230 FOOD AND NUTRITIONAL ANALYSIS / Contaminants Boyce MC (2001) Determination of additives in food by capillary electrophoresis.

Electrophoresis 22: 14471459. Cheng QC and Wang J (2001) Simultaneous determination of articial sweeteners, preservatives, caffeine, theobromine and theophylline in food and pharmaceutical preparations by ion chromatography. Journal of Chromatography A 937: 5764. Dezman DL, Nagy S, and Brown GE (1986) Postharvest fungal decay control chemicals: treatments and residues in citrus fruits. Residue Reviews 97: 3792. Gonzalez M, Gallego M, and Valcarcel M (1999) Gas chromatographic ow method for the preconcentration and simultaneous determination of antioxidant and preservative additives in fatty foods. Journal of Chromatography A 848: 529536. Gordon MH (ed.) (1990) Amino acids and other nitrogencontaining compounds. In: Principles and Applications of Gas Chromatography in Food Analysis. Chichester: Ellis Horwood. Page BD and Charbonneau CF (1989) Liquid chromatographic determination of several antioxidants in dry foods. Journal of the Association of Ofcial Analytical Chemists 72(2): 259265. Pizzoferrato L, Quattnici E, and di Lucco G (1990) Evaluation of an HPLC method for the determination of sulphiting agents in foods. Food Additives and Contaminants 7(2): 189195. Wedzicha BL (1984) Chemistry of Sulphur Dioxide in Foods. London: Elsevier Applied Science. Yang M-H, Lin H-J, and Choong Y-M (2002) A rapid gas chromatographic method for direct determination of BHA, BHT and TBHQ in edible oils and fats. Food Research International 35: 627633.

Contaminants
M OKeeffe, The National Food Centre (Teagasc), Dublin, Republic of Ireland
& 2005, Elsevier Ltd. All Rights Reserved.

Introduction
The common understanding of food contaminants is very broad, covering topics as diverse as natural toxins through agrochemical and veterinary drug residues to industrial contaminants and extraneous matter such as glass and metal objects introduced during food processing. A European Community Council Regulation (No. 315/93) denes a food contaminant as: any substance not intentionally added to food which is present in such food as a result of the production (including operations carried out in crop husbandry, animal husbandry and veterinary medicine), manufacture, processing, preparation, treatment, packing, packaging, transport or holding of such food, or as a result of environmental contamination. Extraneous matter, such as, for example, insect fragments, animal hair, etc. is not covered by this definition. This definition of food contaminants encompasses a very broad range of chemical substances, many of which are covered by other articles in this Encyclopedia, particularly natural toxins, mycotoxins, pesticides, antibiotics, and metals (both individually and collectively). Consequently, this article will give an overview of chemical contaminants in foods, dealing particularly with agrochemical and veterinary drug

residues in food and residues of industrial contaminants, leaving the more in-depth treatment of particular classes of contaminants to the other articles. Chemical contaminants in food are of concern to the consumer, to food manufacturers, and to regulatory authorities. Contamination of food with chemicals may arise from the use of agrochemicals and veterinary drugs in food production or from environmental pollutants and industrial chemicals during food processing and packaging. A wide variety of analytical techniques are used for the detection and quantitation of contaminants in food ranging from microbiological tests through immunoassays to all forms of chromatography. Analyses for contaminants in foods are characterized by the requirements for intricate extraction/cleanup and for sensitivity. This article describes the methodologies used for contaminant residue analyses in food, covering screening and conrmatory techniques, the range of extraction and clean-up methodologies used, and both specific and multiresidue procedures.

Major Analytes
The major source of potential contaminants in food is the direct application of agrochemicals to plants and animals and the use of veterinary drugs for treating food-producing animals. Pesticides, including fungicides, herbicides, and insecticides, are routinely used in most food production to control pests that would otherwise destroy or reduce the food production. Because of the very large

FOOD AND NUTRITIONAL ANALYSIS / Contaminants 231

number of these substances used in food production, analyses are mainly multiresidue, directed to particular classes of pesticides such as organochlorine, organophosphorus, or carbamate substances. Veterinary drugs are used to treat animals for bacterial infections (antibiotics) and parasitic infestations (anthelmintics, coccidiostats), to enhance growth (antibiotics, anabolics, partitioning agents, thyrostats), to control fertility and reproduction (steroid hormones), or to alter behavior (tranquillizers and sedatives). Within each of the drug types there are different classes and methods for residue analysis are typically class specific (e.g., methods for the class of tetracycline antibiotics). In certain cases, multiresidue methods are available that are broader than specific classes. In many cases, maximum residue limits (MRLs) exist for veterinary drugs and the analytical methods must be capable of measuring below such MRLs. Since MRLs vary considerably, the sensitivity requirements of the various methods also vary widely. Typically, MRLs for veterinary drug residues vary from 1 mg kg 1 (ppm) to 1 mg kg 1 (ppb) (Table 1) or, in the case of banned substances, to no detectable residue. Where no detectable residue is specied, it is common to declare an action limit or a limit of decision for use by regulatory agencies; this is not strictly scientifically based and takes into account other aspects such as the sensitivity of a suitable conrmatory technique and policy requirements. A wide range of industrial contaminants may occur in food arising from environmental contamination, from food processing procedures, or from food packaging materials. These substances include polychlorinated biphenyls (PCBs), dioxins and heavy metals, nitrosamines, polycyclic aromatic hydrocarbons (PAHs), and acrylamide, and packaging material monomers and plasticizers.

Table 1 Maximum residue limits (MRLs) specied by EC Regulation (2377/90) for some veterinary drugs Veterinary drug Anti-infectious agents chemotherapeutics Sulfonamides Trimethoprim Target tissues MRL (mg kg 1)

Muscle, liver, kidney, milk Muscle, liver, kidney, fat, milk

100 50

Anti-infectious agents antimicrobials Benzylpenicillin, Ampicillin, Amoxicillin Oxacillin, Cloxacillin, Dicloxacillin Tetracyclines

Spiramycin Tylosin

Muscle, liver, kidney, fat Milk Muscle, liver, kidney, fat Milk Kidney Liver Eggs Muscle, milk Liver, kidney, fat Muscle, milk Muscle, liver, kidney, fat Eggs Milk

50 4 300 30 600 300 200 100 300 200 100 200 50

Antiparasitic agents Ivermectin Febantel, Fenbendazole, Oxfendazole

Liver Fat Liver Muscle, kidney, fat Milk Liver Muscle, kidney, fat

15100a 2040a 500 50 10 100 10

Levamisole

Agents acting on the nervous system Azaperone Carazolol

Sampling and Preparation

Samples and Sampling

Muscle, liver, kidney, fat Liver, kidney Muscle, fat

100 1525a 5

Sampling of food for residue analysis is governed by the rules that apply to all sampling, namely that the analytical sample should contain the residues in the forms and at the levels corresponding to those present in the lot of which the sample is representative. Sampling of food for residue analyses may be very directed (e.g., to establish whether residues of a particular pesticide are present in a particular shipment of fruit) or very broad (e.g., as part of a national monitoring program designed to establish the incidence of use of banned substances in meat-producing animals). For the latter circumstances exact

Glucocorticoids Dexamethasone, Betamethasone

Liver Muscle, kidney Milk

2 0.75 0.3

Depending on species.

sampling plans are required which are statistically based so that an accurate representation of the situation may be obtained from the analyses. The appropriate sample depends to a large extent on the nature of the food. In the case of plant foods such as grain, a composite sample is constructed

232 FOOD AND NUTRITIONAL ANALYSIS / Contaminants

from a number of samples taken at specied points throughout the load, as described in ISO Standard 7002:1986. In the case of plant foods such as fruit and vegetables, the composite sample is constructed from individual items selected according to a dened procedure and is described in ISO Standard 874:1980. In the case of foodstuffs of animal origin, sampling from milk and milk products is covered by a number of ISO standards 707:1997; 5538:1987; 8197:1988. The purpose of these dened procedures is to ensure representativeness of the sample for a variety of nonhomogeneous and homogeneous solid and liquid, bulk and packaged foods. For meat and edible tissues, samples taken from liver or kidney are usually representative, although there may be differences in residue content between the two sections of the kidney (medulla and cortex). Samples from carcasses are normally taken from the same position (muscle or fat) on each carcass and chosen so as to reduce commercial loss as much as is possible. Samples of body uids or excreta are often used to monitor illicit use of prohibited substances or for indirect determination of the presence of contaminants in foods of animal origin. For example, analyses of urine, blood, bile, feces may be used to establish whether specied veterinary drugs have been used on a particular animal. Where well-established ratios exist between the levels of drug residue in body uids and edible tissues (e.g., sulfadimidine levels in blood or urine and in muscle or kidney), analyses of body uid samples may be used to indicate the levels in edible tissues.
Sample Storage

residues. However, homogenization may increase the contact between contaminant residues and endogenous enzymes in the sample matrix (e.g., chloramphenicol in liver, the effect of b-lactamase on penicillin). Sometimes, addition of preservatives to the sample, such as the antimicrobial sodium azide or enzyme inhibitors, may be useful but only where these agents do not interfere with the subsequent analysis.
Sample Preparation

The analytical sample is generated as a subsample from the composite or bulk sample. In the case of fruit and vegetables, cutting/chopping of the composite sample, mixing, and subsampling is required. Grains are normally subsampled, and then ground before analysis. Milk and milk powders, being already very homogeneous, usually require simple mixing before an aliquot is taken for analysis. Milk products, such as cheese, require chopping/grinding and mixing before subsampling. Tissue samples are minced/homogenized and subsampled. Urine samples are mixed and may be centrifuged, to remove solid particles, prior to sampling for analysis. Blood may be clotted and serum removed or collected in heparinized containers and the blood cells separated by centrifugation and plasma removed.

Residue Analysis
Sample Extraction

Samples of foodstuffs taken for analysis for contaminant residues must be stored in such a way as to prevent deterioration of the sample and reduction in the residue content. In the case of relatively dry samples, such as milk powders, or samples protected by an outer skin, such as fruit and vegetables, storage may be at ambient temperature for short periods or at refrigeration temperature. In the case of most animal food materials, storage at refrigeration temperature is satisfactory if analyses are to commence in the short term, but frozen storage (p 181C) is recommended, and is essential for prolonged storage. Even lower storage temperatures of 401C to 701C may be recommended for certain contaminant residues. Homogenization and division of the sample, for example, where required for contra-analysis, may be performed before freezing to avoid the cycle of thawingfreezingthawing, which may be detrimental to the stability of some contaminant

The challenge to the residue analyst is to extract the residue of interest from the sample and to purify the resultant extract to the extent necessary to allow for unequivocal identication and accurate quantication by the selected determination technique. The sample extraction and cleanup procedures are dened, therefore, by the requirements of the determination step. These requirements can vary from as little as no extraction, e.g., use of an intact piece of tissue on an agar plate with a test organism for determination of inhibitory substances; or minimal treatment, e.g., dilution of urine and direct assay for residues of growth enhancing substances by immunoassay; to very extensive extraction and cleanup procedures, e.g., multiple extractions with organic solvent, ltration of extracts, evaporation of combined extracts, liquidliquid partitioning into acid, alkalization, and liquidliquid partitioning with organic solvent, neutralization, and multiple liquidliquid partitioning into organic solvent, evaporation and derivatization, as required in the Association of Ofcial Analytical Chemists (AOAC) method for determination of sulfamethazine residues in tissues by gas chromatographymass

FOOD AND NUTRITIONAL ANALYSIS / Contaminants 233

spectrometry (GCMS) or gas chromatography-electron capture detection (GCECD). With the exception of some limited assays where determination can be made on the intact or diluted sample, most residue determination procedures commence with extraction of the residue(s) from the sample matrix. Organic solvents are commonly used, with selection of the appropriate solvent being based, largely, on the polarity of the residues of interest. Organic solvents cause denaturation of protein and decrease the interaction between residues and proteins. Aqueous extraction is also used particularly for relatively polar residues and has the advantage that lipids present in the sample matrix are not extracted. Modication of the sample homogenate through pH changes may be necessary to make the residue of interest available to the organic solvent extractant; by adjusting pH, the residues may become largely nonionized and hence more readily extracted. Typically, solventsample ratios of 2:1 to 10:1 are used in the extraction procedure and repeated extractions (two or more) are carried out on the sample. The extraction procedure may be by mechanical blending, by manual (separatory funnel), or mechanical (shaker) mixing, by ultrasonics, or by accelerated solvent extraction, which provides efcient extraction under controlled temperature/pressure conditions. In the case of animal tissues, protease digestion prior to extraction may be used to provide better accessibility of the solvent to the residue molecules and to release residues weakly bound to proteins. Where conjugated forms of residues arise in samples, a deconjugation step using enzymes is undertaken prior to extraction; many residues occur as glucuronide or sulfate conjugates, which, being hydrophilic, are difcult to extract. Having extracted the residue into the extractant, it may be separated from the matrix material by centrifugation, ltration, freezing of the aqueous phase, or binding of the aqueous phase using anhydrous salts. Where the extract is an organic solvent, the volume may be reduced or brought to dryness, by controlled heating and distillation of solvent or evaporation under nitrogen or vacuum, prior to cleanup of the extract. Where the extract is aqueous, transfer of the residues to an organic solvent by liquidliquid partitioning or isolation of the residues directly from the aqueous phase by a procedure such as immunoafnity chromatography may be the rst step in cleanup of the extract.
Extract Cleanup

Cleanup of the extract is undertaken by a range of procedures such as liquidliquid partitioning or

column chromatography. Liquidliquid partitioning consists of the intimate mixing of two immiscible solvents such that the residue of interest is accumulated selectively in one phase while certain of the contaminating coextractives are accumulated into the other phase. The classical procedure for liquid liquid partitioning is by manual mixing in a separatory funnel. A form of liquidliquid partitioning, sometimes termed liquidsolid partitioning, may also be undertaken on columns or cartridges packed with inert material, such as diatomaceous earth. An aqueous extract is introduced on to the column and distributes over the packing material into which the water is absorbed. When the second solvent (phase) is added to the column, it mixes intimately with the inert material and the residue is extracted from the column. Some cleanup procedures involve a number of liquidliquid partitioning steps, typically to remove fat and nonpolar coextractives in one step and polar coextractives in another. Movement from one liquid liquid partitioning step to another may be by separation of phases, by evaporation of one phase and resolubilization in another solvent, by freezing of an aqueous phase and decanting of an organic phase, or by changing the pH to alter the afnity of the residue of interest for the aqueous phase. Column chromatography is very widely used as a cleanup procedure. The classical glass column chromatography based on a variety of size-exclusion, polar (normal), nonpolar (reversed phase), and ionexchange materials has been superseded to a considerable extent by microcolumn cartridges available commercially with a very broad range of packing materials or sorbents. This cartridge-based chromatography, known as solid-phase extraction (SPE) is found as a step in many cleanup procedures for residues in foodstuffs. There have been considerable developments during the last decade both in the range of sorbents and the formats available for SPE. In addition to the classical modied silica reversedphase sorbents a number of polymeric sorbents have been developed that have multiphase characteristics. Formats for automated SPE, including 96-well plate systems, are available. Automated systems for size-exclusion, or gel permeation, chromatography are very popular, particularly in pesticide residue analyses. Another form of column chromatography, based on the use of specific antibodies to the analyte(s) of interest, is immunoafnity chromatography (IAC). In this technique, antibodies specific for one or more analytes are covalently bound to an activated support such as Sepharose, which is then packed in a column (Figure 1). This technique is particularly suited for coupling

234 FOOD AND NUTRITIONAL ANALYSIS / Contaminants

with aqueous extraction. Dialysis, particularly in an automated format, has been used to separate the analytes of interest from larger-sized coextractives in an aqueous extract. However, this technique has not been applied very widely in residue analysis. Column-switching techniques, whereby semipuried sample extracts may be cleaned-up and/or concentrated on an SPE-type precolumn, prior to automated switching into the ow to a high-performance liquid chromatography (HPLC) analytical column, are being used and further developed (Figure 2).

Concentration

Elution

Immunoaffinity column Technican pump

+ CH3OH

Some techniques that combine the properties of extraction and cleanup are supercritical uid extraction (SFE) and matrix solid-phase dispersion (MSPD). Supercritical uids, i.e., at a temperature and pressure in excess of their critical point, have unique properties for selective extraction of analytes from a sample. Solid samples are mixed with an inert dispersant, such as hydromatrix, and the mixture packed into the cell of the SFE apparatus. The sample is extracted with supercritical CO2, with or without addition of organic modier, and the extracted analytes may be collected inline or offline on suitable adsorbents (Figure 3). Further cleanup of the sample extract may be performed using SPE. MSPD is based on intimate mixing of animal tissue sample with a bonded silica, such as C18, and packing of the blended material into a column from which interferences can be eluted by washing with solvents and the analytes eluted using a selective solvent.

Residue Determination
= CAP = Extratcted meat components Monoclonal antibody to CAP = Covalently bound to CNBr Activated Sepharose 4B

Figure 1 Schematic representation of immunoafnity column cleanup. CAP, chloramphenicol (an antimicrobial). (Reproduced with permission from Haagsma N (1990) Sample pretreatment in drug residue analysis. In: Haagsma N, Ruiter A, and CzedikEysenberg PB (eds.) Residues of Veterinary Drugs in Food, Proceedings of the EuroResidue Conference, Noordwijkerhout, The Netherlands, 2123 May 1990, p. 47. Utrecht: Rijksuniversiteit.)

Residues are determined in the puried extracts by chromatographic or immunochemical techniques. In the chromatographic systems, thin-layer chromatography (TLC), liquid chromatography (LC), and GC, the analytes are separated on plates or columns and determined by colorimetry, by spectrophotometry (ultraviolet (UV), infrared (IR, Fourier transform infrared (FTIR)), by uorescence, by selective detectors (in GC analysis: ECD, ame photometric (FPD), nitrogen/phosphorus (NPD, TSD), etc.), or by MS. Separations may also be achieved by

Pump 2

Pump 2

Analytical column

Injector Pump 1 (A) (B)

Injector Pump 1

Figure 2 Column switching ow diagram: (A) preconcentration and (B) desorption and chromatographic separation. Reproduced with permission from OKeefe, Residue Analysis in Food, p. 78 (2000) Taylor and Francis.

Analytical column

Precolumn

Precolumn

FOOD AND NUTRITIONAL ANALYSIS / Contaminants 235

Table 2 Food analysis performance assessment scheme (FAPAS) contaminants in food Series Title 2 Veterinary drug residues Analytes Sulfonamides, tetracyclines, chloramphenicol, bagonists, nitrofurans, malachite green Aatoxins B, G, M1 Organochlorine pesticides and PCBs Dioxins, marine toxins, PAHs Pb, Cd, As, Hg, Sn, Fe, Cu, Zn Organophosphorus and pyrethroid pesticides Semicarbazide Nitrate, nitrite Patulin Ochratoxin A Pesticides Deoxynivalenol, fumonisins, zearalenone

Shut-off valve

Inline alumina polar analyte trap Sample + hydromatrix

4 5 Aatoxins OC pesticides and PCBs Environmental contaminants Metallic contaminants OP and pyrethroid pesticides Specific migration Nitrate analysis Patulin Ochratoxin A Pesticides Fusariums

Offline alumina SPE column

6 7

Hydromatrix
9 12 15 16 17 19 22

SF CO2
Figure 3 SFE system congured for offline analyte collection using a standard 6 ml SPE column and inline using a packed sorbent bed. Reproduced with permission from OKeefe, Residue Analysis in Food, p. 96 (2000) Taylor and Francis.

capillary zone electrophoresis, coupled with UV or MS detectors. In the case of immunoassays, the binding of the analyte of interest by an appropriate antiserum is the basis of the determination procedure, using radiolabeled, enzyme, or uorescent markers. Apart from the classical immunoassays in tube or 96-well plate formats, a variety of dip-stick or sol particle systems are available for rapid testing. There have been considerable advances in the application of sensor technologies to high throughput immunoassays for contaminant residues, particularly using surface plasmon resonance (SPR). In SPR-based biosensor systems, antigenantibody binding reactions on a thin gold lm surface may be measured as shifts in the angle of light. Fully automated SPR-based bioTM sensor systems are available e.g., Biacore on which a broad range of rapid analyses have been developed for drug classes such as sulfonamides, coccidiostats, and beta-agonists.
Quality Assurance

establishment as a routine method in the laboratory, ongoing quality control is maintained by assay of negative and positive controls and/or of fortied samples in the assay. These controls may be generated within the laboratory and are supplemented by external certied reference materials (CRMs), certied for specied levels of the analyte(s) of interest. Such CRMs are produced by the EC Institute for Reference Materials and Measurements, by international and national agencies, such as NIST, and by commercial companies. Quality assurance for residue analysis may be provided also by involvement of a laboratory in appropriate prociency testing schemes, operated according to ISO standards (ISO Guides 43-1, 43-2: 1997). One example of such is the Food Analysis Performance Assessment Scheme organized by the Central Science Laboratory, UK, which produces samples on a number of occasions each year covering a particular analyte or class of analytes (Table 2). Quality assurance of residue analysis is frequently maintained within a laboratory accreditation framework, to standards such as ISO 17025.

Methods are validated using samples fortied with the analyte of interest at appropriate levels (i.e., normally levels similar to levels found in the samples), levels related to MRLs for permitted substances, or levels related to minimum required performance limits for prohibited or unapproved substances. Following validation of the method and its

Quantitative Methods
Pesticides

This topic is covered in detail elsewhere in the Encyclopedia so only brief mention of relevant methods will be made here. A number of multiresidue procedures for organochlorine (OC), organophosphorus

236 FOOD AND NUTRITIONAL ANALYSIS / Contaminants

(OP), and carbamate pesticides are available as standard methods by the AOAC covering both nonfatty and fatty foods. These methods are based on solvent extraction and a variety of column chromatographic cleanup procedures with determination by GC using selective detectors. The International Dairy Federation has approved similar methods for OC and OP pesticide residues in milk and milk products. The AOAC have standard methods for specific pesticides or groups of pesticides that are approved for certain foodstuffs; these methods involve colorimetric, spectrophotometric, or gas chromatographic determination. In the case of methylcarbamate pesticides, a liquid chromatographic method is approved.
Veterinary Drugs

There are relatively few standard methods for veterinary drug residues although very large numbers of methods are available for the different classes of compounds. The United States Department of Agricultures (USDA) Food Safety and Inspection Service publish a Chemistry Laboratory Guidebook that, in addition to some AOAC methods, contains a number of validated methods. Within the European Union, an alternative approach has been adopted to that of standard methods. Commission Decision 2002/657/EC species a range of performance criteria for analytical methods to be used to test for the presence of

veterinary drug and other residues in foods of animal origin, under Council Directive 96/23/EC. This decision denes suitable analytical methods and the performance criteria that should be determined for methods, depending on whether they are classied as qualitative or quantitative and screening or conrmatory. Validation procedures are specied for performance characteristics such as specicity, trueness, ruggedness, stability, calibration, recovery, repeatability, reproducibility, decision limit, and detection capability. Conrmatory methods should preferably be based on MS, providing direct information on the molecular structure of the analyte(s), especially for residues of prohibited substances or substances for which no MRLs are set. However, validated chromatographic methods with specific detectors (e.g., photodiode array, uorescence) or using two or more different chromatographic separation systems may be used, particularly for contaminants for which MRLs are specied. There have been major developments in LC coupled with MS (LCMS/MS) during the last 510 years and this technology has been applied very widely for veterinary drug residue analysis (Figure 4). LCMS/MS is particularly suitable for veterinary drug residues in that derivatization, commonly required for GCMS analysis, is not necessary. In addition, it can provide robust and sensitive determination.

100 GCMS LCMS 80 Number of presentations

60

40

20

0 ER (1990) ER (1993) IS (1994) ER (1996) IS (1998) ER (2000) IS (2002) European residue conferences, 1990-2002 (ER EuroResidue, The Netherlands; IS International Symposium, Belgium)
Figure 4 Increase in use of mass spectrometry in veterinary drug residue analysis (19902002), presentations at major European conferences.

FOOD AND NUTRITIONAL ANALYSIS / Contaminants 237

Antibiotics The AOAC has listed methods for sulfamethazine residues in swine tissues with determination either by GCMS or GCECD of methylated derivatives and for sulfamethazine (and for the class of sulfonamides) in milk with determination by HPLCUV. There is an AOAC method for the class of sulfonamide antimicrobials in animal tissues using solvent extraction and liquid partitioning with determination by TLC and uorimetric scanning. For analysis of tetracyclines, AOAC describes methods based on buffer extraction from tissue samples and SPE (C18) cleanup, or metal chelate afnity binding from milk samples, with determination in both cases by HPLCUV. USDA/FSIS methods include: (1) a method (similar to the AOAC GCMS method for sulfamethazine) for conrmation of sulfonamide residues in edible tissues using solvent extraction and multiple liquid partitioning with determination of the methylated derivatives by GCMS; (2) methods for determination and conrmation of chloramphenicol in muscle by solvent extraction, liquid partitioning, and determination of the trimethylsilane (TMS) derivative by GCECD and GCMS, respectively; and (3) a method for determination of the beta-lactam antibiotic amoxicillin by aqueous extraction, cleanup by tricarboxylic acid precipitation, and ether extraction and formation of a uorescent derivative for determination by LC. Methods for the determination of residues of most of the important classes of antimicrobials (sulfonamides, chloramphenicol, tetracyclines, beta-lactams, quinolones, and nitrofurans) have been developed based on HPLC or LCMS/MS determination. These methods are applicable to edible tissues, milk, eggs, urine, and other matrices. The methods typically involve sample workup procedures of solvent extraction and single or multiple SPE cleanup prior to chromatographic determination. For example, extraction of sulfonamides and chloramphenicol may be achieved using aqueous methanol, acetonitrile, or ethyl acetate with cleanup using SPE on reversed phase (e.g., C18) and/or cation exchange (e.g., benzenesulfonic acid) sorbents. In the case of tetracyclines, extraction is commonly performed with EDTA/McIlvane buffer and cleanup on C18, TM Oasis HLB and/or SCX (propylsulphonic acid) SPE cartridges. Determination may be by HPLC with UV detection, by uorescence detection following postcolumn derivatization, or by MS. In the case of the prohibited nitrofurans, which occur as proteinbound metabolites in edible tissues, methods are directed at releasing these metabolites by acid hydrolysis, derivatization with nitrobenzaldehyde, extraction with ethyl acetate, and determination by LCMS/MS.

Growth-enhancing substances An AOAC method is approved for the growth promoter melengestrol acetate in animal tissues by solvent extraction of fatty tissue and cleanup on a Florisil column with determination by GCECD. Numerous methods has been validated for the determination of anabolic agents (natural and synthetic estrogens, androgens, and gestagens) in edible tissues and in other samples that can be used for food monitoring such as urine, bile, and feces. These methods involve determination of analytes by LCMS/MS, by GCMS of suitable derivatives, or by immunological assay, with or without prior LC separation. Typical steps in the extraction, cleanup, and isolation of the analytes are the following: 1. enzymatic hydrolysis of conjugated residues; 2. mechanical homogenization, or enzymatic digestion with proteolytic enzymes; 3. solvent extraction direct, or on diatomaceous earth; 4. cleanup using SPE, e.g., coupled C18 and amino cartridges for anabolic steroids in urine; 5. analyte isolation by IAC or HPLC; 6. determination of residues by LCMS/MS, or by GCMS following derivatization. Methods for determination of beta-agonists (repartitioning agents) in biological samples involve aqueous extraction using ultrasonics, enzymatic hydrolysis of conjugated residues, solvent extraction, isolation of analytes by IAC, and determination by LCMS/MS, or of derivatives by GCMS. For example, a semiautomated quantitative method is described which is capable of screening and conrmation of 22 steroids in urine. This method involves enzymatic deconjugation followed by dual SPE on C18 and amino cartridges, using an automated system (ASPEC XL4, Gilson Ltd.). Screening is undertaken by LCMS/MS, with an electrospray interface, and the column efuent is split between the mass spectrometer and a fraction collector. Where the presence of steroid residues is determined, conrmation is achieved by reanalysis of the collected fraction using either LCMS/MS, while monitoring additional transition products, or GChigh resolution MS, after derivatization. Anthelmintics The major classes of anthelmintics, or antiparasitic drugs, are benzimidazoles and macrocyclic lactones (avermectins and milbemycins). USDA/FSIS have approved a method based on LC/ uorescence detection for determination of the anthelmintics albendazole and ivermectin in tissues (albendazole extracted with ethyl acetate and cleanup

238 FOOD AND NUTRITIONAL ANALYSIS / Contaminants

by liquid partitioning and SPE on a C18 cartridge; ivermectin extracted with iso-octane and cleanup by liquid partitioning, formation of a uorescent derivative by heating with imidazole reagent and purication by SPE on a silica cartridge). Multiresidue methods for benzimidazoles and macrocyclic lactones are applied in many testing laboratories. For example, benzimidazoles may be determined by extraction from tissue with ethyl acetate, removal of fat by liquid/liquid partitioning between acidic ethanol and hexane, cleanup by automated SPE on C18 or C8 cartridges, and HPLC chromatography with UV detection at 298 nm. In the case of the macrocyclic lactones, samples are typically extracted using acetonitrile and modied with watertriethylamine prior to purication on C8 or C18 SPE cartridges. Residues may be determined, after derivatization, by HPLC with uorescence detection (excitation 365 nm, emission 475 nm) or determined without derivatization by LCMS/MS. Other veterinary drugs Other veterinary drugs of importance are the anticoccidial feed additives such as the ionophores, narasin, monensin, salinomycin, apramycin, lasalocid, and nicarbazin. The USDA/ FSIS have a method for the major ionophores in tissue samples, which is based on purication of sample extracts by silica gel, alumina, or ion-exchange column chromatography and determination by TLC with detection by bioautography. A number of alternative methods based on immunoassays, biosensor technology, and HPLC have been developed. Tranquillizers and beta-blockers are important potential contaminants in edible tissues, particularly of pigs, where they may be used to counteract the stress susceptibility of some breeds during transport, which results in poor quality meat and even mortality. Some of these substances are prohibited for use in foodproducing animals while others have MRL values set. A simple method for seven of these compounds (azaperol, azaperone, carazolol, chlorpromazine, acepromazine, propionylpromazine, xylazine) involves extraction of kidney or muscle samples with acetonitrile, modication of the extract with a sodium chloride solution (10%), and SPE on an Oasis HLBTM cartridge. The eluate (acetonitrile) is evaporated, reconstituted in aqueous acetonitrile, and residues determined by LCMS/MS, with atmospheric pressure chemical ionization. Two additional steps are incorporated into this method to increase the numbers of sample extracts that may be analyzed without affecting the mass spectrometer source; rst, an inline column (Biomatrixs) is used ahead of the analytical column to remove any macromolecules in the sample extracts and, second, a switching value is

used ahead of the mass spectrometer that allows the eluate from the analytical column to pass through the mass spectrometer only during analyte elution.
Industrial and Processing Contaminants

Typical contaminants of concern are PCBs and polychlorinated dibenzo-p-dioxins (dioxins), which are highly persistent environmental contaminants arising from industrial processes, detergents, and disinfectants (especially in milk) that are used in cleaning processes, process contaminants such as PAHs and acrylamide, and food packaging monomers and residual solvents from extraction processes or printed packaging material. The AOAC standard method for PCBs in foods is similar to that for organochlorine pesticide residues, involving extraction with lipophilic solvent and cleanup by liquid partitioning and column adsorption chromatography with determination by GC with ECD and/or potassium chloride thermionic detector. Current methods for dioxins and coplanar PCBs involve determination by GC coupled with high-resolution MS, to achieve the sensitivity and specicity required. Typically, the sample preparation consists of fat/lipid extraction and isolation of dioxins/PCBs by gel permeation chromatography with extract purication by column chromatography. A method that may be used to screen samples for the presence of dioxins/PCBs is the CALUXs cellbased assay. This bioassay technique Chemical Activated LUciferase eXpression is based on use of rat or mouse hepatoma cells, modied to show an increased production of luciferase in response to binding of dioxins and dioxin-like PCBs to a particular cell receptor. Because many detergents and disinfectants contain phosphate, chloride, or iodide, it is difcult to determine the presence of these substances against the naturally occurring levels of these salts in milk or other food. Contamination with detergents or disinfectants is suspected, therefore, by the determination of much higher levels than normally would be found in the food. Ion-selective electrodes are commonly used. Food packaging monomers and solvents are assayed by headspace GC linked with various detection systems such as FID, AFID, and MS. Nitrosamines are carcinogenic substances that may form in foods preserved with high levels of sodium nitrite. The AOAC have approved two methods for analysis of nitrosamines in fried bacon, one for volatile N-nitrosamines by mineral oil vacuum distillation and extraction into dichloromethane, and the other for N-nitrosopyrrolidine by extraction

FOOD AND NUTRITIONAL ANALYSIS / Contaminants 239

from a column of Celite with dichloromethane. Nitrosamine levels in the extracts are determined by chemiluminescence, using a gas chromatograph tted with a thermal energy analyzer. Acrylamide was identied as a contaminant in heat-treated starchy foods (e.g., potatoes, bread, snack foods) in 2002. Methods for determination of acrylamide in food are based mainly on GCMS (without derivatization or with derivatization by bromination) or LCMS/MS. Acrylamide is generally extracted from the food matrix with water, sometimes with defatting using hexane, and cleanup of the extract is achieved using mixed-mode or multiple SPE cartridges.

Screening Methods
Apart from the methods described above that are used for monitoring of residues in food, meat and milk supplies are assayed routinely for the presence of antimicrobial or inhibitory substances. In the EC the microbiological Four Plate Test method is used widely as a screening technique for antimicrobial residues in edible tissues (mainly kidney). The principle of this method is a measurement of the extent of inhibition of the growth of test organisms due to the presence of antimicrobial residues in the tissue sample. Modications and extensions of this methodology have been developed such as a single plate New Dutch Kidney Test and a 12plate antimicrobial residue identication procedure used in Denmark. A tube-based test, the PremitestTM, also uses the principle of inhibition of growth of a test organism but with a colorimetric endpoint, similar to the widely used tests for antimicrobial residues in milk. A sample of uid is obtained from the tissue sample and incubated in the tube, containing agar seeded with the test organism and a colored substrate sensitive to lowering of pH caused by microbial growth. Presence of inhibitory substances (antimicrobial residues) in the tissue uid prevents growth of the test organism with no consequent lowering of pH and color change. The USDA/FSIS have a number of rapid tests for determining the presence of antimicrobial substances in carcasses or in live animals. The swab test on premises and the live animal swab test are developed for kidney (or liver) and urine samples, respectively, using an inhibition agar plate assay. The sulfa on site test is used on farm or at slaughter plant to screen samples of serum, urine, or feed for sulfadimidine residues. The test is based on TLC and uses well-dened ratios between concentrations of sulfadimidine in serum or urine and in edible tissues

to establish whether the concentration in edible tissues is above the MRL of 100 mg kg 1. Enzyme immunoassays, in a card-test format, are also used for screening carcasses by testing of urine for specific antimicrobial residues such as sulfadimidine and chloramphenicol. The sol particle immunoassay uses adsorption of specific antibodies to dyed colloidal particles (e.g., carbon) as a mechanism for making directly visible the presence of residues (i.e., antigen) in the sample. A lateral ow membrane device is used to identify residue-positive samples through the absence of color formation at the test capture zone (Figure 5). Milk samples are screened for the presence of antimicrobial residues with inhibition tests in different formats and using a variety of test organisms. Specific antimicrobial identication tests, based on more extensive inhibition tests, enzymatic tests, immunoassays, and receptor assays (Table 3) are also used for milk samples. These tests show considerable

Positive result

Figure 5 Lateral ow membrane device for residue testing. S, sample well; T, test line; C, control line. (Reproduced from OKeeffe M, Crabbe P, Salden M, et al. (2003) Preliminary evaluation of a lateral ow immunoassay device for screening urine samples for the presence of sulphamethazine. Journal of Immunological Methods 278: 117126.)

C T S

Negative result

T S

240 FOOD AND NUTRITIONAL ANALYSIS / Contaminants

Table 3 Detectable concentrations (mg ml 1) of antimicrobials in milk for various identication tests Antimicrobial (class) b-Lactams Tetracyclines Chloramphenicol Aminoglycosides Macrolides Sulfonamides Penzyme 0.0030.04 Delvotest P 0.0030.02 0.30.7 8 0.515 0.45 Three-plate 0.0030.03 0.4 2 0.320 0.1 0.11 Six-plate 0.0040.04 0.40.5 15 1328 2 1 Charm II 0.0030.03 0.10.5 0.1 0.1 0.02 0.01

variation in sensitivity for different classes of antimicrobials and for individual antimicrobials within a class.

Limitations of Current Analytical Procedures

The aim and purpose of residue analysis for contaminants in food is to ensure the safety of food to the consumer. Because of the large number of potential contaminants and the relatively difcult, expensive, and lengthy procedures required for residue analysis in complex matrices such as food, it is difcult to achieve this aim. Multiresidue procedures, use of a combination of screening and conrmatory techniques, use of robotics and automated systems, and soundly based statistical sampling are all factors that contribute to enhanced control of food safety. Particular problems are (1) the need for sensitive, multiresidue techniques, giving equivalent, high recovery for a range of analytes of diverse chemistry, (2) the difculties associated with extracting and determining polar metabolites and/or bound residues, and (3) the inability of the analytical procedures to keep pace with food manufacturing. Realtime and/or online analysis, based on highly automated systems, has the potential for solving some of these problems. In other aspects of food analysis (e.g., near-infrared spectroscopy for online compositional analysis) such systems have been developed. Biosensor-based techniques offer good possibilities here. In the area of contaminants in food, developments in automated systems using rapid extraction and cleanup based on SPE, IAC, etc., and determination by MS offers the possibility for real-time and multiresidue analyses.
See also: Dioxins. Extraction: Solvent Extraction Principles; Supercritical Fluid Extraction. Food and Nutritional Analysis: Antioxidants and Preservatives; Pesticide Residues; Mycotoxins; Packaging Materials. Gas Chromatography: Overview. Hormones: Steroids.

Immunoassays, Applications: Food. Liquid Chromatography: Food Applications. Nitrosamines. Polychlorinated Biphenyls. Quality Assurance: Reference Materials; Production of Reference Materials. Sampling: Theory. Sensors: Overview. Surfactants and Detergents.

Further Reading
Botsoglou NA and Fletouris DJ (2001) Drug Residues in Food. New York: Dekker. Gilbert J (ed.) (1996) Progress in Food Contaminant Analysis. London: Blackie Academic & Professional. Haagsma N and Ruiter A (eds.) (1996) Residues of Veterinary Drugs in Food, Proceedings of the EuroResidue III Conference, Veldhoven, The Netherlands, 68 May 1996. Utrecht: University of Utrecht. Moats WA and Medina MB (eds.) (1996) Veterinary Drug Residues: Food Safety. Washington DC: American Chemical Society. OKeeffe M (ed.) (2000) Residue Analysis in Food: Principles and Applications. Amsterdam: Harwood Academic Publishers. Stephany RW and Bergwerff AA (eds.) (2004) Residues of Veterinary Drugs in Food, Proceedings of the EuroResidue V Conference, Noordwijkerhout, The Netherlands, 1012 May 2004. Bilthoven: National Institute of Public Health and the Environment (RIVM). Turnipseed SB and Long AR (eds.) (1998) Analytical Procedures for Drug Residues in Food of Animal Origin. West Sacramento: Science Technology System. Van Ginkel LA and Ruiter A (eds.) (2000) Residues of Veterinary Drugs in Food, Proceedings of the EuroResidue IV Conference, Veldhoven, The Netherlands, 810 May 2000. Bilthoven: National Institute of Public Health and the Environment (RIVM). Van Peteghem C (ed.) (1998) Third International Symposium on Hormone and Veterinary Drug Residue Analysis, Proceedings of Symposium, Bruges, Belgium, 25 June 1998. The Analyst 123(12): 24012801. Van Peteghem C (ed.) (2003) Fourth International Symposium on Hormone and Veterinary Drug Residue Analysis, Proceedings of Symposium, Antwerp, Belgium, 47 June 2002. Analytica Chimica Acta 473(12): 1196; 483(12) 1432. Watson DH (ed.) (2001) Food Chemical Safety, Volume 1: Contaminants. Cambridge: Woodhead Publishing.