Journal of Chromatography A, 1216 (2009) 80168034

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

Review

Options for veterinary drug analysis using mass spectrometry

Bruno Le Bizec , Gaud Pinel, Jean-Philippe Antignac
Laboratoire dEtude des Rsidus et Contaminants dans les Aliments (LABERCA), Ecole Nationale Vtrinaire de Nantes (ENVN), USC INRA 2013, BP 50707, F-44307 Nantes Cedex 3, France

a r t i c l e

i n f o

a b s t r a c t
Several classes of chemical compounds, exhibiting many different chemical properties, are classied under the generic term of veterinary drugs, among which are the antimicrobial medicines such as antibiotics or dyes, and drugs exhibiting growth promoting properties like steroids, -agonist compounds, thyrostats or growth hormones. For food safety purposes, the resort to these substances in animal breeding has been submitted to strict regulation within the European Union for more than 15 years. Systems of control have therefore been set up within the same period of time to ensure compliance with the regulation. The current strategy relies on targeted analytical approaches focusing on the detection of residues of the administered compounds or their metabolites in different kinds of feed, food or biological matrices. If screening methods, which provide rapid discrimination between compliant and suspect samples, may be based on several techniques such as immunoassays or mass spectrometry, conrmatory methods mainly rely on the latter, which provides adequate specicity and sensitivity for unambiguous identication of the target analytes in biological matrices at trace level. The present article reviews the main mass spectrometric strategies, from the very rst, nonetheless still efcient, single MS and multidimensional and high-resolution MS through to advanced isotope ratio MS. Several applications in the eld of residue analysis illustrate each of these approaches and focus on the balance between issues related to the compounds of interest (chemistry, matrix, concentration, . . .) and the large offer of mass spectrometric-related technical possibilities, from the choice of the ionization conditions (EI, NCI, PCI, reagent gases, ESI+, ESI), to the mass analyzers (single quadrupole, triple quadrupole, ion traps, time-of-ight, magnetic sectors, isotope ratio mass spectrometer) and corresponding acquisition modes (full scan, LRSIM, HRSIM, SRM, precursor scan, . . .). All the displayed strategies, from the importance of sample preparation to MS analysis to potential derivatization steps and chromatographic separation parameters are discussed in that context. Besides the advantages of each strategy, main issues associated to such MS approaches are commented with an emphasis not only on such critical points as ion suppression and resolution, but also on the adequacy of the current regulation regarding the evolution of the technology. Finally, future trends which may lead to strong and positive impacts in the eld of residue analysis are presented, including latest developments and improvements in chromatography or software dedicated to signal acquisition and data analysis. 2009 Elsevier B.V. All rights reserved.

Article history: Available online 14 July 2009 Keywords: Mass spectrometry Residues Hormones Growth promoters Veterinary drugs

Abbreviations: APCI, atmospheric pressure chemical ionization; APPI, atmospheric pressure photo ionization; CC, decision limit; CC, detection capability; CID, collision-induced dissociation; CRL, community reference laboratory; ECD, electron capture detection; EI, electron impact; ERC, endogenous reference compound; ESI , positive/negative electrospray ionization; FTICR, Fourier transform ion cyclotron; FWHM, full-width half-maximum; GCMS, gas chromatography coupled to mass spectrometry; HRMS, high-resolution mass spectrometry; IAC, immuno afnity chromatography; IRMS, isotopic ratio mass spectrometry; IT, ion trap analyzer; ITD, ion trap detector; LCMS, liquid chromatography coupled to mass spectrometry; LOD, limit of detection; LOI, limit of identication; LIT, linear ion trap; MRM, multiple reaction monitoring; MRL, maximum residue limit; MRPL, minimum required performance level; MS, mass spectrometry; MS/MS, tandem mass spectrometry; MS2 , tandem mass spectrometry; MSn , multidimensional mass spectrometry; NCI, negative chemical ionization; PCI, positive chemical ionization; QqQ, triple quadripole; RPLC, reversed phase liquid chromatography; SIM, single ion monitoring; SPE, solid phase extraction; SRM, single reaction monitoring; TOF, time-of-ight; TMS, trimethylsilyl; UPLC, ultra performance liquid chromatography; UV, ultra violet. Corresponding author. Tel.: +33 2 40 68 78 80. E-mail address: laberca@vet-nantes.fr (B. Le Bizec). 0021-9673/$ see front matter 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2009.07.007

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8017

Contents 1. 2. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8017 Mass spectrometric approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8018 2.1. Single MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8018 2.1.1. Basic introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8018 2.1.2. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8018 2.1.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8023 2.2. Multidimensional MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8023 2.2.1. Basic introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8023 2.2.2. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8023 2.2.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8027 2.3. High-resolution MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8027 2.3.1. Basic introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8027 2.3.2. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8027 2.3.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8029 2.4. Isotope ratio MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8029 2.4.1. Basic introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8029 2.4.2. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8029 2.4.3. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8030 Pitfalls and future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8030 3.1. Issues related to the application of the 2002/657/EC decision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8030 3.2. Chromatographic separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8031 3.3. Ionization (matrix effect) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8031 3.4. Ion characterization (resolution, mass accuracy, speed scanning, multiresidue analysis, cross talk) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8032 3.5. Software (acquisition, data analysis) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8033 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8033 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8033

3.

4.

1. Introduction Drug residues regulation in animal-derived food is an integral component of food safety programmes worldwide. Analytical methods used to monitor veterinary drugs in feed and food are essential to help protect human and animal health, monitor consumer exposure to the drugs, reduce the impact of chemicals on the environment, support the enforcement of laws and regulations and facilitate international trade of animal food products. Most veterinary drugs are not of acute toxicological concern. Some substances, such as diethylstilbestrol, nitrofurans or chloramphenicol, have been banned in most developed countries due to their demonstrated carcinogenicity. Increased cases of allergic reactions to antibiotics, as well as the growing current concern for pathogenic microorganisms becoming antibiotic-resistant, are also important reasons for setting maximum residue limits in food. Endocrine disruption properties of residues in food have become another major issue justifying the regulation of certain veterinary drugs. The occurrence of unwanted residues in edible products can be the result of illegal use, in the cases of prohibited medicines, or of failure to respect the proper withdrawal times before butchering in the case of authorized medicines. The European Union (EU) has strictly regulated controls on the use of veterinary drugs, including growth promoters, particularly in foodanimal species, by publishing different Regulations and Directives. The use of veterinary drugs is regulated through EU Council Regulation (2377/90/EC) [1], which describes the procedure for establishing MRLs for veterinary medicinal products in foodstuffs of animal origin. Four annexes cover substances with MRL values (I), substances for which it is not considered necessary to establish MRL values (II), substances with provisional MRL values (III) and substances for which no MRL values could be established so that the corresponding compounds are prohibited (IV). The prohibition of the use of growth promoters is laid down in two Council Directives (96/22/EC, 96/23/EC) [2,3] that contain guidelines for controlling veterinary drug residues in animals and their products with all the necessary information to set up national monitoring plans. For any type of animal or food, there are two main groups

of substances that must be monitored: Group A comprises prohibited substances in conformity with Directive 96/22 and Annex IV of Regulation 2377/90; Groups B1 and B2 comprise all registered veterinary drugs in conformity with Annexes I and III of the 2377/90 regulation and Group B3 comprises contaminants of the environment, such as phytosanitary products, heavy metals, dyes or mycotoxins. In contrast to other areas of food control or to what is enforced in most non-EU countries, in the EU there is no obligation to use standardized methods in the residue control of food-producing animals. Instead, a criteria approach applies, which lays down performance characteristics, limits and criteria that have to be met by the methods used. A signicant advantage of this approach is the high degree of exibility. It allows the ready adaptation of analytical methods to technical developments and offers the possibility to react rapidly to newly emerging problems. Recent examples are the presence of chloramphenicol in shrimps or honey and medroxyprogesterone acetate in feed. Technical guidelines related to the analytical performance criteria (e.g., detection level, selectivity and specicity) and validation procedures of methods used for residue control in the framework of Directive 96/23/EC are described in a dedicated Commission Decision (2002/657/EC) [4]. Next to the general performance lines, additional and new requirements are described in this reference document, especially for conrmatory methods, by introducing the concept of identication points (IPs) and unambiguous identication criteria (maximal variability authorized for ion ratio intensities). During conrmatory analyses, a specic number of IPs has to be collected. For conrmation of the identity of Group A substances, a minimum of four IPs is required. For conrmation of the identity of Group B substances, a minimum of three IPs is required. Besides the traditional biological measurement approaches based on RIA, ELISA or biosensors, which provide rapid diagnostic of compliant versus suspect samples (i.e., screening stage), mass spectrometry (MS) is today frequently used as conrmatory technique. Indeed, MS combined either with gas or liquid chromatography (LC) is a clear and valuable tool of choice to identify and quantify residues in feed and food [58]. Nowadays, analytical strategies based on LCMS supplant those relying upon GCMS, even if they

8018

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

should be considered as complementary to one another. When considering compounds belonging to Group A1 (stilbens), A2 (thyrostats), A3 (steroids), A4 and A5, the preferred methods used to be based on GCMS. The detractors today are opposing the needs for a derivatization step (e.g., silylation or acylation) prior to GC separation, but they probably minimize, if not neglect, the incomparable chromatographic separation observed on capillary GC. But the fact that LCMS techniques provide a universal approach applicable to the widest number of veterinary drugs is a reality. Performances have been widely improved these past 10 years, particularly because of the introduction of new mass analyzers authorizing by principle/automatically authorizing specic acquisition techniques such as high-resolution selected ion monitoring (HRSIM, on double sector), high-resolution scanning (HRSCAN, on time-of-ight or orbital trap instruments) or selected reaction monitoring (SRM, on triple quadrupole or ion trap). The minimum required performance levels (MRPL) as introduced by the 2002/657/EC decision compelled to improve the limits of detection, both in providing more abundant signal and decreasing noise even in difcult matrices such as hair, urine or feeding stuff. To full these constraints of performance, single MS is gradually replaced by multidimensional MS which gives a much higher degree of certainty in analyte identication and quantication. Among the different mass analyzers usually applied for target analysis, triple quadrupole (QqQ) is the most widely used for measuring and quantifying residues of veterinary drugs. Conventional ion trap (IT) provides sensitive full scan mass spectra and has the ability to perform MSn experiments, which improves its ability to characterize structurally and to conrm the identity of metabolites. Quadrupole linear IT (QqLIT) is a system that combines fully functional QqQ and LITMS within the same instrument. In addition to standard QqQ capabilities (i.e., very sensitive SRM), QqLIT is capable of producing MSn spectral information, useful for structure elucidation. A recent tendency towards the popularization of high-resolution mass analyzers, and especially TOF instruments, is undoubtedly observed; the mass accuracy, the high speed to deliver mass spectrometric signals and the non-a priori approach is really tempting. The consequence is the growing of analytical development dedicated to multiple residues, sometimes exceeding 100, with major applications in the eld of pesticides [9] or veterinary drugs [5,6,8]. After years characterized by a continuously growing focus on specic acquisition modes (SIM, SRM, MRM, . . .), one tendency is more recently noticed for a renewed interest in full scan MS acquisition modes that enable retrospective analysis without reinjecting the sample. Through the new limits achieved in terms of sensitivity in this mode by the latest generations of instruments, it represents a powerful tool for the identication of non-targets and unknowns in food samples that have not previously been included in any of the routine target multiresidue methods. However, the use of these new types of high-resolution instruments providing low-ppm mass accuracy (e.g., OrbitrapTM and Fourier transform ion cyclotron resonance (FTICR)) still remains exceptional for target approaches [8]. In this general context, this review addresses one highly relevant topic: the residue analysis of veterinary drugs and growthpromoting agents in food-producing animals. The aim of this article is to provide helpful background to laboratories contemplating a review of veterinary drug analysis in feed, food of animal origin, as well all in any material which can be collected from the animal and help demonstrate the administration of a drug. It should be stressed that the paper is not intended to present a model for drafting a unique analytical strategy dedicated to veterinary drugs in general. Each reader will realize that for each topic, in various contexts, different technologies are available to the analyst, each providing its own characteristics, which can sometimes be advantages or occasionally limitations.

The paper begins with a comparison of single mass spectrometry versus multidimensional and high-resolution mass spectrometry. Illustrations will be given either on growth promoters, authorized or forbidden veterinary drugs. Different categories of mass analyzers, chromatography couplings, ionization and acquisition techniques will be critically depicted in this context. This review will then consider the particular context of natural hormones, with a description of Gas chromatographycarbonisotope mass spectrometry (GCCIRMS) as an emerging analytical tool in the eld. Finally, the main pitfalls and new trends in the control of veterinary drugs will be pointed out, including particular comments related to the current issues with to the application of the 2002/657/EC decision as regards new analytical technologies. It should be noted that the present review does not aim at presenting an exhaustive list of applications in the eld and do not therefore represent a comprehensive guide to applications; in each section, a selection of three to four examples has been realized for illustration purposes. 2. Mass spectrometric approaches 2.1. Single MS 2.1.1. Basic introduction Historically, the control of hormone residues and veterinary drugs was based on chromatography coupled to non-specic technologies such as uorimetric detector, ultra violet detector, electron capture detector or on immunochemical assays. The rst introduction of mass spectrometry in the 1980s was immediately considered as a revolution in the domain due to its outstanding specicity and sensitivity. The implementation of gas chromatography coupled to mass spectrometry occurred rstly in the eld of hormone residues analysis and led to a revolution regarding the ght against the misuse of forbidden compounds. Few applications related to liquid chromatography coupled to single mass spectrometry have also been reported in the literature; however, their performances were not in accordance with regulatory requirements (identication points, specicity). They will not be covered in this section. 2.1.2. Applications 2.1.2.1. Stilbens (GCMS, Q, EI, NCI, SIM). The bad reputation of hormones is mainly caused by the harmfulness of diethylstilbestrol (DES). The properties of DES as a growth promoter were discovered in 1954 in the USA. It has been widely used as a feed additive or as an implant into the ear to promote growth and increase feed efciency in cattle and sheep (until 1979 in the USA). DES is a non-steroidal oestrogenic agent which does not occur naturally. Its oestrogenic effect is mainly attributed to the same spatial distance of the two OH groups as for oestradiol. Hexestrol and dienestrol are two other examples of synthetic non-steroidal oestrogens that are structurally related to DES. Conversely, resveratrol is a naturally occurring stilben compound that may be found in various vegetal feed, grapes in particular. Based on their carcinogenity, the EC banned the use of stilbens in food manufacturing. Consequently, no residues of these anabolics should be present in animal products imported or produced within the EC (Directive 96/22/EC). Urine and hair samples are normally ad hoc biological matrices since they are readily available at both slaughterhouse and farm. Extraction and purication of stilbens in biological uids and tissues are generally based on at least one solid phase extraction (SPE) column. Reversed phase grafted silica C18 and C8 are the most popular stationary phases used for the rst stage of extractionpurication. But normal phase NH2 grafted silica or pure silica (SiOH) are also occasionally used to purify the biological extract. As stilbens can be ionized at alkaline pH (formation of phenolate ionic species), SAX (strong anion exchange) columns

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8019

are sometimes found in the literature for the isolation of this group of compounds. Stilben conjugated phase II metabolites are usually hydrolyzed either by enzymatic (glucuronidase) or chemical approaches (solvolysis or methanolysis), especially when the monitoring is conducted in urine, liver or kidney. This operation is indeed not required for muscle, hair or faeces. When GCMS is used, stilbens are quite always derivatized, the most popular approach being silylation, and acylation (acetylation, peruoroacylation). N-methyl-N-(trimethylsilyl)-triuoro-acetamide (MSTFA) is often used for this purpose. When LCMS is exploited, derivatization such as uoroacylation, even not mandatory, is sometimes used to enhance the signal and/or improve the specicity. In the 1990s, mass analyzers were mainly single quadrupoles and rst generation ion trap devices with an internal ionization source. Single MS is nowadays almost systematically replaced by multidimensional approaches, either by triple quadrupole (QqQ) and more recent generations of ion trap (external sources) either 3D or 2D (LIT). Few applications on high-resolution mass spectrometers either double sector or time-of-ight instruments are also reported in the literature, probably because stilbens are not characterized by a remarkable mass defect/excess so that the increase in resolution does not impact directly onto the S/N ratio. One of the major drawbacks of stilben analysis in GCMS is probably the difculty to obtain sufcient diagnostic signals especially for hexestrol. Indeed, the TMS-derivative of this compound completely fragments when ionized in EI. The major ion is observed at m/z 207, which is a very noisy diagnostic signal as far as the stationary phase is a methylpolysiloxane derivative. Two strategies are followed in the conrmatory process, the rst being the TBDMS derivatization, the second one consisting in the modication of the ionization process or the electron energy. Fig. 1 clearly shows the modication of the mass spectrum of hexestrol depending on the preferred strategy. A specic mass accuracy may be obtained for these target compounds through the introduction of a specic chemical modication so that the increase of the resolution either on TOF, FTMS or BE instruments has the direct consequence of an efcient mass clean-up. For example, the introduction of several halogen atoms on the target analytes during the derivatization step was proven to engender a sufcient mass defect, so that matrix interferences are not detected at a higher resolution [10,11].

2.1.2.2. Chloramphenicol (GCMS, ITD, NCI, SIM). Chloramphenicol (CAP) has been widely adopted as an effective broad spectrum antibiotic to treat many kinds of animal diseases. Because of some toxicity evidences that have been extensively demonstrated in humans, chloramphenicol has been prohibited for use in foodproducing animals and the maximum residue limit has been established at a zero tolerance level in edible tissues in many countries. For screening purposes, capillary electrophoresis, micellar electrokinetic chromatography and surface plasmon resonance biosensor assays have been used for the multiresidue analysis of fenicols in different matrices. Liquid chromatography coupled to UV detector as well as GCECD have also been used for the determination of CAP in bovine, swine, poultry muscle, sh, milk, and shrimp tissues. However, methods using chromatography coupled to mass spectrometry remain the current standards to conrm unambiguously the presence of the target analytes in suspect samples. If GCMS methods based on electron ionization (EI) have historically been used for this purpose [12], the resulting sensitivity sometimes remains insufcient. Negative chemical ionization (NCI) is more commonly used because particularly well adapted for these halogenated substances which exhibit intense electronic capture properties [1316]. Puried extracts are usually derivatized using silylating agents prior GC separation. In this case, the trimethylsilyl (TMS) derivative of CAP leads to 4 diagnostic ions (m/z 466/468 and m/z 376/378) which allow complying with the identication criteria (Council Decision 2002/657/EC) xed at the European level (Fig. 2). The detection limits achieved with this approach are typically in the sub 0.1 g kg1 (ppb) range in biological tissues. Either quadrupole or ion trap mass analyzers operating in single ion monitoring (SIM) acquisition mode can be used for this purpose, usually with methane as reagent gas. The same strategy can be successfully applied for measuring other related compounds such as thiamphenicol (TAP) or orfenicol (FF). 2.1.2.3. -agonistic drugs (GCMS, Q, PCI, full scan, reagent gas). -agonist compounds, the so-called -sympathomimetics, are characterized by structural and pharmacological properties which are very close to those of catecholamines. The control of -agonists misuse received extra attention after several food poisoning cases in 1990 in Spain caused by consumption of bovine liver, and later

Fig. 1. Different mass spectra obtained for hexestrol depending on the ionization technique used (top left corner EI 70 eV, top right corner PCICH4 , bottom left corner PCINH3 , bottom right corner NCINH3 ).

8020

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

Fig. 2. Typical GCNCIMS mass spectrum obtained for chloramphenicol TMS derivative.

on in China (20082009). This was the rst time that pharmacological residues found in slaughtered cattle were found to have caused acute intoxication in consumers. In muscle tissue, -agonistic drugs promote lipolysis. This may result in a reduction of up to 40% of carcass fat (lipolysis) and an increase of up to 40% of carcass protein (by increasing protein synthesis and reducing proteolysis in striated muscle bres). Because of their ability to shift nutrients towards protein instead of lipid anabolism, such molecules were gathered under the generic name of repartitioning agents. While the therapeutic treatment of cattle with respiratory diseases is permitted, the use of -agonists as growth promoters in cattle is forbidden in the EU. Regarding sample preparation, solid phase extraction (SPE) is among the rst choice for multiresidue -agonist extraction, preferably with mixed phase sorbents such as C8 and cation exchange stationary phase. Alternatively, immuno-afnity cleanup (IAC) can be used for efcient purication of the extracts. In this last case, different antibodies should be coupled to the column so that the system may capture the widest range of -agonist compounds taking into account their difference in terms of chemical structure, i.e., from clenbuterol to ractopamine, or from isoxsuprine to zilpaterol. Care should be taken for reutilization of IAC one sample to another to prevent any carry over. Molecular imprinted polymer (MIP) is a more recent alternative to IAC. MIPSPE is very selective, but the production of a constant quality material is sometimes reported as questionable as is the reproducible extraction of the analyte from the cartridge, especially when biological matrices are used. Now regarding the measurement techniques, radio and enzyme immunoassay screening tools were developed in the past and are still being used in many countries for the control of -agonists. More recent biological tests based on surface plasmon resonance (SPR), optical biosensor or competition binding assay have been also proposed for screening -agonists. However, the sensitivity of these tests was sometimes rather limited to comply with the requirement of low residue levels as found in urine and tissue samples. Consequently, several methods based on MS instrumentation have been set up, among which GCMS and LCMS (often reinforced by MS/MS or HRMS congurations) have been recognized today as the most powerful approaches for measuring -agonist compounds [1618]. Concerning GCMS techniques, trimethylsilylation or tert-butyldimethylsilylation are commonly used derivatization methods. But extra strategies based on different derivatives and/or ionization modes have also been set up. For instance, cyclization of the side chain either with boroxime or DMCS reagents clearly stabilizes the compound and provides intense high m/z ions even in EI at 70 eV. However, this approach is disappointingly only applicable to clenbuterol-like compounds. An alternative approach to the derivatization is the use of milder ionization conditions. Positive chemical ionization (PCI) is indeed an interesting option providing a wide panel of different mass spectra by varying the nature of the reagent

gas in the ionization source. Fig. 3 illustrates the different mass spectra for mabuterol when EI, PCI (CH4 ), PCI (iBu), PCI (NH3 ) are operated respectively. In the EI mass spectrum, the molecular ion is missing and the only valuable ion is the m/z 86 characteristic of the side chain moiety of the compound. According to the same principle, detection at sub-ppb levels of most -agonists is achievable on the basis of one single ion, but does not authorize the condent identication of these residues according to ofcial criteria. Clearly, chemical ionization conducted in the positive mode provides more characteristic information especially when the proton afnity inbetween -agonists and reagent gas decreases (from methane to ammonia). The energy transfer then drops and the fragmentation is drastically reduced. The PCIisobutane mass spectrum is probably the one providing the best combination in-between specicity (number of high m/z ion) and sensitivity (reduced fragmentation so that the percentage of a given ion is standing for a substantial percentage of the total ionic current). The PCINH3 mass spectrum of mabuterol is almost condensed to the quasimolecular ion (M+H)+ , and has to be considered as highly attractive on MS/MS instrument when isolated as precursor ion (all the TIC is concentrated into on high mass ion) and fragmented in the collision cell to generate product ions. If GCMS instruments were historically the more widely used for various classes of residues, LCMS appears today as the method of choice and the major actual investment for many laboratories, especially for the analysis of polar compounds. Undoubtedly, reversed phase LC and positive ESI is the method of choice for most -agonistic drugs nowadays. 2.1.2.4. Antibiotics (GCMS or LCMS, Q, EICI or ESI, full scan or SIM). Bioassay techniques which are widely used to screen for antibiotics in food and tissues do not generally allow a distinction between members of classes of antibiotics, therefore providing a semi-quantitative estimate of total residues level. Suspect samples consequently need to be analyzed by sufciently selective and sensitive conrmatory MS-based methods. Even if the latest MS instruments present in laboratories allow for MSn detection, single MS can still be considered as an efcient acquisition mode for antibiotics [5]. Few application methods are reported based on GCMS, mainly intended for chloramphenicol analysis after CI or EI ionization [19], while in general, the use of RPLCESIMS is very efcient for several classes of antibiotics. In this context several applications are reported for some antibiotic peptides (avoparcin, bacitracin, . . .) [20], tetracylines [21], chloramphenicol [22], sulphonamides [2326], -lactams for which sensitive ionization is generally achieved by ESI(), however, this ionization mode is not suitable for the amphoteric ones such as amoxicillin and ampicillin for which ESI(+) is preferred [27]. Macrolides are also efciently analyzed by ESI(+)MS [26,27] on a single quadrupole mass spectrometer. The mass spectra of some macrolides (spiramycin (SPI), tilmicosin (TILM), oleandomycin (OLE), erythromycin (ERY),

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8021

Fig. 3. Different mass spectra obtained for mabuterol depending on the different ionization modes used (from top to bottom, EI, PCICH4 , PCIiBu, PCINH3 ). Clearly the energy involved is decreasing from CH4 to NH3 , whereas the relative abundance of the quasimolecular ion increased.

8022

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

Fig. 4. Typical mass spectra obtained for macrolides compounds after ESI+ ionization (cone voltage 55 V) [28]. SPI: spiramycin, TILM: tilmicosin, OLE: oleandomycin phosphate, ERY: erythromycin, TYL: tylosin tartrate, KIT: kitasamycin, ROX: roxithromycin, JOS: josamycin.

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8023

tylosin (TYL), kitasamycin (KIT) and josamycin (JOS)) obtained in the full scan mode at the selected conditions (Fig. 4), showed that the protonated molecular ion [M+H]+ is the predominant ion, except for SPI and OLE. The base peak for SPI was [M+2H]2+ (m/z 422), whereas the predominant ion for OLE was [MC7 H13 O3 +2H]+ (m/z 544), which corresponds to the loss of sugar moieties from [M+H]+ [28]. Sensitive measurements are generally carried out in SIM acquisition mode focusing on the pseudo-molecular [M+H]+ ion for each macrolide. In this case, the most abundant ion is used for quantication and the second and third ones are used for conrmatory purpose, except in the case of josamycin for which only two ions are signicantly observed in the mass spectra (Fig. 4). The expected quantication limit of macrolides compounds with this approach is around 25 g kg1 of dry weight of animal muscle (Fig. 5), which is compliant with established MRLs [27]. 2.1.3. Conclusion Single mass spectrometry approaches in the low-resolution mode are currently almost neglected for trace residue analysis. The release of ofcial identication criteria by means of the 2002/657/EC decision combined with the continuous technology innovation in the eld ionization interface, ion transmission, mass analyzer and ion detection gently move forward LRMS. Ion trap detectors and above all triple quadrupoles come to substitute single quadrupoles in this exercise, providing incomparable spectrometric signals with specic and sensitive ion chromatograms with tremendous S/N on at least two diagnostic traces even at the low pg level loaded onto the system. 2.2. Multidimensional MS 2.2.1. Basic introduction Tandem (MS2 ) or multidimensional (MSn ) mass spectrometry techniques today present incomparable advantages in the eld

of residue analysis at trace levels in complex biological matrices. Indeed, the fragmentation of the target compounds for detecting only specic product ions rather than the entire molecule permits to considerably increase the signal to noise ratio of the target diagnostic signal by decreasing to a major extent the interferences due to other compounds present in the nal extract with the same or very close molecular weight as the analyte of interest. Various mass analyzers offering these capabilities are already used routinely in combination with gas or liquid chromatography, among which triple quadrupoles (QqQs) and ion traps (ITDs), are in common use. More recent technologies are linear ion trap (LITs), orbital trap (OrbitrapTM ) and new-generation of hybrid instruments, e.g., quadrupole time-of-ight (QqTOF), quadrupolelinear trap (QqLITs) or linearorbital traps (LTQOrbitrapTM ), which are gaining widespread acceptance in several application areas. All these recent instruments offer advantages such as high scanning speeds, accurate mass measurement (QqTOF, LTQOrbitrapTM ) and increased sensitivity (LITs and new-generation of QqQs). The application range of multidimensional MS is today extremely wide, both in terms of target compounds and in terms of possible different acquisition modes. This last capability authorises not only very sensitive and specic quantitative target measurements, but also powerful untargeted shing approaches based on the detection of typical product/precursor ions or neutral species belonging to a class of substances.

2.2.2. Applications 2.2.2.1. Corticosteroids (LCMS/MS, QqQ, ESI, SRM, precursor scan). Natural corticosteroids (i.e., cortisol, cortisone) are hormones secreted by the adrenal cortex. Their anti-inammatory properties have led to the chemical synthesis of more active articial corticosteroids used in many veterinary therapeutic drugs. At the same time, these compounds increase weight gain (water and fat retention), reduce feed conversion ratio, and have a synergetic

Fig. 5. Diagnostic SIM chromatograms obtained for different macrolides in a meat sample extract spiked at 100 g kg1 [29].

8024

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

Fig. 6. Specic extracted chromatograms of dexamethasone obtained for a liver sample extract spiked at 0.1 ng g1 acquired in LC(ESI)MS/MS [MRM] (A) or LC(ESI)MS [neutral loss 90] (B) modes. (Waters UPLC-XEVOTM TQ MS instrument).

effect when combined with other molecules like -agonists or anabolic steroids. Thus, corticosteroids have been illegally used as growth promoters in cattle, administered through livestock food or by injection. From a regulatory point of view, some of them are authorized for therapeutic treatments (with a withdrawal period between treatment and slaughtering and established maximal residue levels in milk and muscle), but their use as growth promoters has never been allowed. Many authors have proposed methods based on gas chromatographymass spectrometry for the detection of corticosteroids, including chemical oxidation or derivatization. Although providing good sensitivity, these methods modify the chemical structure of the molecule, reducing information and specicity. Today, liquid chromatography coupled to multidimensional mass spectrometry (LCMSn ) represents a powerful alternative for this class of compounds, combining speed, specicity and sensitivity, especially when associated to the negative ionization mode [3037]. Indeed, the main ionic species usually observed in negative ESI or APCI is the pseudo-molecular ion [MH] or a [M+acetate] or [M+formate] adduct if traces of acetic or formic acid are added in the mobile phase, respectively. This last ion usually appears very intense and its fragmentation leads to the [MCH2 OH] product ion which corresponds to a cleavage of the C17 side chain characteristic of this class of compounds (loss of formaldehyde). This fragmentation pathway appears extremely efcient for the measurement of a large number of corticosteroids at trace residue level (i.e., in the sub 0.1 g kg1 range) in biological matrices. Triple quadrupole or ion trap instruments operating in multiple or selected reaction monitoring (SRM/MRM) acquisition modes are usually used for this purpose. But this particular mass spectrometric behaviour of corticosteroid is also particularly adapted to use alternative acquisition modes offered by triple quadrupole devices, such as neutral loss scan [36,38]. In this case, the loss of 60 or 90 mass units, which correspond to the fragmentations [M+acetate] > [MH] or [M+acetate] > [MCH2 OH] , respectively, represents an efcient strategy to screen potential corticosteroids with unknown or non-targeted structures. Of course, this strategy appears really efcient only with the latest generation of instruments reaching a good sensitivity even in scan mode (Fig. 6). Similarly, precursor ion scan may also be used for direct measurement of corticosteroid phase II metabolites by focusing on fragment ions characteristic of glucuronides and/or sulphate forms [39]. 2.2.2.2. Nitrofurans (LCMS/MS, QqQ, ESI+, SRM). A number of methods are currently available for the analysis of nitrofurans

in a variety of matrices. Nitrofurans are characterized by their rapid metabolism within a few hours after administration, leading to protein-bound metabolites which may persist for a considerable period of time in edible animal tissues. Therefore, detection methods for nitrofurans should focus on metabolites of the parent drugs such as 3-amino-2-oxazolidinone (AOZ), 3-aminomorpholinomethyl-2-axozolidinone (AMOZ), semicarbazide (SEM) and 1-aminohydantoin (AHD), corresponding respectively to the metabolites of furazolidone, furaltadone, nitrofurazone and nitrofurantoin. Most methods described in the literature are based on the release of protein-bound nitrofuran metabolites under acidic conditions followed by derivatization with 2-nitrobenzylaldehyde (2-NBA) and determination by liquid chromatography coupled to mass or tandemmass spectrometry [4042]. RPLCESI(+)QqQ is in this context the preferred technique to conrm the identity of nitrofuran metabolites which are monitored using the SRM transitions corresponding to the fragmentations of the [M+H]+ ion of the derivatized metabolite (Fig. 7): NPSEM: 209.1 > 165.9, 209.1 > 191.9; NPAOZ: 236.0 > 133.6, 236.0 > 130.6; NPAMOZ: 335.1 > 291.1, 335.1 > 262.0; NPAHD: 249.0 > 133.6, 249.0 > 130.6 [5,4245]. Labelled internal standards such as d4 AOZ, d5 AMOZ, 13 C3 AHD or 13 C, 15 N2 SEM, are available and have been recently introduced in the methods to overcome problems such as matrix suppression during electrospray ionization [43,45]. 2.2.2.3. Malachite green (LCMS/MS, QqQ, ESI+, SRM). Conrmatory methods developed for malachite green (MG) and its main metabolite leuco-malachite green (LMG) in sh tissues originally involved in the mid 1990s GCMS detection methods. Selected ion monitoring was then performed based on 5 diagnostic ions (m/z 330, 329, 253, 210 and 165 in the case of LMG) [46]. More recently, the development of LCMS coupling combined with instruments allowing specic acquisition through tandem mass spectrometry has led to efcient detection methods which are now widely reported in the eld. Chromatographic separation of MG and LMG is generally performed on phenyl phases using either a gradient of acidic acetonitrile (0.1% FA)/water or an isocratic mixture of ACN/acetate buffer (70/30, v/v) as mobile phases [47,48]. C18 phases with 50 mM ammonium acetate/ACN or acidic water/ACN as eluents have also been reported [4951]. Atmospheric pressure chemical ionization has shown to be very efcient for MG ionization which is recovered as [M]+ with a molecular ion at m/z 329 [52]. Electrospray ionization also allows monitoring MG as [M]+ while LMG is recovered as [M+H]+ . The use of ion trap (3D or linear) as mass analyzer is

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8025

Fig. 7. LCESI(+)MS/MS diagnostic ion chromatograms obtained for 5-nitroimidazoles and nitrofuran metabolites in a sample spiked with at 0.5 g kg1 and internal standards at 5.0 g kg1 [43].

reported as being a sensitive and selective technique for the analysis of MG and LMG since it allows for full collection of product scan data, further increasing the analytical selectivity of the method [53]. High collision energy, typically around 50%, is needed to obtain signicant abundance of these ions [47,54]. SRM methods with triple quadrupole as mass analyzers typically report the following transitions 329.3 > 165.0, 329.3 > 208.0, 329.3 > 313.3 and 331 > 239 and 331 > 316, for MG and LMG, respectively: [4951]. Isotopic internal standards (d5 MG and 13 C6 LMG) are available and commonly introduced in the methods to overcome problems such as matrix

suppression during electrospray ionization [55]. An example of the analysis of MG and LMG in sh tissue using one of these methods is shown in Fig. 8. 2.2.2.4. Zeranol (GCMS/MS, QqQ, ESI, SRM). Zeranol (-zearalanol, -ZAL), a synthetic non-steroidal anabolic agent, had been widely used for increasing the rate of body weight gain and improving feed conversion in ruminants. Zeranol is structurally related to the naturally occurring mycotoxin zearalenone (ZEN) also known as the Fusarium spp. toxin (F2-toxin). ZEN, -ZAL and related compounds,

8026

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

Fig. 8. Typical SRM diagnostic ion chromatograms obtained for malchite green (MG), d5 MG, leuco-malachite green (LMG) and 13 C6 -LMG in a sh sample extract (corresponding to 2 g kg1 MG and 7 g kg1 LMG) monitored on a LCESI+QTrap instrument. [55].

including -zearalanol (-ZAL), -zearalenol (-ZEL), -zearalenol (-ZEL), and zearalanone (ZAN), are referred to as resorcylic acid lactones (RALs). As a consequence of their occurrence in the food chain, considerable attention has been paid to their potential risk for human health. Zeranol has been widely used as a growth pro-

moter in the United States since 1969 to improve fattening rates of cattle. Its application has been banned in the European Union (EU) since 1985 (Group A of Council Directive 96/23/EC). Methods for urine and tissue have been published [5459]. Several methods using gas chromatography with mass spectrometry (GCMS) or tandem mass spectrometry (GCMS/MS) for the analysis of resorcylic acid lactones in biological samples have been reported but required chemical derivatization [60]. Liquid chromatography combined with tandem mass spectrometry (LCMS/MS) has proven to be the ad hoc technique for the determination of the whole range of resorcylic acid lactones. Triple quadrupole instruments with electrospray (ESI) or atmospheric pressure chemical ionization (APCI) interfaces either in the negative or the positive mode are applied. Zeranol and zearalenone are known to give identical metabolites which explains why these metabolites, including zeranol itself, can also occur naturally in ovine urine and bovine bile after metabolism of zearalenone. To differentiate an illegal use of zeranol from consumption of food contaminated with Fusarium spp. Toxin, an exhaustive monitoring of main RALs is necessary. The in vivo metabolism of zearalenone and zeranol has been investigated in several animal species and in humans. It has been shown that the anabolic agent zeranol is predominantly metabolized into its diastereoisomer -zearalanol (taleranol) and to a minor extent into zearalanone. The mycotoxin zearalenone is preferentially transformed into - and -zearalenol [61]. The objective of the Natural Zeranol project FAIR5-CT-1997-3443 was to study and to establish criteria for discriminating between illegal treatment with zeranol containing preparations and the presence of Fusarium spp. toxins. A statistical model was developed after screening and conrmation

Fig. 9. Compared HPLC separation of zeranol metabolites and precursors (mycotoxin origin). On the left, a C18 (X) column 50 mm 2 mm, 3 m has been used; on the right, a Hypersil Gold C18 column 100 mm 2.1 mm, 1.9 m shows its ability to separate the complex mixture of resorcylic acid lactones. Positive ESI has been used to ionize the analytes, and SRM acquisition on a QqQ was found the most adapted to reach the sub-ng mL1 in urine, and to full identication criteria as xed by the 2002/657/EC decision.

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8027

of 8008 samples collected from different parts of Europe. The model developed is a tool based on comparing the sum of the zeranol and taleranol concentrations with the sum of zearalenone and its two major metabolites, namely - and -zearalenol. Validated quantitative methods are necessary for practical applications. An example of a LCMS/MS characterization is proposed in Fig. 9, pointing out that crucial importance has to be paid to the stationary phase for efcient separation between metabolites (close chemical structure) as well as with other secondary metabolites and interferences. Electrospray ionization can be successfully applied for this class of compounds both in the positive or negative mode. SRM acquisition mode provides the ad hoc specicity (at least four identication points) and guarantees the necessary sensitivity (CC < 0.1 ng mL1 in urine) to achieve reliable control of zeranol in cattle. 2.2.3. Conclusion Nowadays, the combination of a wide range of chromatographic separation systems (GC or LC, conventional or more recent fastLC) associated to numerous possible ionization techniques (either EI, CI or API), and mass analyzers (QqQ, ITDs, TOF or QqTOF, Orbitrap system) provides the necessary tool for efcient monitoring of vet drug residues. Among them, multidimensional MS is probably today the most widespread strategy implemented in the eld. The considerable gain in terms of both sensitivity and specicity as compared to single MS performances clearly improves the result interpretation. However, a clear tendency towards the democratization of high-resolutionhigh mass accuracy measurements strategies has been observed for a couple of years. 2.3. High-resolution MS 2.3.1. Basic introduction Reliable identication of analytes in complex mixtures necessarily requires robust mass spectrometers with high resolving power, mass accuracy, sensitivity and dynamic range. High-resolution MS and accurate mass high-resolution MS (HRMS) are increasingly becoming more popular in laboratories, either in the form of TOFMS, or magnetic sector, and more recently FTICR or OrbitrapTM mass spectrometer. If the target analyte contains chemical elements with an outstanding mass defect, then HRMS becomes efcient in the suppression of matrix interference (also known as mass clean-up). Whereas hydrogen adds only +0.008u mass defect per H atom, each Cl atom has mass defect of 0.032u, and Br has an even greater mass defect of 0.083u. These relatively large mass defects induced by Cl and Br atoms cause a significant shift from the centre of the mass distribution for a given molecular weight, which clearly improves the diagnostic signal selectivity acquired in HRMS. However, the greater capabilities of HRMS have been fully demonstrated for decades in the analysis of dioxins, PCBs, and OCs. Other common elements in organic molecules have the following mass defects (u): C = 0, N = 0.003,

O = 0.005, F = 0.002. In recent years, time-of-ightmass spectrometry (TOFMS) has represented one of the fastest growing areas of mass spectrometry. The introduction of the reectron TOFMS which is used to compensate for energy spread from the initial ion velocities has resulted in mass resolving power approaching m/ m 10,000 [ m = full-width half-maximum (FWHM)]. Mass accuracy can currently reach values better than 10 ppm in routine conditions with external calibration. It is known that the selected width of the extraction mass window is critical regarding the false negative score of the method. Co-elution of isobaric compounds can result in signicant deviations in exact mass measurements, with possible assignment of the detected mass at the average with the accurate mass of the co-eluting compound. Hybrid mass spectrometers such as linear quadrupole ion trap/Fourier transform ion cyclotron (LIT/FTICR) or linear quadrupole ion trap/Fourier transform orbitrap (LTQOrbitrapTM ) devices combine large trapping capacity, MSn capability with unsurpassed high mass accuracy, resolving power, sensitivity and dynamic range (better than the one observed on a TOFMS instrument). The high resolving power (>150,000) and excellent mass accuracy (specied as 25 ppm, but demonstrated to be as low as 0.2 ppm under favourable conditions) signicantly reduces false positive identication. 2.3.2. Applications 2.3.2.1. Boldenone (LCHRMS, Orbitrap, ESI). The applicability of the OrbitrapTM technology to the measurement of steroid-related compounds in general, and conjugated phase II metabolites of boldenone in particular, in complex biological matrices has already been demonstrated [62,63]. Boldenone (androsta-1,4-dien-17ol-3-one) is an anabolic steroid synthesized by dehydrogenation of testosterone. Its use as growth promoter for cattle fattening is banned within the EU. Until 1996, the identication of either 17- or 17-boldenone in urine was considered exclusively as the result of an exogenous administration of boldenone or analogues. Nevertheless, some observations reported by ofcial laboratories underlying the almost systematic presence of 17-boldenone in urine with values in the low g L1 range [64], have led to consider a possible endogenous production/excretion of boldenone in bovine/ovine. Then a non-unambiguous piece of evidence became necessary to allow the competent authorities to take appropriate action. A screening criterion based on concentration level of 17-boldenone levels (>2 g L1 ) in urine has been set up. For conrmatory needs, the presence of 17-boldenone conjugates (glucuronide or sulphate metabolite) in urine has been recognized as a denite indicator to differentiate treated from untreated animals. LCMS approaches became necessary to point out these hydrophilic residues, based either on MS/MS or HRMS signal acquisition after electrospray ionization operated in the negative mode. For high-resolution acquisition on the OrbitrapTM system (Fig. 10), FTMS resolution was set at 30,000 (FWHM at 400 m/z). Mass spectra were recorded from m/z 100 to 400. The method validation focused on blank calf urine samples was run to assess the specicity

Fig. 10. Ion chromatograms (ESI, LCHRMS, R 30,000 FWHM) for 17-boldenone sulphate in (a) a blank urine sample, (b) a spike urine at 1 g L1 and (c) a urine sample collected 4 h after administration of boldenone to a calf. [62].

8028

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

of the method, which was concluded satisfactory since no interferences were eluted at the retention time (extended both sides to 5 times the peak width at half height of 17-boldenone sulphate, 11.6 0.5 min). A linear regression-based approach provided estimated values of decision limit (CC) and detection capability (CC) equal to 0.2 and 0.4 g L1 , respectively. Response linearity was found adapted to the needs (R2 = 0.9925). Identication relies upon the monitoring of two ions (precursors and products) acquired in high-resolution mode (R = 30,000 FWHM), which provided four identication points according to EU Decision 2002/657/EC. Comparable performances have been observed on a last generation but more conventional triple quadrupole instrument (QqQ, SRM acquisition) in terms of sensitivity, specicity and potential identication points. The main advantage of QqQ was its ability to provide fast scanning information. When the OrbitrapTM is limited to 78 acquisition points (30,000 resolution), the triple quadrupole was able to generate 15 to 20 points without any signicant loss in sensitivity, making fast LC approach achievable.

2.3.2.2. Thyrostats. Specic detection procedures for the analysis of this group of drugs have been described in the literature involving a long time ago TLC protocols [6567]. Then GCMS and LCMS approaches appeared [6870], and more recently LCMS/MS [71,72] and GCMS/MS methods which allowed an increase in the performance of thyrostat detection in various biological matrices and extending the monitoring to a wide range of compounds. The characterization of thyrostats in biological matrices remains a current difcult challenge because of their low molecular weight, their high polarity and the existence of several tautomeric forms. The derivatization step prior to any MS analysis was considered until today as the most efcient way to extract and analyze the compounds; derivatization agents like benzylchloride [73]. Pentauorobenzylbromide [67,7275], NBDCl [77] or 3-BrBBr or 3-IBBr [78] have been described for this purpose. Such derivatization induces a stabilization of the thyrostats chemical structure, a reduction of their polarity improving their separation, and an increase of their molecular weight, all these elements allowing lower decision limit and detection capabilities [72]. The newly developed protocols are efcient for the detection and identication of thyrostat compounds in biological uids and edible tissues in the g kg1 or g L1 range which is in accordance with the requirements of

the European Union provisional minimum required performance limit (MRPL) suggested at 10 g L1 (CRL guidance paper, 2007). Since high urine concentrations of residues ( 100 g L1 ) would be generated upon drug administration aimed at increasing animal weight, the occasional occurrence in the range 110 g L1 of thiouracil in urines collected on food-producing animals raised the question of the origin of the molecule and/or the origin of the associated signals [76,78]. Therefore resorting to highresolution mass spectrometry enabled deep investigation of the target compound proving beyond doubt its identity in the urine sample. Conrmatory analysis were performed by gas chromatography coupled to either negative chemical ionization or electron ionization high-resolution mass spectrometry (GC(NCI)HRMS) or (GC(EI+)HRMS) following derivatization with pentauorobenzyl bromide (PFB). The rst approach (GCNCIHRMS) led to an intense fragment ion at m/z 199.0361 corresponding to the elemental composition C7 H11 ON2 SSi (Fig. 11). The second approach (GCEIHRMS, PFB/TMS derivative) led to a molecular ion M + appearing at m/z 380.0438 corresponding to the elemental composition C14 H13 OF5 N2 BrSSi. Both ions led to unambiguous identication of thiouracil and can therefore be considered as further diagnostic ions for the identication of thiouracil in urine samples (Fig. 11).

2.3.2.3. Antibiotics. HRMS is becoming more popular in laboratories, particularly in the form of time-of-ightmass spectrometry (TOFMS) technology which can be implemented as an alternative to MS2 instruments providing high specicity due to both mass accuracy and high mass resolution. HRMS can also be found in magnetic sector, Fourier transform (FT) MS and OrbitrapTM exhibiting similar mass accuracy compared to TOF but a higher resolving power. In the eld of antibiotic residues analysis, this particular property might for example be of interest to distinguish two anthelminthic drugs with different regulatory LODs and exhibiting the same nominal molecular weight [i.e., albendazole sulphone (MW = 297.328) and hydroxyl mebendazole (MW = 297.313)] which are co-eluted and therefore undistinguishable using the classical RPLCMS2 methods implemented in control laboratories [79]. More generally, and as an emerging trend in veterinary drug residue analysis, several applications based on HRMS measurements have recently been reported for the development of

Fig. 11. GCNCIMS full scan mass spectrum of the PFB/TMS thiouracil derivative (left) and GCNCIHRSIM (R = 10,000) diagnostic ion chromatogram obtained for a standard solution (TU 10 g L1 ) and a urine sample.

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8029

multiresidue methods, especially for antibiotics analysis. However, multiclassmultiresidue methods used to be scarce, not because they were not needed, but because of a number of analytical challenges which had to be overcome, among which the lack of volatility and chemical stability of most drugs, the wide polarity range encountered and the trace level at which they occur in the matrices. Very interesting examples of the use of full scan MS for multiresidue screening of antibiotics in various biological matrices (urine, kidney, muscle, liver, milk) have been described by [9,80]. The methods generally cover more than 100 different veterinary drugs belonging to the following classes: quinolones, sulphonamides, nitroimidazoles, penicillins, macrolides, benzimidazoles, tetracyclines and tranquilizers. After extraction, the samples are analyzed by UPLC coupled to TOFMS. A common cone voltage for all the compounds measured is generally used, even if additional sensitivity could be gained by using dedicated voltages for every single compound. Ion optic voltages are set to values permitting high ion transitions while preventing CID. Temperature and gas ow parameters of the interface ensure proper desolvatation of all compounds without causing degradation of thermo labile ones. The combination of UPLC and TOF provides sufcient sensitivity and selectivity; furthermore the possibility of retroactive analysis is extremely valuable. 2.3.3. Conclusion The TOFMS full scan accurate mass screening procedure may enable the analysis of hundreds veterinary drugs and metabolites [9]. The main advantage is the theoretically unlimited number of compounds to be screened simultaneously at low concentration levels. Therefore LCTOFMS is used by the authors as a quantitative screening method. The samples suspected of containing a drug at a concentration level above the MRL or MRPL is further analyzed by an LCQqQMS method for the nal conrmation of the identity. Most of these HRMS approaches (except double sectors) have the clear advantage of retrospective analysis: careful examination of old raw data sets for ions of novel residues without re-analysis of samples is feasible. The accurate mass capability of LC/HRMS allows the reconstruction of highly selective accurate mass chromatograms of target residues in complex matrices. Accurate mass determination and elemental composition data calculated thereof can be used for structure elucidation as well. This characteristic of HRMS instruments is expected nowadays to be more and more used in metabolomic studies, with the main objective to identify and to further use diagnostic biomarkers reecting the biological effect of a group of substances. 2.4. Isotope ratio MS 2.4.1. Basic introduction To meet the endogenous steroid challenge, stable carbon isotope ratio (13 C) analysis of urinary steroids may be performed to satisfy the rigorous demands of regulatory control. The measurement of 13 C/12 C ratios by isotope ratio mass spectrometry (IRMS) can be a powerful tool to elucidate the origin of steroids. They are derived from phytosterols by partial synthesis, where solely C3 plants (Yams: Dioscorea spp., Soy: Glycine max.) seem to represent the primary source. Becchi et al.[81] suggested to take advantage of this fact and to use the 13 C/12 C ratio of testosterone precursors and metabolites to detect misuse of this steroid in sport [81]. The ability to measure isotopic distribution at natural abundance with high accuracy and precision has increased the application of gas chromatographycombustionisotope ratio mass spectrometry (GCCIRMS) in recent years. It was not until 1990, however, that GCCIRMS instruments became commercially available. Modern GCCIRMS instruments are designed for compounds to elute from the capillary column of the GC into combustion interface consisting of an oxidation furnace (reaction of Cu with high purity O2 gas to

form CuO at 600650 C), water remover (either by cryo-trapping or Naon dryer) and open split (sampling of the gas stream). The oxidation furnace converts all the carbon in each compound to CO2 which are further ionized (EI > 120 eV) before separation of CO2 + ions according to m/z on a magnetic sector. Ions m/z 44, 45 and 46, representing 12 C16 O16 O, 13 C16 O16 O and 12 C16 O18 O, are collected in three separated Faraday cups. The ratio of m/z 45 to m/z 44 is determined for each target analyte. An algorithm corrects m/z 45 for isotopic abundances contributed from 12 C16 O17 O by assumption that a xed relationship exists between 18 O and 17 O. 13 C/12 C ratios are expressed as 13 C values against the international standard (std), Vienna Pee Dee Belemnite (VPDB), based on the equation: (13 C/12 C)sample (13 C/12 C)std (13 C/12 C)std

13 C[] =

1000

The original Pee Dee Belemnite (PDB) was assigned the value of zero on the corresponding 13 CPDB scale. Most biogenic carbon compounds were 13 C depleted in relation to it, and therefore, had negative 13 CPDB values. 2.4.2. Applications 2.4.2.1. Natural steroids (GCCIRMS, EI). The demonstration of misuse associated with the natural steroids testosterone, oestradiol or cortisol is still problematic as no ofcial concentration threshold or list of discriminative metabolites have ever been accepted and published by the community reference laboratories (CRL) or the European Commission (96/23/EC). Then, the analyst is faced with the challenge of developing methods to provide unequivocal discrimination between the natural presence of an endogenous hormone and its presence in the biological matrix arising from administration. One of the unequivocal approaches currently used by some laboratories, whether in the eld of antidoping or food safety is to detect anabolic steroid administration by demonstrating the presence of the proprietary steroid esters at injection sites or in hair Nowadays, the most promising approach for the control of natural steroid hormones is the 13 C/12 C ratio measurement of steroids by gas chromatographycombustionisotope ratio mass spectrometry (GCCIRMS). It has been shown that the administration of natural hormones (testosterone, oestradiol, cortisol) leads to an alteration of the 13 C/12 C ratio of their metabolites whereas the isotopic composition of precursor steroids (upstream in steroid metabolism pathway) remains unchanged. If the difference between the 13 CVPDB values of a steroid or its metabolites and the ERC exceeds a given limit, this is considered as an evidence for the presence of exogenous steroids. [] = 13 CERC 13 CTARGET ANALYTE In the antidoping world in sport, an ofcial threshold has been set up and published by the World Antidoping Agency97 . . . the results will be reported as consistent with the administration of a steroid when the 13 C/12 C value measured for the metabolite(s) differs signicantly, i.e., by 3 delta units or more from that of the urinary reference steroid chosen . . .. If this threshold is not exceeded, the difference between the ERC and metabolite is then attributed to a natural variability; the sample is then concluded compliant. No ofcial criterion regarding is available in the food safety community even if Bichon et al. [82] suggested a set at 3 for discriminating testosterone and oestradiol treated animals from untreated ones [82]. An example of urinary 13 C/12 C proles of three metabolites (i.e., etiocholanolone, 17-testosterone and 5-androstane-3,17-diol) determined by GCCIRMS are shown in Fig. 12 following the administration of testosterone enanthate and 4-androstenedione to a bovine. The average value of the ERC [23.7 0.6 (n = 43)] demonstrated

8030

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

Fig. 12. Isotopic deviation measurement (expressed in 13 CVPDB in ) of ERC (DHEA and 5-androstene-3,17-diol) and metabolites of testosterone and 4-androstenedione (etiocholanolone, 5-androstane-3,17-diol and 17testosterone) during 10 days after testosterone enanthate administration (250 mg, IM injection) and 12 days after 4-androstenedione injection (100 mg).

of such a strategy for growth hormones characterization come up against several critical points such as the thermolability of the considered macromolecule and the trace level at which it occurs in biological matrices. The only attempt in this way was reported by Karlsen et al. [88] where the 13 C/12 C isotope ratio of recombinant bovine somatotropine (rbST) and of the endogenous growth hormone have been analyzed via on-line determination of 13 C/12 C isotope ratios after HPLC separation and isotope ratio MS measurement via the LC IsoLink interface [88]. Basically, the samples are oxidized within the aqueous solvent eluting from the HPLC and the generated CO2 is separated from the liquid phase and fed into the isotope ratio MS. The measured delta 13 C/12 C values of endogenous and recombinant bovine somatotropine analyzed by LCIRMS were shown signicantly different on standards, 13 C/12 C () = 20.94 and 16.69, respectively. The analysis of high level spiked plasma samples (10 g 100 L1 ) showed lower difference in the measurements (13 C/12 C () = 24.66 and 22.94 for bST and rbST, respectively) and therefore the inuence of the matrix and the need for improved purication before this strategy can be considered as successful. 2.4.3. Conclusion The IRMS methodology is nowadays becoming more and more popular for conrmation purpose regarding the determination of natural occurring growth promoters in cattle. The method is ofcially applied in antidoping and food safety to control testosterone, oestradiol, and research studies are currently running for nandrolone, boldenone and cortisol. An ofcial threshold is now proposed for testosterone and oestradiol in cattle, but it has not been accepted ofcially by the European authorities. Improvements are expected in the sensitivity of the instrument to facilitate the 13 C/12 C measurement of other natural steroids and to allow the technique to be applied to other biological matrices such as tissue samples. Extension to other isotopes would be benecial to ensure the unambiguous character of the conclusion; 2 H/1 H is probably the next item. A possible further development would be the GC GC approach to improve the chromatographic separation, as it remains a critical limitation in the robust determination of isotope composition. Finally, for other class of compounds such as growth hormone (somatotropine) new coupling such as LCIRMS would be benecial to the domain. 3. Pitfalls and future trends 3.1. Issues related to the application of the 2002/657/EC decision The validation of analytical methods has been subjected to various debates for decades among analytical chemists, but not only. In each application eld, specic discussions arise. As far as food safety is concerned and chemical hazard involved, the application of a specic decision is mandatory in Europe (2002/657/EC decision), except when a more specic regulation has been published as it is for other residues such as PCDD/F, PCB-dl (see 1883/2006 regulation). The Decision 2002/657/EC denes the performance criteria for analytical residue methods both for validation of the method and identication of the target analytes. Other specic expectations in term of performances have been published by CODEX, IUPAC, ISO, FDA, AORC, or WADA. Often met in these institutions, the limit of detection, limit of identication and limit of quantication, have been replaced by the concept of decision limit (CC) and detection capability (CC). For Group A compounds, i.e., forbidden substances, the concept of the minimum required performance limit (MRPL) has been introduced; it corresponds to the minimum content of an analyte in a sample that has to be detected and conrmed. The Decision introduces as well the concept of IPs, which

the robustness of the analytical methodology and the good homogeneity of the endogenous steroid isotopic deviation (when the diet, i.e., hay in this case, is kept constant). Fig. 12 illustrates the depletion of the metabolite isotopic deviation following injection of 17-testosterone and shows a difference in-between ERC and metabolites (etiocholanone or 5-androstane-3,17-diol) for over one week after injection. Buisson et al.[83] studied the efciency of such an approach to demonstrate oestradiol administration to cattle after oestradiol valerate injection [83]. The 13 CVPDB values of the both ERC (i.e., DHEA and 5-androstene-3,17-diol) and the main oestradiol metabolite (17-oestradiol) were measured in urine samples collected in different animals, treated versus non-treated, gender (male, female versus castrated), age (sexually mature and immature) and feedings (grass or maize). The ERC 13 C/12 C ratio was not affected by the oestradiol treatment and found very repeatable one animal to another when feed was remained constant (maize 17.8). The () was found equal to 14 after treatment; as the relative concentration of endogenous 17-oestradiol was weak compared to exogenous residues, the () remained almost constant over the period of time. This difference is substantial and far above the 3 threshold (uncertainty of measurement + interanimal variability) and unambiguously allowed the differentiation in-between treated and non-treated animals. 2.4.2.2. GH. Somatotropin, also known as Growth Hormone, is a protein hormone exhibiting a molecular mass around 22 kDa and containing 191 amino acids in the bovine specie. Several pharmaceutical companies have developed large scale production of this protein using recombinant techniques. In several countries somatotropine is thus available on the drug market and used in food-producing animals as a general growth promoter or to increase milk production in dairy cattle. The use of somatotropin in the European Union is forbidden and therefore efcient control methods have to be set up in order to differentiate between endogenous and recombinant forms of the hormone. One way for analysis these substances is based on the mass difference between endogenous and recombinant somatotropine and successful strategies have recently been developed to solve this issue [8487]. However and since one of the available recombinant growth hormones is the same as the natural product, only isotope ratio analysis can be used for the determination of the origin of the growth hormones. Stable isotope ratio analysis can provide distinctive ngerprints in order to determine the origin of natural materials and to authenticate pharmaceuticals. If this strategy is now very efcient for natural steroid hormones identication through GCCIRMS, the implementation

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8031

corresponds to the minimal number of credits to be reached before any conclusion compliant/non-compliant for a sample. When 4 IPs are necessary to prove the identity of forbidden compounds, 3 IPs only are expected for Group B substances. These numbers of credits depends mainly on the specicity of the signals generated by the analytical method used; for all these reasons, chromatography coupled to mass spectrometry is unavoidable. The attribution of IP relies upon tolerable variability in the relative ion-abundance ratios; the tolerance is wider in positive chemical ionization mode and atmospheric pressure ionization than in electron ionization for instance. Necessary discussion and possible improvements may concern: - Mass accuracy. The recent introduction of TOF and FTMS in the eld makes urgent a discussion regarding the minimal exigencies linked to the denition of mass accuracy. A strong link with MS resolution has to be kept in mind, as it is known that an insufcient resolution leads to inappropriate mass accuracy and so direct and clear consequence onto false compliant results. - High-resolution. The HR criteria for mass spectrometer is applicable when resolution is greater than 10,000 for the entire mass range at 10% valley denition (approximately 20,000 FWHM). It has been determined at the origin for magnetic sectors where lock mass were used; in the 2002/657/EC decision, nothing refers to TOF, or FTMS instruments. - Very high resolution, i.e., R above equal or above 80,000100,000. It concerns mainly FTMS instruments. - Fingerprinting approaches. Several research projects are currently on-going, and nal results are expected soon. These metabolomic approaches are mainly conducted to ght against the illegal use of forbidden growth promoters. To prepare the introduction of these new screening strategies in the ofcial control, new specic criteria should be discussed. Special attention should be paid at least to the approved denition given to a biomarker, the number of target molecules to build a robust metabolomic model, etc. Some new validation criteria for this multiparametric approaches will be also to be invented and implemented, as well as some methods of estimation for false positive and negative rate. - Isotope ratio mass spectrometry. The needs are growing up in the eld of natural hormones control; the approach is now ofcially used in at least one National Reference Laboratories in Europe. Some guidelines have to established regarding the 13 C/12 C determination; it may concern at least the apparatus calibration, the signal integration, the linearity of the response, the calculation of the isotopic deviation difference between endogenous reference compounds and target metabolites, etc. - Minor additional points should be discussed, and they may concern numerous items, e.g., the consequence of the introduction of last generation of MS instruments for which the noise is not always measurable (zero amplitude), or the denition of a reference sample (spiked, standard . . .) to be used for nal identication of an analyte. Another important issue not covered in this reference document is the quantitative aspect associated by denition to the concept of CC and CC. Indeed, even with qualitative methods, the comparison of a measurement result with the decision limit value determined during the validation process imply a quantitative estimation of the concentration present in the considered sample. And no ofcial indication is given regarding this quantitative estimation: how many and what concentration levels for the calibration curve?, what fortied sample (representative sample, mixture of different sample)?, what quality control and procedures (preparation of standard solution, metrology)? . . . The quantitative verication of ion ratio is also in some extend incompletely discussed, considering that no precise indication is given regarding the sample nature and

concentration level that have to be used as reference value (ion ratio are largely dependent of either matrix effects and analyte concentration so that the reference sample have to be representative of the analyzed sample with a similar concentration). 3.2. Chromatographic separation Chromatographic techniques play a signicant role in the determination of an analyte in a complex biological matrix. For residue control in food, gas chromatography and liquid chromatography are the two main chromatographic techniques in use for routine analysis. A growing interest in fast GC and LC separations is currently observed. As the run time is continuously decreasing, the detector ability to scan faster and faster is expected. GC and LC methods can be speeded up by employing higher mobile phase ows, shorter or dedicated columns such as shorter megabore or microbore for GC purposes or smaller particle size (1.7 and 1.8 m) or monolithic columns for LC purposes. Fast separation plays an important role in two-dimensional (2D) separations, particularly in comprehensive 2D chromatography which represents a major improvement in comparison to GCMS techniques [89,90]. GC GC present the advantages to increase peak capacity, increased sensitivity and selectivity, independently from retention processes, and nally provide two independent retention times for each analyte. Basically, the separation of many unresolved components from the rst dimension column is achieved in the second dimension. Primary columns typically used in these systems are generally 1530 m 0.25 mm 0.251.0 m lm thickness. The rst column is often a non-polar stationary phase but not always. The second dimension separation must be very fast and performed with a stationary phase that is different from the one used in the rst dimension. Typical dimension characteristics for the secondary column is 0.51.5 m 0.1 mm 0.10.25 m. To be able to characterise, the multitude of narrow peaks generated by the system, the MS detector must be characterized by a high scanning rate; TOF instruments are appropriate to achieve this exercise [91,92]. The GC GCTOFMS system presents not only a superb separation power, but also reliable data for identication, which is obtained from continuous acquisition of full mass spectra. Very recently, Silva et al. [91] demonstrated the identication of anabolic agents (clenbuterol, norandrosterone, epimetendiol, two methyltestosterone metabolites and 3 -hydroxystanozolol) contained in a spiked urine sample at 2 ng mL1 . Special emphasis was given to 3 -hydroxystanozolol, mainly considering the difculty in its detection. In contrast to conventional GCMS approaches that must use single ion monitoring, the GC GCTOFMS method enabled the identication of that metabolite through the deconvolution of the full mass spectrum and also resolved the co-eluted peaks of 3 hydroxystanozolol and an endogenous component (Fig. 13). 3.3. Ionization (matrix effect) During a rst development period (1980searly 1990s), a common enthusiasm was shared by the new users of LCMS-related techniques. Limited or no sample preparation was often presented as a possibility even for complex matrices with a guarantee of repeatable results even with high-throughput ambitions. At the early stage of LCAPIMS, the chromatographic system was often considered as a loading system only. However, during a second period (1990stoday), various studies started to report some troubleshooting associated to these techniques. Overall, the main source of analytical problems encountered by LCMS users was related to matrix effects problems. For many years, the composition of a sample extract and the presence of interfering compounds have been recognized to have major inuence on the analyte signal, whatever the detection technique used. But in the specic case of

8032

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

Fig. 13. GC GCTOFMS TIC chromatogram of a human urine sample spiked with anabolic agents and with three-dimensional (3D) expanded regions of the anabolic agents in the spiked and in the blank urine for the key compounds (clenbuterol, stanozolol, 17-methyl-5-androst-1-en-3-17-diol, epimetendiol (EMD), 17-methyl 5-androstane-3,17-diol (methyltestosterone M1), 17methyl 5-androstane-3,17-diol (methyltestosterone M2)) [from Silva et al. [91]].

mass spectrometry, the so-called ion suppression phenomenon appears as one particular tricking manifestation of matrix effect. It represents certainly one of the main source of pitfall for the analyst, affecting many aspects of the method performances such as detection capability, repeatability, or accuracy. The possible origins of ion suppression are multiple [93,94]. The main one universally reported is the presence of endogenous substances, i.e., organic or inorganic molecules present in a sample and still present in the nal extract. Among this rst group of ion suppressor agents, ionic species can be included (inorganic electrolytes, salts), but highly polar compounds (phenols, pigments), and various organic molecules including carbohydrates, amines, urea, lipids, peptides, analogous compounds or metabolites with a similar chemical structure should be considered as well. Finally, a wide range of molecules can lead to ion suppression, especially when they are present in high concentration in the extract and eluted in the same elution range than the analyte of interest. A second source of difculty is the presence of exogenous substances, i.e., molecules not present in the sample but coming from various external sources during the sample preparation. Among this second group, plastic and polymer residues [95] phthalates, detergent degradation products (alkylphenols), ion pairing reagents [9698], proton exchanges promoting agents such as organic acids [98100], calibration products, buffers, or material released by the solid phase extraction (SPE), LC or GC stationary phases can be listed. Different mechanisms have been proposed to explain ion suppression [101,102]. In the case of LCMS, the main phenomenon corresponds to the decrease of the evaporation efciency due to the presence of matrix components. Indeed, the presence of interfering compounds in high concentration can increase the viscosity and the surface tension of the droplets produced in the ESI or atmospheric pressure chemical ionization (APCI) interfaces, and may reduce the capability of the analytes to reach the gas phase. The co-precipitation of the analytes with non-volatile material such as macromolecules can also limit their transfer in the gas phase. Another proposed mechanism is the competition between analytes and interfering compounds regarding the maximal ionization efciency of the technique [103]. A last possible mechanism involves neutralization processes linked to the relative basicity in the gas phase of the analytes and interfering substances, as well as to the stability of the produced ions in the

gas phase. The consequences of ion suppression are numerous, all affecting the different aspects of the analytical result. The detection capability is clearly reduced due to the decrease of the analyte signal intensity. The repeatability is also affected, because the degree of suppression may vary in a large extent from one sample to another. Ion ratio, linearity, and quantication, are also affected due to the variability of this unpredictable and not always repeatable phenomenon. Another side-effect of ion suppression is the difculty to perform database searching, because of the modication of the typical mass spectra patterns. Finally, ion suppression may lead to the non-detection of a given analyte, to the weak estimation of its real concentration, or to the non-fullment of the identication criteria, with immediate consequences on the false compliant score. If affecting the internal standard rather than the analyte, ion suppression may also lead sometimes to an overestimation of the analyte concentration with a clear risk on false non-compliant results; it concerns mainly maximum residue limit compounds. A rst possible action to overcome ion suppression troubles would consist into the modication of the mass spectrometric conditions, when possible. Indeed, the occurrence of ion suppression may differs between different ionization techniques (ESI, APCI, APPI), ionization modes (positive or negative), or between equipments with different source design [95,97,104106]. A second possibility is to improve the chromatographic separation efciency in order to shift the retention time of the analytes of interest far away from the area affected by ion suppression [105,107,108]. A third level to overcome this problem is to use adequate internal standard [109111]. 2 H- or better 13 C-labelled corresponding standards permit to reduce to a great extent the signal variability observed for the analyte and consequently to improve the repeatability of the measurement. The previously described action levels should permit to limit the consequences of ion suppression, but not to eliminate the risk as the cause is not deeply treated. Obviously, the best strategy is to take care of the sample preparation and purication to limit the presence of interfering compounds in the nal extract. Numerous authors demonstrated the evidence of such approach [108,112116]. Therefore, it should be strongly suggested to check the matrix effects resulting from different sample treatment procedures systematically during method development. In other word, the usual tendency to consider the recovery of the target analyte as a main performance indicator should be moderated by the necessity to evaluate also the method efciency in terms of removing interfering compounds. 3.4. Ion characterization (resolution, mass accuracy, speed scanning, multiresidue analysis, cross talk) Many signicant improvements have appeared since the last couple of years in the eld of mass spectrometric data acquisition, with direct positive impact on the instrumentation capabilities and performances. Some of these latest improvements may have some clear advantages in the specic eld of residue and contaminant analysis. Besides the introduction of new types of mass analyzers such as linear ion trap or orbital trap [117], a rst noticeable tendency is the increase of both resolution and mass accuracy. Current resolution better than to 30,000 FWHM may be today achieved either on the very last generation of time-of-ight or orbital trap instruments, which offers new capabilities for separating very complex mixtures containing isobaric compounds. Continuous progresses on electronic and computing devices also permits to increase the scanning speed of the corresponding MS instruments, which direct consequence on multi residue monitoring, and coupling with fast, high-resolution or two-dimensional chromatography. These improvements are benecial both for target classical approaches (focused approaches on already known target compounds), but to more global mass spectrometric approaches

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034

8033

involving ngerprinting, e.g., through metabolomic approaches [118121]. From this point of view, the usefulness of combined high sensitivity in full scan mode and high signal specicity through high-resolution and mass accuracy are obvious. And this tendency explains in some extent the novel wide interest for TOF instruments in particular. A direct consequence of these new screening developments is the demand for new chemical structure elucidation tools. Thus, the simultaneous possibility of multidimensional mass spectrometry and high-resolution ion characterization for both precursor and product ions undoubtedly appears as benecial in the scope of helping the determination of the chemical structure of unknowns. Thus, hybrid instruments such as quadrupoletimeof-ight (QTOF) or linear ion traporbital trap (LTQOrbitrapTM ) may be already considered as the current standard in the eld. 3.5. Software (acquisition, data analysis) Another eld of intense recent development is related to the software dedicated to the signal acquisition or processing. The simultaneous recording of different signals on a mass spectrometer is nowadays possible; for instance, the concomitant acquisition in the full scan mode with additional scanning events such as the precursor ion scan, the product ion scan, the neutral loss, or MSn is today achievable. This strategy may be of valuable interest for new shing approaches. These dependent scan capabilities are now implemented on various instruments and constitutes a key innovation in mass spectrometry for the last very few years. A second trend in terms of MS data analysis is the introduction of software packages specically dedicated to the processing of complex data les. Metabolomic experiments generate global and complex ngerprints which are characterized by their huge volumes of data impossible to handle manually or with conventional methods. Thus, such integrated software are designed to support the user for extracting the useful information from these large data sets, for example by automatically checking several ag parameters associated to ions to be fullled before being underlined, by comparing the abundances observed for these ions between two or more sub-groups of samples to be compared, or by proposing included statistical analysis modules for descriptive purposes [122127]. Some of these last generation software also propose a direct link with databases accessible via the Web for further structural elucidation of unknown potential biomarkers. 4. Conclusion After a very exiting period of intense analytical method development in the laboratories during the 1990s where classical mass spectrometry (quadrupole and ion trap detector) became the analytical standard for monitoring residues in food, the second period (up to 2009) saw the progressive introduction of new major technological improvements. MS/MS or MSn replaced MS1 in the late 1990s for target strategies, and progressively high-resolution MS was introduced in the eld, rst for its capacity to conrm difcult cases of residue traces in food (magnetic sectors), but more recently to authorize an open strategy for the simultaneous analysis of various classes of veterinary drugs including sometimes more than 100 compounds by using full scan MS techniques (TOF instruments). The current and future use of HRMS is undoubtedly the one consisting in the nding of biomarkers of a given class of compounds (so, characteristic of the biological effect) for further use as screening criteria to suspect the administration of a compound belonging to the corresponding pharmacological category; this strategy would be an answer to the new supposed strategies of misuse and would permit to detect the use of unknowns and low dose cocktails.

References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] Council Regulation 2377/90/EC, OJ L 224, 18.8.1990, p. 1. Council Directive 96/22/EC. Council Directive 96/23/EC, OJ L 125, 29.4.1996, p. 1024. Council Decision 2002/657/EC, OJ L 221, 17.8.2002, p. 836. A.A.M. Stolker, U.A.Th. Brinkman, J. Chromatogr. A 1067 (2005) 15. A.A.M. Stolker, T. Zuidema, M.W.F. Nielen, TrAC Trends Anal. Chem. 26 (2007) 967. A.A.M. Stolker, P. Rutgers, E. Oosterink, J.J.P. Lasaroms, R.J.B. Peters, J.A. Van Rhijn, M.W.F. Nielen, Anal. Bioanal. Chem. 391 (2008) 2309. M.W.F. Nielen, M.C. van Engelen, R. Zuiderent, R. Ramaker, Anal. Chim. Acta 586 (2007) 122. I. Ferrer, E.M. Thurman, J. Chromatogr. A 1175 (2007) 24. R. Bagnati, M. Paleologo Oriundi, V. Russo, M. Danese, F. Berti, R. Fanelli, J. Chromatogr. 564 (1991) 493. F. Courant, J.-P. Antignac, D. Maume, F. Monteau, A.M. Andersson, N. Skakkebaek, F. Andr, B. Le Bizec, Anal. Chim. Acta 586 (2007) 105. L. Santos, J. Barbosa, M.C. Castilho, F. Ramos, C.A.F. Ribeiro, M.I.N.d. Silveira, Anal. Chim. Acta 529 (2005) 249. J. Shen, X. Xia, H. Jiang, C. Li, J. Li, X. Li, S. Ding, J. Chromatogr. B 877 (2009) 1523. P. Li, Y. Qiu, H. Cai, Y. Kong, Y. Tang, D. Wang, M. Xie, Chin. J. Chromatogr. 24 (2006) 14. H.-Y. Shen, H.-L. Jiang, Anal. Chim. Acta 535 (2005) 33. M.-P. Montrade, B. Le Bizec, F. Monteau, B. Siliart, F. Andr, Anal. Chim. Acta 275 (1993) 253. M.-P. Montrade, B. Le Bizec, F. Monteau, F. Andr, Food Addit. Contam. 12 (1995) 625. F. Andr, B. Le Bizec, M.-P. Montrade, D. Maume, F. Monteau, P. Marchand, Analyst 119 (1994) 2529. A. Gantwerg, I. Shisbani, M. Hoffman, Anal. Chim. Acta 483 (2003) 125. C.A.J. Hajee, J.A. Van Rijhn, J.J.P. Laasroms, H.J. Keukens, J. de Jong, Analyst 126 (2001) 1332. F. Bruno, R. Curini, A. Di Corcia, M. Nazzari, M. Pallagrosi, Rapid Commun. Mass Spectrom. 16 (2002) 1365. M. Ramos, P. Munoz, A. Aranda, I. Rodriguez, R. Diaz, J. Blanca, J. Chromatogr. B 791 (2003) 31. S. Bogialli, A. Curini, A. Di Corcia, M. Nazzari, M. Sergi, Rapid Commun. Mass Spectrom. 17 (2003) 1146. S. Bogialli, A. Curini, A. Di Corcia, M. Nazzari, R. Samperi, Anal. Chem. 75 (2003) 1798. S. Bogialli, A. Curini, A. Di Corcia, M. Nazzari, M.L. Polci, J. Agric. Food Chem. 51 (2003) 4225. M.Y. Haller, S.R. Mller, C.S. McArdell, A.C. Alder, M.J.-F. Suter, J. Chromatogr. A 952 (2002) 111. A. Di Corcia, A. Nazzari, J. Chromatogr. A 974 (2002) 53. R. Codony, R. Compano, M. Granados, J.A. Garcia-Regueiro, M.D. Prat, J. Chromatogr. A 959 (2002) 111. H. Berrada, F. Borrull, G. Font, R.M. Marc, J. Chromatrogr. A 1208 (2008) 83. K. Croes, L. Goeyens, W. Baeyens, J. Van Loco, S. Impens, J. Chromatogr. B 877 (2009) 635. R. DiFrancesco, V. Frerichs, J. Donnelly, C. Hagler, J. Hochreiter, K.M. Tornatore, J. Chromatogr. B 859 (2007) 42. J.H. Andersen, L.G. Hansen, M. Pedersen, Anal. Chim. Acta 617 (2008) 216. G.N.W. Leung, E.W. Chung, E.N.M. Ho, W.H. Kwok, D.K.K. Leung, F.P.W. Tang, T.S.M. Wan, N.H. Yu, J. Chromatogr. B 825 (2005) 47. G. Brambilla, F. Buiarelli, G.P. Cartoni, F. Coccioli, C. Colamonici, A. Fagiolo, C. Giannini, B. Neri, J. Chromatogr. B 755 (2001) 265. R. Draisci, C. Marchiafava, L. Palleschi, P. Cammarata, S. Cavalli, J. Chromatogr. B 753 (2001) 217. J.-P. Antignac, B. Le Bizec, F. Monteau, F. Poulain, F. Andr, Rapid Commun. Mass Spectrom. 14 (2000) 3339. F. Bvalot, Y. Gaillard, M.A. Lhermitte, G. Ppin, J. Chromatogr. B 740 (2000) 227. E. Sangiorgi, M. Curatolo, W. Assini, E. Bozzoni, Anal. Chim. Acta 483 (2003) 259. J.-P. Antignac, A. Brosseaud, I. Gaudin-Hirret, F. Andr, B. Le Bizec, Steroids 70 (2005) 205. R.J. McCracken, G. Kennedy, J. Chromatogr. A 771 (1997) 349. R.J. McCracken, G.J. Kennedy, J. Chromatogr. B 691 (1997) 87. A. Conneely, A. Nugent, M. OKeeffe, P.P.J. Mulder, J.A. Van Rhijn, L. Kovacsics, A. Fodor, R.J. McCracken, D.G. Kennedy, Anal. Chim. Acta 483 (2003) 91. X. Xia, X. Li, S. Zhang, S. Ding, H. Jiang, J. Li, J. Shen, J. Chromatogr. A 1208 (2008) 101. G. Balizs, A. Hewitt, Anal. Chim. Acta 492 (2003) 105. X. Xia, X.W. Li, S.Y. Ding, S.X. Zhang, H.Y. Jiang, J.C. Li, J.Z. Shen, J. Chromatogr. A 1216 (2009) 2587. S.B. Turnipseed, J.E. Roybal, J.A. Hurlbut, A.R. Long, J. AOAC Int. 78 (1995) 971. W.C. Andersen, S.B. Turnipseed, J.E. Roybal, J. Agric. Food Chem. 54 (2006) 4517. K. Mitrowska, A. Posyniak, Jan Smudzki, J. Chomatogr. A 1207 (2008) 94. A.A. Bergwerf, P. Scherpenisse, J. Chomatogr. B 788 (2003) 351. A.A. Scherpenisse, Bergwerf, Anal. Chim. Acta 529 (2005) 173.

8034

B. Le Bizec et al. / J. Chromatogr. A 1216 (2009) 80168034 [88] C. Karlsen, G. Balizs, D. Juchelka, A. Lampen, Proceeding of the 17th International Mass Spectrometry Conference, Praha, 2006. [89] P. Mariott, R. Shellie, TrAC Trends Anal. Chem. 21 (2002) 573. [90] R. Ong, P.J. Marriott, J. Chromatogr. Sci. 40 (2002) 276. [91] A.I. Silva, H.M.G. Pereira, A. Casilli, F.C. Conceicado, F. Aquino Neto, J Chromatogr. A 1216 (2009) 2913. [92] J. Dallge, R. Vreuls, J. Beens, U.A.Th. Brinkman, J. Sep. Sci. 25 (2002) 201. [93] J.-P. Antignac, K. de Wasch, F. Monteau, H. De Brabander, F. Andr, B. Le Bizec, Anal. Chim. Acta 529 (2005) 129. [94] T.M. Annesley, Clin. Chem. 49 (2003) 1041. [95] H. Mei, Y. Hsieh, C. Nardo, X. Xu, S. Wang, K. Ng, W.A. Korfmacher, Rapid Commun. Mass Spectrom. 17 (2003) 97. [96] S.A. Gustavsson, J. Samskog, K.E. Markides, B. Langstrm, J. Chromatogr. A 937 (2001) 41. [97] M. Holcapek, K. Volna, P. Jandera, L. Kolarova, K. Lemr, M. Exner, A. Cirkva, J. Mass Spectrom 39 (2004) 43. [98] C.R. Mallet, Z. Lu, J.R. Mazzeo, Rapid Commun. Mass Spectrom 18 (2004) 49. [99] J. Eshraghi, S.K. Chowdhury, Anal. Chem. 65 (1993) 3528. [100] A. Apffel, S. Fischer, G. Goldberg, P.C. Gooley, F.E. Kuhlmann, J. Chromatogr. A 712 (1995) 177. [101] R. King, R. Bonglio, C. Fernandez-Metzler, C. Miller-Stein, T. Olah, J. Am. Soc. Mass Spectrom. 11 (2000) 942. [102] P.T.A. Reilly, A.C. Lazar, R.A. Gieray, W.B. Whitten, J.M. Ramsey, Aerosol Sci. Technol. 33 (2000) 135. [103] P. Kebarle, L. Tang, Anal. Chem. 65 (1993) 972A. [104] E.T. Gangl, M. Annan, N. Spooner, P. Vouros, Anal. Chem. 73 (2001) 5635. [105] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez-Eng, Anal. Chem. 70 (1998) 882. [106] Y. Hsieh, M. Chintala, H. Mei, J. Agans, J.M. Brisson, K. Ng, W.A. Korfmacher, Rapid Commun. Mass Spectrom. 15 (2001) 2481. [107] P.R. Tiller, L.A. Romanyshyn, Rapid Commun. Mass Spectrom. 16 (2002) 92. [108] C. Mller, P. Schfer, M. Strtzel, S. Vogt, W. Weinmann, J. Chromatogr. B 773 (2002) 47. [109] R. Kitamura, K. Matsuoka, E. Matsushima, Y. Kawaguchi, J. Chromatogr. B 754 (2001) 113. [110] M.J. Avery, Rapid Commun. Mass Spectrom. 17 (2003) 197. [111] L. Gomidez Freitas, C.W. Gtz, M. Ruff, H.P. Singer, S.R. Mller, J. Chromatogr. A 1028 (2004) 277. [112] M.W.J. van Hout, H.A.G. Niederlnder, R.A. de Zeeuw, G.J. de Jong, Rapid Commun. Mass Spectrom. 17 (2003) 245. [113] R. Bonglio, R.C. King, T.V. Olah, K. Merkle, Rapid Commun. Mass Spectrom. 13 (1999) 1175. [114] D.L. Buhrman, P.I. Price, P.J. Rudewicz, J. Am. Soc. Mass Spectrom. 7 (1996) 1099. [115] G. Shi, T.L. Lloyd, S.K.B. Sy, Q. Jiao, A. Wernicki, A. Mutlib, T.A. Emm, S.E. Unger, H.J. Pieniaszek Jr., J. Pharmaceut. Biomed. 31 (2003) 937. [116] C.H.P. Bruins, C.M. Jeronimus-Stratingh, K. Ensing, W.D. van Dongen, G.J. de Jong, J. Chromatogr. A 863 (1999) 115. [117] R.J. Qizhi Hu, R.J. Noll, A. Hongyan Li, A. Makarov, M. Hardman, R.G. Cooks, J. Mass Spectrom. 40 (2005) 430. [118] J. Nicholson, J. Lindon, E. Holmes, Xenobiotica 29 (1999) 1181. [119] D. Wishart, Trends Food Sci. Technol. 19 (2008) 482. [120] G. Theodoridis, H. Gika, I. Wilson, Trends Anal. Chem. 27 (2008) 251. [121] W. Lu, B. Bennett, J. Rabinowitz, J. Chromatogr. B 871 (2008) 236. [122] B. Warrack, S. Hnatyshyn, K.-H. Ott, M. Reily, M. Sanders, H. Zhang, D. Drexler, J. Chromatogr. B 877 (2009) 547. [123] D. Yuan, Y. Liang, L. Yi, X. Qingsong, O. Kvalheim, Chemometr. Int. Lab. Syst. 93 (2008) 70. [124] S. Mahadevan, S. Shah, T. Marrie, C. Slupsky, Anal. Chem. 80 (2008) 7562. [125] J. Trygg, E. Holmes, T. Lundstedt, J. Proteome Res. 6 (2007) 469. [126] H. Lu, Y. Liang, W. Dunn, H. Shen, D. Kell, Trends Anal. Chem. 79 (2007) 2745. [127] C. Smith, E. Want, G. OMaille, R. Abagyan, G. Siuzdak, Anal. Chem. 78 (2006) 779.

[51] L.A. Perez-Estrada, A. Aguera, M.D. Hernando, S. Malato, A.R. Fernandez-Alba, Chemosphere 70 (2008) 2068. [52] L. Valle, C. Diaz, A.L. Zanocco, P. Richter, J. Chromatogr. 1067 (2005) 101. [53] D.R. Doerge, M.I. Churchwell, T.A. Gehring, P. Yu Ming, S.M. Plakas, Rapid Commun. Mass Spectrom. 12 (1998) 1625. [54] K.C. Lee, J.L. Wu, C. Zongwei, J. Chomatogr. B 843 (2006) 247. [55] Z. Hall, C. Hopley, G. OConnor, J. Chromatogr. B 874 (2008) 95. [56] E.O. van Bennekom, L. Brouwer, E.H.M. Laurant, H. Hooijering, M.W.F. Nielen, Anal. Chim. Acta 473 (2002) 151. [57] F.M. Launay, P.B. Young, S.S. Sterk, M.H. Blokland, D.G. Kennedy, Food Addit. Contam. 21 (2004) 52. [58] F.M. Launay, L. Ribeiro, P. Alves, V. Vozikis, S. Tsitsamis, G. Alfredsson, S.S. Sterk, M. Blokland, A. Iitia, T. Lovgren, M. Tuomola, A. Gordon, D.G. Kennedy, Food Addit. Contam. 21 (2004) 833. [59] J. Jodlbauer, J. Zllner, W. Lindner, Chromatographia 51 (2000) 681. [60] M.H. Blokland, S.S. Sterk, R.W. Stephany, F.M. Launay, D.G. Kennedy, L.A. Van Ginkel, Anal. Bianal. Chem. 384 (2006) 1221. [61] M. Kleinova, P. Zllner, H. Kahlbacher, W. Hochsteiner, W. Lindner, J. Agric. Food Chem. 50 (2002) 4769. [62] B. Destrez, E. Bichon, L. Rambaud, F. Courant, F. Monteau, G. Pinel, J.P. Antignac, B. Le Bizec, Steroids 74 (2008) 803. [63] B. Destrez, G. Pinel, E. Bichon, F. Monteau, R. Lafont, B. Le Bizec, Rapid Commun. Mass Spectrom. 22 (2008) 4073. [64] H.F. De Brabander, S.S. Poelmans, R. Schilt, R.W. Stephany, B. Le Bizec, R. Draisci, S.S. Sterk, L.A. Van Ginkel, D. Courtheyn, N. Van Hoof, A. Macri, K. De Wash, Food Addit. Contam. 21 (2004) 515. [65] H.F. De Brabander, R. Verbeke, J. Chromatogr. 108 (1975) 141. [66] H.F.De Brabander, The determination of thyreostatic drugs in biological material, PhD Thesis, University of Ghent, Ghent, Belgium, 1986. [67] H.F. De Brabander, P. Batjoens, J. Van Hoof, J. Planar Chromatogr. 5 (1992) 124. [68] W.J. Blanchower, P.J. Hughes, A. Cannavan, M.A. McCoy, D.J. Kennedy, Analyst 122 (1997) 967. [69] B. Le Bizec, F. Monteau, D. Maume, M.-P. Montrade, C. Gade, F. Andr, Anal. Chim. Acta (1997) 340. [70] J.W. Pensabene, S.J. Lehotay, W. Fiddler, J. Chromatogr. Sci. 39 (2001) 195. [71] K. De Wasch, H.F. De Brabander, S. Impens, M. Vandewiele, D. Courtheyn, J. Chromatogr. A 912 (2001) 311. [72] G. Pinel, E. Bichon, K. Pouponneau, D. Maume, F. Andr, B. Le Bizec, J. Chromatrogr. A 1085 (2005) 247. [73] S. Floberg, K. Lanbeck, B. Lindstrm, J. Chromatogr. 489 (1980) 63. [74] Q.-H. Zou, Y. Liu, M.-X. Xie, J. Han, L. Zhang, Anal. Chim. Acta 551 (2005) 184. [75] L. Zhang, Y. Liu, M.-X. Xie, Y.-M. Qiu, J. Chromatogr. A 1074 (2005) 1. [76] G. Pinel, S. Mathieu, N. Cesbron, D. Maume, H. De Brabander, F. Andr, B. Le Bizec, Food Addit. Contam. 23 (2006) 974. [77] P. Batjoens, H.F. De Brabander, K. De Wasch, J. Chromatogr. A 750 (1996) 127. [78] G. Pinel, D. Maume, Y. Deceuninck, F. Andr, B. Le Bizec, Rapid Commun. Mass Spectrom. 20 (2006) 3183. [79] S.J. Lehotay, K. Mastovska, A. Amirav, A.B. Fialkov, T. Alon, P.A. Martos, A. DeKok, A.R. Fernandez-Alba, TrAC Trends Anal. Chem. 27 (2008) 1070. [80] A. Kaufmann, P. Butcher, K. Maden, M. Widmer, J. Chromatogr. A 1194 (2008) 66. [81] M. Becchi, R. Aguilera, Y. Farizon, M.-M. Flament, H. Casabianca, P. James, Rapid Commun. Mass Spectrom. 8 (1994) 304. [82] E. Bichon, F. Kieken, N. Cesbron, F. Monteau, S. Prvost, F. Andr, B. Le Bizec, Rapid Commun. Mass Spectrom. 21 (2007) 2613. [83] C. Buisson, M. Heberstreit, A. Preiss-Weigert, K. Heinrich, H. Fry, U. Flenker, S. Banneke, S. Prevost, F. Andr, W. Schanzer, E. Houghton, B. Le Bizec, J. Chromatogr. A 1093 (2005) 69. [84] G. Pinel, F. Andr, B. Le Bizec, J. Agric. Food Chem. 52 (2004) 407. [85] L. Bailly-Chouriberry, G. Pinel, P. Garcia, M.-A. Popot, Y. Bonnaire, B. Le Bizec, Anal. Chem. 80 (2008) 8340. [86] M.-H. Le Breton, S. Rochereau-Roulet, G. Pinel, G. Rychen, T. Goldmann, B. Le Bizec, Rapid Commun. Mass Spectrom. 22 (2008) 3130. [87] M.-H. Le Breton, S. Rochereau-Roulet, G. Pinel, N. Cesbron, T. Goldmann, B. Le Bizec, Anal. Chim. Acta 637 (2009) 121.