, Brazil), trichloroacetic acid (TCA) (Isofar, Brazil), acetic acid (HOAc) (Acro´s Organics, Belgium), disodium of ethylenediaminetetraacetic acid (Na2EDTA) (Sigma), potassium phosphate monobasic (KH2PO4) (Sigma), and HPLC-grade methanol (Tedia, USA). The water was purified in a Milli-Q Gradient (Millipore, Brazil) system. A solid-phase extraction (SPE) cartridge BakerBond SPE Wide Pore CBX, 500 mg, 6 ml was supplied by J.T. Baker (USA).
Standard solutions Stock solutions of aminoglycosides standards at concentration of 200 mg ml1 diluted in distilled deionised water were prepared and stored in freezer at 23 >23 >23 >23 >22 >22 >22 >22 >8 >8
1 2 3 4 5 6 7 8 Blank Acetic acid
>22 >22 >23 >23 >22 >22 >24 >24 >8 >8
1 2 3 4 5 6 7 8 Blank Acetic acid
Figure 3. Chromatogram of poultry kidney juice (incurred sample) with enrofloxacin at 9.78 mg ml1 and ciprofloxacin at 180.8 mg ml1.
production from Geobacillus metabolism, occurring only when no inhibitory substance is present. This is visible by a colour change (purple to yellow) in the indicator, bromocresol purple, added to the solid agar medium. When antibacterial compounds are present above the limit of detection (LOD), no growth occurs and the colour remains purple. Aiming to improve method sensitivity by sample concentration, improved protocols suggested by the manufacturer were tested. A extraction with acetonitrile/acetone (70:30, v/v) was performed. Briefly, 2 0.1 g of sample (muscle or kidney) was homogenised, extracted with organic solvent, centrifuged and the supernatant evaporated in a water-bath at 40–45 C
under a gently flow of nitrogen. The dry residue was reconstituted in sterile BHI broth (200 ml) and applied to the ampoules. Kidney, muscle and liver of cattle and swine were analysed using the PremiÕ Test. Preliminary results were unsatisfactory, since all results were positive. This find lead us to consider solvent interference, which caused a false-positive response. However, as these samples, assumed to be blank samples, were not analysed for a complete antibiotics profile (just for sulfonamides, tetracyclines, fluorquinolones and quinolones), no conclusions could be drawn. Indeed, the concentration protocol adds several steps, such as extraction, homogenisation, centrifugation and evaporation under nitrogen. This increases analysis complexity, taking 3–5 h just to prepare samples prior to testing (this already takes 3.5 h for incubation and colour reading). Changes in colour were sometimes difficult to interpret, which may be reflected in the number of false positive results obtained. In addition, incubation time for the PremiÕ Test appeared to be important, as continued incubation at 64 C past the time at which a negative control turned yellow could lead, eventually, to positive samples turning yellow (negative) as well. To avoid subjective interpretation by visual inspection, results were obtained using a scanner coupled with DSM interpretation software, which gives an assigned numerical z-value for each colour based in an
584
R. Hoff et al.
Table 6. Comparative data for blank samples obtained from visual and scanner analysis. The z-value is a value given by the DSM software that use an algorithm based on negative values (0.0) for negative samples. Description
z-value
Scanner
Visual
A1 B1 C1 D1 E1 E2
2.27 2.23 2.31 5.89 9.91 0.28
NEG POS POS POS POS NEG
POS POS POS POS POS NEG
Bovine liver Bovine kidney Bovine liver Swine kidney Poultry kidney Blank
algorithm. Values above zero indicate a negative result. Blank samples of bovine, poultry and swine liver and/or kidney, previously analysed by LC–MS/MS methods for sulfonamides, tetracyclines, fluorquinolones and quinolones, were applied to test several false positive results (Table 6). Again, an unknown inhibitor substance present in the samples (endogenous or exogenous) could be the reason. Cantwell and O’Keeffe (2006) found that the PremiÕ Test was rugged with respect to species, ampoule age and ampoule batch. In disagreement with this work, we found that the PremiÕ Test and FAST showed a high species and tissue dependence, varying performance between species and even between distinct tissues within the same species.
In-house developed tube tests E. coli tube test The E. coli tube test was designed to screen fluorquinolones in kidney. For this purpose, eight analytes were tested (NOR, CIPRO, ENRO, NALID, OXO, FLU, DIFLO and SARA). Considering just aqueous solutions from each fluorquinolone, E. coli was able to detect CIPRO and ENRO at a concentration of 0.25 MRL. NOR, NALID, OXO, FLU, DIFLO and SARA produced positive results at a concentration of 0.5 MRL. When spiked kidney was analysed, the results show a strong matrix effect. Here, the influence of agar pH was evaluated. The medium was tested at three different pH values (6.0, 7.0 and 8.0). Enrofloxacin and ciprofloxacin were correctly detected in bovine kidney and muscle at pH 7.0. Difloxacin was positive at pH 6.0 and 7.0. The other analytes, although detected when in solution, were not detected when in the matrix. To test specificity, eight vials were incubated with an aqueous solution of sulfadiazine, penicillin G, tetracycline, amoxicillin, neomycin, streptomycin and erythromycin at concentrations of 0.5, 1.0 and
1.5 MRL. All results were negative, with the exception of neomycin, which was detected at all tested concentrations. S. aureus tube test The tube test with S. aureus was developed to deal with other classes of antibiotics. Drugs tested were -lactams (ampicillin, amoxicillin, penicillin G, penicillin V, cephalexin), aminoglycosides (streptomycin and neomycin), macrolides (erythromycin and tylosin), tetracyclines (tetracycline) and sulfonamides (sulfadiazine). For LOD determination using separate aqueous solution of each antibiotic, S. aureus was capable of detecting at the 0.5 MRL level in bovine kidney for all analytes tested. However, when tested in a spiked matrix of bovine kidney, it was unable to detect any antibiotic at any of the three deferent pH values. Despite the low cost and speed of preparation and analysis, both in-house-developed tube tests had some disadvantages. Firstly, as E. coli and S. aureus do not produce spores, the tube tests have a short shelf-life and must be used within 48 h after preparation, preferentially. Vials must be maintained under refrigeration until the moment of analysis. Another serious drawback was observed when we tested other matrices or other species. For milk, results were negative for all testes substances. For poultry muscle and kidney, the tube tests did not reproduce the same results obtained for bovine muscle and kidney.
Conclusions Screening kits improve a laboratory’s operational capability with reduced costs. For a large number of samples, only positive samples are submitted to specific confirmation analysis. There are different kit methodologies for food analysis, which are specific for certain veterinary drugs or for multi-class analysis. These kits should be used according to laboratory demands and must be precise. The objective of this paper was to compare the performance of various kits available on the market for veterinary drugs analysis in edible tissues and to evaluate the possibility of developing screening methods ‘‘in loco’’ for the same purpose. The results, associated with kit costs, will serve to select the best method to be implemented and validated in MAPA laboratories. A general comparison between all discussed methods are presented in Table 7. In this study, we evaluated two well-established methods based on the bioactivity of antibiotics: FAST and PremiÕ Test. A new, in-house-developed method using E. coli or S. aureus as sensitive microorganism was also presented. LC–MS/MS multi-residue and multi-class analysis was also performed and compared with the bioactivity-based screening methods. In contrast to some literature reports, we found that
Food Additives and Contaminants
585
Table 7. General comparison of bioactiviy-based screening methods and LC–MS/MS.
Criteria Sample preparation (time and difficult/easy)
Time of analisys Costs (low, medium or high)
Results confidence
FAST
Premi-Test
20 min for sample preparation; requires asseptic work and microbiology expertise Results in 18 h Sample preparation – medium Analytical system – not required Non-specific, non-selective
2 h for sample preparation; easily performed
Results interpretation
Visual inspection, subjective
General
Complex and laborious (spores suspension standardisation and process control)
Results in 6 h Sample preparation – high Analytical system – not required Non-specific, medium selectivity if use inhibitors (PABA or -lactamase) Visual inspection or optical lecture (scanner), subjective Commercial available
the MIC values for several target analytes were above the MRL values (Table 1). Thus, the screening method is discarded if it is not fit-for-purpose or concentration steps (solid–liquid extraction, organic solvent extraction and posterior concentration) are applied to achieve adequate detection capability. False-positive results were found in a high number of cases for all bioactivity-based methods and several causes were identified. Physiological pH differences between species, endogenous inhibitory substances and the matrix were found to be major causes. In comparison with LC–MS/MS screening methods, bioactivity-based methods are less expensive in terms of instrumental techniques. However, considering that every positive sample in a microbiological test must be analysed by a confirmatory methods, the use of such methods that can provide conclusive results, even for screening, may be the most rational choice. For laboratories that already have LC–MS/MS systems installed, direct analysis in these systems eliminates most false-positive results, permits metabolite identification, and provides qualitative results. Moreover, LC–MS/MS screening methods can be converted into semi-quantitative or quantitative methods. Notwithstanding, for antibiotics with high MRL values, such as aminoglycosides and some macrolides, bioactivity-based methods seems to be an appropriate approach, since these classes of antibiotics can be easily detected at values below the MRL, even without any pre-concentration strategy. This does not eliminate the high potential for
E. coli/S. aureus tube test
LC–MS/MS multiclass
2 h for sample preparation; requires asseptic work and microbiology expertise Results in 4 h Sample preparation – low Analytical system – not required Non-specific, nonselective
1 h for a batch preparation; easily performed
Visual inspection or optical lecture (scanner), subjective Low stability
15 min for each run Sample preparation – low Analytical system – high High selectivity and specificity
Confirmatory and objective Requires high technology and specialised technicians
false-positive results. For large scale monitoring programmes, fast responses and high confidence levels can be obtained using rapid screening methods based on mass spectrometry, followed by quantification of analytes via class-specific methods. Currently, this is the strategy applied in our laboratory for routine analysis in the National Residues and Contaminants Control Plan. References Althaus R, Berruga MI, Montero A, Roca M, Molina MP. 2009. Evaluation of a microbiological multi-residue system on the detection of antibacterial substances in ewe milk. Anal Chim Acta. 632:156–162. Berendsen BJA, Pikkemaat MG, Stolker LAM. 2011. Are antibiotic screening approaches sufficiently adequate? A proficiency test. Anal Chim Acta. 685:170–175. Bittencourt MS, Martins MT, de Albuquerque FGS, Barreto F, Hoff R. 2011. High-throughput multi-class screening method for antibiotic residue analysis in meat using liquid chromatography–tandem mass spectrometry: a novel minimum sample preparation procedure. Food Addit Contam A. DOI:10.1080/19440049.2011.606228. Boove TFH, Pikkemaat MG. 2009. Bioactivity-based screening of antibiotics and hormones. J Chromatogr A. 1216:8035–8050. Cantwell H, O’Keeffe M. 2006. Evaluation of the PremiÕ Test and comparison with the One-Plate Test for the detection of antimicrobials in kidney. Food Addit Contam A. 23(2):120–125.
586
R. Hoff et al.
Dey BP, White CA, Reamer RH, Thaker NH. 1998. Detection of antimicrobial residues in meat and poultry tissue by screen tests. In: USDA/FSIS Microbiology laboratory guidebook. Chap. 33. 3rd ed. Washington (DC): United States Department of Agriculture. p. 33-1–33-57. European Commission. 1990. Council Regulation (EEC) No. 2377/90 of 26 June 1990: laying down a Community procedure for the establishment of maximum residue limits of veterinary medicinal products in foodstuffs of animal origin. Off J Eur Commun. L224:1–8. European Commission. 2002. Commission Decision 2002/657/EC of 12 August 2002: implementing Council Directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Off J Eur Commun. L221:8–36. European Commission. 2010. Commission Regulation 37/2010 of 22 December 2009: on pharmacologically active substances and their classification regarding maximum residue limits in foodstuffs of animal origin. Off J Eur Commun. L15:1–72. Kinsella B, O’Mahony J, Malone E, Moloney M, Cantwell H, Furey A, Danaher M. 2009. Current trends in sample preparation for growth promoter and veterinary drug residue analysis. J Chromatogr A. 1216:7977–8015. Korpela MT, Kurittu JS, Karvinen JT, Karp MT. 1998. A recombinant Escherichia coli sensor strain for the detection of tetracyclines. Anal Chem. 70. 4457–4462. Mauricio AQ, Lins ES, Alvarenga MB. 2009. A national residue control plan from the analytical perspective. The Brazilian case. Anal Chim Acta. 637(1–2):333–336.
Pengov A, Kirbis A. 2009. Risks of antibiotic residues in milk following intramammary and intramuscular treatments in dairy sheep. Anal Chim Acta. 637:13–17. Pikkemaat MG, Rapallini ML, Karp MT, Elferink JWA. 2010. Application of a luminescent bacterial biosensor for the detection of tetracyclines in routine analysis of poultry muscle samples. Food Addit Contam A. 27(8):1112–1117. Pikkemaat MG, Rapallini ML, van Dijk SO, Elferink JWA. 2008. Comparison of three microbial screening methods for antibiotics using routine monitoring samples. Anal Chim Acta. 637:298–304. Pikkemaat MG, van Dijk SO, Schouten J, Rapallini ML, Kortenhoeven L, van Egmond HJ. 2008. Nouws antibiotic test: validation of a post-screening method for antibiotic residues in kidney. Food Control. 20:771–777. Schneider MJ, Lehotay SJ. 2008. A comparison of the FAST, Premi and KIS tests for screening antibiotic residues in beef kidney juice and serum. Anal Bioanal Chem. 390:1775–1779. Schneider MJ, Mastovska K, Lehotay SJ, Lightfield A, Kinsella B, Shultz C. 2009. Comparison of screening methods for antibiotics in beef kidney juice and serum. Anal Chim Acta. 637:290–297. Stead SL, Sharman M, Stark J, Geijp EML. 2009. Improvements to the screening of antimicrobial drug residues in food by the use of the PremiTestÕ . J Chromatogr A. 1216:8035–8050. Zvirdauskiene R, Salomskiene J. 2007. An evaluation of different microbial and rapid tests for determining inhibitors in milk. Food Control. 18:541–547.
Food Additives and Contaminants Vol. 29, No. 4, April 2012, 587–595
Optimisation and validation of a quantitative and confirmatory method for residues of macrolide antibiotics and lincomycin in kidney by liquid chromatography coupled to mass spectrometry C.P. Rezende*, L.F. Souza, M.P. Almeida, P.G. Dias, M.H. Diniz and J.C. Garcia Ministry of Agriculture, Livestock and Food Suply – MAPA, Laborato´rio Nacional Agropecua´rio – LANAGRO/MG, Brazil (Received 22 November 2010; final version received 16 December 2011) A solid phase extraction followed by a liquid chromatography (LC)-tandem mass spectrometry (MS/MS) detection method for the confirmatory analysis of lincomycin (LIN), clindamycin (CLI), tilmicosin (TIM), erythromycin (ERI) and tylosin (TYL) residues in kidney were optimised and validated for monitoring and controlling the use of these antibiotics in food producing-animals. The method optimisation was carried out by testing changes in the extraction buffer pH and in the ammonium/acetonitrile concentrations on SPE eluent solutions. The optimised extraction procedure involved the extraction of the analytes with a pH 8 phosphate buffer, clean-up on a reversed-phase mixed-cation exchange cartridge, followed by the elution of the analytes in a 98:2 acetonitrile/ammonia solution, concentration in air flow and re-dissolved with an 1:1 methanol/water solution. The analytes were detected in an LC-MS/MS system in electrospray positive ionisation mode. The validation was performed according to the European Commission Decision 2002/657/EC. Linearity was established for all analytes using the method of least weighted squares and CC values ranged from 5.3% to 21.1% higher than the minimum residue limit (MRL) values. The addition levels varied from 0.5 to 1.50 MRL for all analytes, with recoveries exceeding 92.5%. The relative standard deviations (RSD%) in terms of repeatability (n ¼ 54) and reproducibility (n ¼ 108) for all analytes were less than 21.6% and 21.4%, respectively. The uncertainties were calculated by simplified methods using the calibration curve uncertainty and the intermediate precision to obtain the combined measurement uncertainty. The results of the validation process demonstrated that this method is suitable for the quantification and confirmation of antibiotic residues for the Brazilian Residue and Contaminant Control Plan (PNCR). Keywords: macrolides; kidney; lincomycin; validation; LC-MS/MS; SPE
Introduction Methods for the quantification and confirmation of residues are very useful tools to ensure the safety of animal products consumed domestically and in demanding foreign markets all over the world. Thus, monitoring antibiotic residues in the food supply chain plays an important role in the field of veterinary medicine. Macrolides are broad-spectrum antibiotics widely used in veterinary medicine for the treatment of respiratory and enteric infections in cattle, sheep, pigs and poultry. These compounds are effective against gram-positive and some gram-negative bacteria, as well as against members of the group of Chlamydia (Berrada et al. 2007; McGlinchey et al. 2008). Lincomycin is an antibiotic of the lincosamide group used to control certain gram-positive bacteria and exerts its antibacterial action by inhibiting RNA-dependent protein synthesis by acting on the 50 S subunit of the ribosome, used in monopreparations and combined with other antibiotics
such as spectinomycin, sulfadimidine and gentamicin, orally, intra-muscularly or sometimes in the feed or drinking water (EMEA 2008). Incorrect use of these antibiotics may leave residues in edible tissues, causing toxic effects to consumers such as allergic reactions, or problems due to the induction of resistant strains of bacteria (Moats 1996 cited by Draisci et al. 2001). The chemical structures of some macrolides are shown in Figure 1 (Codony et al. 2002). There are several detection methods for macrolides and lincomycin in animal tissues. Microbiological assays are used to screen samples, although they are time-consuming, with poor specificity and selectivity, leading to false-positive results (Granelli et al. 2009). More specific methods are used, normally LC methods in combination with many kinds of detectors like UV, fluorometric, chemiluminescent and electrochemical. The increasing use of highly selective techniques such as mass spectrometry (MS), tandem mass spectrometry (MS/MS) (Adams et al. 2009) and time-of-flight mass spectrometry (TOF/MS) (Peters et al. 2009), coupled with advances in chromatographic technology,
*Corresponding author. Email:
[email protected] ISSN 1944–0049 print/ISSN 1944–0057 online ß 2012 MAPA – BRASIL http://dx.doi.org/10.1080/19440049.2011.652196 http://www.tandfonline.com
588
C.P. Rezende et al.
Figure 1. Chemical structures of macrolides: clindamycin, erythromycin, lincomycin, tylosin and tilmicosin.
have made it possible to develop multi-residue methodologies covering many trace contaminants (Stubbings and Bigwood 2009). MS/MS has the advantage of simpler extraction procedures, higher sensitivity and efficiency (Draisci et al. 2001), although the complexity of tissue matrices in terms of content and analyte-tissue interactions must be considered in the development and optimisation of methods for the analysis of veterinary drug residues. Extracting antibiotics from kidney and liver is a critical step in this method because of the interferences caused by the high protein and fat content of these matrices. The matrix effect is also important in antibiotic recoveries. Macrolide extraction involves protein and fat removal by organic solvents and preconcentration in SPE cartridges (Berrada et al. 2007). A simple method for both confirmatory and quantitative multi-residue analysis of lincomycin (LIN), clindamycin (CLI), tilmicosin (TIM), erythromycin (ERI) and tylosin (TYL) residues in bovine kidney samples has been optimised and validated for the Brazilian Residue and Contaminant Control Plan (PNCR). The antibiotics have been extracted with a pH 8.0 phosphate buffer, clean-up on reversed-phase mixed-cation exchange and detection/quantification by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Method optimisation has been carried out by single comparisons among the original and optimised extraction steps. Instrumental parameters
have been validated according to the European Union Commission Decision 2002/657/EC with a subsequent extension of the validation for bovine, horse and chicken matrices, by evaluating the matrix effects, linearity, CC and CC intercomparison for each species. The measurement uncertainty of the method has been estimated from the linearity and precision data using a simplified methodology in accordance with the requirements of ISO/IEC 17025:2005.
Material and methods Reagents and chemicals Tylosin tartrate (mixed isomers), clindamycin hydrochloride, lincomycin hydrochloride, tilmicosin and erythromycin with a minimum purity of 99.5% were obtained from Sigma (St Louis, MO, USA). Acetonitrile, methanol and hexane were HPLC grade and supplied by Tedia (Farfield, USA). Formic acid (98%) was mass spectrometry grade and supplied by Fluka. Potassium dihydrogen phosphate (KH2PO4) and mono-hydrogen potassium phosphate (K2HPO4) were reagent grade and supplied by Sigma (St Louis, MO, USA). Ammonium hydroxide solution (30%) was reagent grade and obtained from Vetec (Rio de Janeiro, Brazil). All water was filtered in a Milli-Q Gradient system (Millipore Corporation, Billerica, MA, USA).
Food Additives and Contaminants
589
Table 1. MS/MS parameters for the determination of LIN, CLI, ERI, TIM and TYL. Analyte
Precursor ion (m/z)
Product ions (m/z)
Cone voltage (V)
Collision energy (eV)
Dwell time (s)
Lincomycin
407.5
30
869.6
Clindamycin
425.5
Erythromycin
734.6
Tylosin
916.6
23 19 40 42 40 27 22 27 17 40 27
0.05
Tilmicosin
126 359 132 174 696 126 377 158 576 174 772
35 25 30 45
0.05 0.05 0.05 0.05
Note: Subscribed ion is the quantitative ion.
Preparation of standard solutions Individual stock solutions were prepared with methanol at concentrations of 100 mg mL1 and stored at
