Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 PRINCIPLES AND METHODS FOR THE RISK ASSESSMENT OF CHEMICALS IN FOOD 2 3 CHAPTER 3. CHEMICAL CHARACTERIZATION, ANALYTICAL 4 METHODS AND THE DEVELOPMENT OF SPECIFICATIONS FOR 5 HAZARD IDENTIFICATION AND CHARACTERIZATION AND RISK 6 MANAGEMENT 7 8 3.1 INTRODUCTION...... 1 9 3.2 CONSIDERATIONS FOR LABORATORIES ...... 2 10 3.2.1 Guidance from Codex Alimentarius Commission (CAC/GL 27 - 1997)...... 2 11 3.2.2 Best practices in analytical measurement ...... 2 12 3.3 CONSIDERATIONS FOR SELECTION OF VALIDATED ANALYTICAL METHODS ...... 3 13 3.3.1 General considerations ...... 3 14 3.3.2 Multilaboratory method trials and collaborative studies ...... 3 15 3.3.3 Routine regulatory methods...... 4 16 3.3.4 Validation of methods in a single laboratory (the criteria approach)...... 4 17 3.3.5 Method validation process...... 6 18 3.4 DEVELOPMENT OF ANALYTICAL METHODS...... 6 19 3.4.1 Defining the analytical measurement requirements (fitness for purpose)...... 6 20 3.4.2 Analytical performance characteristics...... 6 21 3.5 GENERAL CONSIDERATIONS FOR USE OF ANALYTICAL METHODS IN A REGULATORY PROGRAMME FOR 22 FOODS...... 7 23 3.6 FOOD ADDITIVES—SPECIFICATIONS ...... 7 24 3.6.1 General considerations ...... 7 25 3.6.2 Formulation of specifications and information requirements ...... 9 26 3.6.3 Stability and fate of additives in food ...... 10 27 3.6.4 Methods of analysis...... 10 28 3.7 PESTICIDES—SPECIFICATIONS ...... 11 29 3.7.1 General considerations ...... 11 30 3.7.2 Identity and purity ...... 13 31 3.7.3 Stability...... 14 32 3.7.4 Physical and chemical properties...... 14 33 3.7.5 Analytical methods ...... 14 34 3.7.6 Analytical problems and challenges...... 16 35 3.8 VETERINARY DRUG RESIDUES...... 17 36 3.8.2 Analytical methods ...... 17 37 3.8.3 Analytical problems and challenges...... 18 38 3.9 CONTAMINANTS ...... 18 39 3.9.1 Sampling plans ...... 19 40 3.9.2 Analytical methods ...... 20 41 3.9.3 Reactions and fate of contaminants in food...... 20 42 3.10 SUBSTANCES CONSUMED IN LARGE AMOUNTS ...... 21 43 3.11 REFERENCES ...... 21 44 45 3.1 Introduction 46 Chemical characterization plays a critical role in risk assessment. Risk characterization brings 47 together the results of hazard characterization with the estimated human exposure, and these 48 should be based on data for the same chemical entity or a more complex material that has 49 been properly characterized and quantified. The report of the WHO Scientific Group on 50 Procedures for Investigating Intentional and Unintentional Food Additives indicated in 1967 51 that “adequate specifications for identity and purity should be available before toxicological 52 work is initiated. Toxicologists and regulatory bodies need assurance that the material [to be] 53 tested corresponds to that to be used in practice” (WHO, 1967). 54 Analytical methods have to be fit for purpose with respect to defining 1) the nature 55 and chemical purity of the material investigated during in vitro and in vivo hazard

1 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 identification and characterization studies and 2) the concentrations of the chemical in foods 2 in relation to regulatory limits or exposure surveys. 3 The dietary exposure (intake) of any chemical depends on the concentrations present 4 in foods and the amounts of the relevant foods consumed. For residues of pesticides and 5 veterinary drugs, the concentrations used for exposure estimates are the MRLs and maximum 6 limits (MLs) established by the JMPR and JECFA, respectively. The derivation of MRLs is 7 discussed in chapter 8. For food additives, the exposure assessment and risk characterization 8 are based on the proposed use levels for an additive in different foods. Risk characterization 9 of contaminants differs from that for other types of chemical, because there is not an 10 approved, accepted or proposed concentration in food and the exposure estimate has to use 11 actual analytical data rather than a recommended maximum limit. Methods of exposure 12 assessment are described in chapter 6. 13 Analytical requirements of the JECFA and the JMPR for food additives, pesticides, 14 veterinary drug residues, contaminants and substances consumed in large amounts are given 15 in sections 3.6, 3.7, 3.8, 3.9 and 3.10, respectively. 16 17 3.2 Considerations for laboratories 18 3.2.1 Guidance from Codex Alimentarius Commission (CAC/GL 27 - 1997) 19 The CAC has provided guidance for laboratories involved in the import/export testing of 20 foods (FAO/WHO, 1997a). This includes the recommendations that such laboratories should: 21 22 • use internal quality control procedures that comply with the Harmonized Guidelines for 23 Internal Quality Control in Analytical Chemistry Laboratories (Thompson and Wood, 24 1995); 25 • participate in proficiency testing schemes designed and conducted in accordance with the 26 International Harmonized Protocol for Proficiency Testing of (Chemical) Analytical 27 Laboratories (Thompson and Wood, 1993); 28 • become accredited according to International Organization for Standardization 29 (ISO)/International Electrotechnical Commission (IEC) Guide 25 on general requirements 30 for the competence of calibration and testing laboratories and its subsequent replacement 31 (ISO/IEC-17025, 2005); and 32 • whenever available, use methods that have been validated according to the principles laid 33 down by the CAC. 34 35 3.2.2 Best practices in analytical measurement 36 The OECD established Principles of Good Laboratory Practice (GLP) in 1978 under the 37 Special Programme on the Control of Chemicals. The Principles were formally recommended 38 for use in Member countries by the OECD Council in 1981. They were developed primarily 39 in relation to the performance of toxicology (safety) studies, but the principles related to 40 facilities, personnel, methods, data recording and quality assurance are equally applicable to 41 analytical methods. The Principles were updated in 1997 and are freely available in electronic 42 form on the OECD website 43 (http://www.oecd.org/document/63/0,3343,en_2649_201185_2346175_1_1_1_1,00.html). 44 Six general principles have been elaborated for the best practice of analytical 45 measurements (EURACHEM, 1998) and should be considered in the evaluation of data for 46 use in risk characterization and assessment (adapted as regards the feasibility of principle no. 47 6): 48 49 1. Analytical measurements should be made to satisfy a defined objective, and the method 50 used should be fit for that purpose.

2 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 2. Organizations making analytical measurements should have well defined quality control 2 and quality assurance procedures. Where possible, these should be based on 3 internationally recognized standards. 4 3. Analytical measurements should be made in appropriate facilities with methods and 5 equipment that have been tested to ensure they are fit for the analyses performed. 6 4. Staff making analytical measurements should have appropriately demonstrated 7 qualifications and competence to undertake the task. 8 5. There should be a regular independent assessment of the technical performance of the 9 analysing laboratory. 10 6. Analytical measurements made in one location should to the extent possible be consistent 11 with those made elsewhere, through multilaboratory testing or other means of validation 12 and quality assurance. 13 14 The safety assessment and evaluation of food additives and residues of veterinary 15 drugs and pesticides in foods requires adequate analytical data. Codex has elaborated 16 guidelines for analytical methods that are primarily intended for the verification of provisions 17 in Codex standards (FAO/WHO, 2005). These guidelines address the general principles for 18 the best practices of analytical measurements listed above. 19 20 3.3 Considerations for selection of validated analytical methods 21 3.3.1 General considerations 22 Various types of methods are available to food safety agencies and testing programmes to 23 conduct analyses that may be consistent with their requirements. Decisions on the use of a 24 specific analytical method should be based on the intended objectives of the regulatory 25 programme and the analytical performance requirements. Methods that are suitable for 26 determining compliance with MRLs and MLs are those that have been validated for the 27 analysis of specific analytes, such as pesticide or veterinary drug residues or other 28 contaminants, in specific types of samples (matrices). These methods provide analytical 29 results for either quantification or confirmation that are appropriate to support regulatory 30 action without the need for additional analyses. In some cases, these methods may be 31 considered reference methods, but reference methods frequently are not those selected for 32 routine use. 33 34 3.3.2 Multilaboratory method trials and collaborative studies 35 Relatively few of the analytical methods used in regulatory control programmes have 36 successfully completed a multilaboratory study, which provides information on method 37 performance in the hands of different analysts in different laboratories. Collaboratively 38 studied methods are subjected to a properly designed interlaboratory study with analysts in 39 independent laboratories, so that different sources of reagents, chromatographic media and 40 equipment are used by the participants. Collaborative studies of qualitative methods currently 41 require a minimum of 10 participating laboratories. Quantitative methods studied 42 collaboratively according to the revised harmonized protocol adopted in 1995 have been 43 evaluated in a minimum of eight laboratories, unless highly complex equipment or other 44 unusual requirements were identified (in such cases, a minimum of five participating 45 laboratories is required). For methods that have been subject to multilaboratory trials, the 46 performance characteristics, such as recovery and precision, are defined through the results 47 obtained during the study. Multilaboratory trials that do not meet the criteria for collaborative 48 studies may still provide useful information on the expected performance of the method 49 tested, and such information can be important in assessing the validity of concentration data 50 used in exposure assessments.

3 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 Multilaboratory and collaborative studies of methods usually do not encompass all 2 possible combinations of the analyte and the sample material to which the method may 3 subsequently be applied. These methods may be extended to related analytes and sample 4 materials not included in the original multilaboratory study by completing additional properly 5 designed within-laboratory studies. Whenever possible, analytical results obtained using 6 methods that have not been validated by interlaboratory study should be correlated and 7 compared with results obtained using a method that has been validated through a 8 collaborative or multilaboratory study. The comparison should be based on a statistically 9 acceptable study design using portions of the same (homogeneous) samples. The data from 10 such studies should be independently reviewed by a qualified third party (such as a quality 11 assurance unit, a peer group of regulatory scientists or auditors of national accreditation body) 12 to determine the comparability of method performance. 13 14 3.3.3 Routine regulatory methods 15 Most laboratories conducting analyses for regulatory purposes rely on analytical methods that 16 have not been subjected to an interlaboratory study. Factors that have contributed to this 17 situation include a requirement for specialized expertise or equipment, cost of such studies, 18 lack of suitable collaborating laboratories, analyte and/or sample instability and rapidly 19 changing technologies. Historically, the equivalency of analytical results was assured by the 20 use of standardized methods with performance characteristics based on collaborative studies. 21 At the present, it is the responsibility of the individual accredited laboratory to demonstrate 22 that the methods used and the analytical results produced meet performance criteria 23 established in consultation with a client. 24 Typical requirements would include that the methods are capable of detecting the 25 compounds included in the regulatory programme in the target samples with analytical 26 recovery and precision that meets the customer needs. In addition, the methods should be 27 used within an established laboratory quality assurance system that is consistent with the 28 principles on internal quality control (Thompson and Wood, 1995). When methods that have 29 not been subjected to a multilaboratory performance trial are used in a regulatory programme, 30 the quality control and quality assurance procedures applied with these methods require 31 careful definition, implementation and monitoring. Unlike methods established through 32 multilaboratory trials, the performance characteristics of a method validated within a single 33 laboratory are defined by the data generated by analysts within that laboratory. The ongoing 34 performance of the method and analysts must then be monitored through the quality system 35 in place in the laboratory. 36 37 3.3.4 Validation of methods in a single laboratory (the criteria approach) 38 The “criteria approach” to method validation is now defined in general terms in the Codex 39 Procedural Manual (FAO/WHO, 2005). It describes a process by which data may be 40 generated to establish the performance characteristics of an analytical method within a single 41 laboratory, with a recognized quality control system, and to demonstrate the “fitness for 42 purpose” of the method for routine analysis of samples. This approach was developed for use 43 in laboratories operating within the guidelines of CAC/GL 27-1997 and recognizes that 44 reliance solely on multilaboratory trials for method validation may no longer be practical. 45 Single laboratory validation has become increasingly important in recent years, 46 especially in the analysis of pesticide residues, for a number of reasons: 47 48 • Technological advancement is much faster, so that a method becomes out of date almost 49 as soon as it has been tested collaboratively.

4 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 • Laboratories have different equipment; although they may all have sophisticated 2 equipment such as gas chromatography (GC)–mass spectrometry (MS) and liquid 3 chromatography (LC)–tandem mass spectrometry (MS/MS), they may be fitted with 4 different types of injectors and the detectors may exhibit differences in sensitivity and 5 selectivity. 6 • In practice, only a small range of commodity/pesticide/residue level combinations can be 7 subject to multilaboratory method trials. 8 • The cost of multilaboratory method trials has increased because most contract 9 laboratories now require payment to participate. 10 11 On advice from the 10th Meeting of the CCRVDF (FAO/WHO, 1997c), the FAO 12 convened an expert consultation on “methods validation for food control purposes” in 1997 13 (FAO/IAEA, 1998), which recommended that “all methods used for determining compliance 14 with international or other standards which have not been subjected to a full collaborative 15 study should be subject to a form of independent review, which may include any of the 16 following options: a) a multi-laboratory validation study involving a smaller number of 17 laboratories, b) second laboratory verification, in a laboratory operating under Good 18 Laboratory Practices (GLP) or c) validation in a laboratory which has been recognized under 19 ISO/IEC Guide 17025 (2005) or equivalent.” The Consultation recommended the following 20 principles that should be applied in cases “where collaborative or other inter-laboratory 21 studies are impractical or impossible to carry out”: 22 23 • Laboratories carrying out the validation studies operate under a suitable quality system 24 based upon internationally recognized principles (Thompson and Wood, 1995). 25 • Laboratories have in operation a third-party review of the whole validation process (e.g. 26 GLP registration, accreditation according to ISO/IEC 17025 or equivalent, or peer 27 review). 28 • Analytical methods are assessed in respect to the Codex general criteria for selection of 29 methods of analysis (FAO/WHO, 2003), with emphasis on the assessment of the limit of 30 quantification rather than the limit of detection. 31 • The validation work should be carefully documented in an expert validation report that 32 states unambiguously the purposes (matrices and analyte concentrations) for which the 33 method has been found to perform in a satisfactory manner. 34 • Evidence of transferability should be provided for all methods intended for Codex use for 35 food control purposes. 36 37 Subsequently, requirements for “single laboratory method validation” were 38 considered by a working party of the International Union of Pure and Applied Chemistry 39 (IUPAC), resulting in the publication of a general guidance document (Thompson et al., 40 2002). An expert consultation jointly organized by IUPAC, AOAC International, the 41 International Atomic Energy Agency (IAEA) and FAO has provided specific guidance for the 42 validation of methods intended for the determination of trace residues of pesticides and 43 veterinary drugs in foods (Alder et al., 2000; Fajgelj and Ambrus, 2000). The contents of the 44 IUPAC guidance document have been applied generally by the Codex Committee on 45 Methods of Analysis and Sampling (CCMAS) for inclusion in the Codex framework for 46 method requirements, while the consultation report has been used by experts from both the 47 CCPR and the CCRVDF in revising requirements for the validation of recommended 48 analytical methods for pesticide residues and veterinary drug residues.

5 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 Based on recommendations from the CCMAS, the following conditions have been 2 included in the Codex Procedural Manual (FAO/WHO, 2005), to define where single- 3 laboratory validated methods can be used: 4 5 a) No interlaboratory validated method is appropriate. 6 b) The single-laboratory validated methods must fulfil the following criteria: 7 i) the method is validated according to an internationally recognized protocol (for 8 example, the IUPAC protocol referenced above); 9 ii) the use of the method is subject to a quality assurance system under accreditation; 10 iii) when available, external support is given by systematic participation in 11 proficiency schemes, by calibration using reference materials and by comparison 12 of results with those obtained using other methods. 13 14 3.3.5 Method validation process 15 Method validation is a process using a defined set of experiments to establish the 16 performance criteria that should be achieved by an analyst using the method and that provides 17 both users of the method and their customers a means to assess the reliability of results 18 obtained. The necessary experiments must be conducted to establish the various performance 19 criteria defined in the Procedural Manual and in other relevant Codex documents, including 20 analytical recovery, precision at various concentrations within the defined analytical range of 21 the method and detection capabilities related to selectivity and limits of detection and 22 quantification. 23 24 3.4 Development of analytical methods 25 3.4.1 Defining the analytical measurement requirements (fitness for purpose) 26 Developing an analytical method requires suitably experienced analysts, laboratory space, 27 equipment and financial support. Analytical methods used to determine compliance with 28 MRLs or other regulatory standards or specifications should be effective and practical. The 29 results should provide the qualitative and quantitative information required to demonstrate 30 compliance with the regulatory standard. Applications may include: 31 32 • analysis of randomly selected survey samples in a national programme to determine 33 compliance with established standards; 34 • analysis of targeted samples where there is reason to suspect non-compliance with these 35 standards; 36 • analysis of samples to meet a commercial requirement; or 37 • analyses used to estimate consumer exposure to residues or contaminants through food. 38 39 3.4.2 Analytical performance characteristics 40 Methods that provide quantitative results must perform with good precision and accuracy 41 within an analytical range that covers the MRL or other regulatory limit. Methods applied in 42 studies to assess daily intake of a selected residue or contaminant may be required to 43 accurately measure concentrations orders of magnitude below the MRL. For such 44 applications, the limit of quantification (LOQ) and linearity of response over an extended 45 analytical range become primary considerations. A key element of the work undertaken 46 during development of a regulatory method is ruggedness testing (Youden and Steiner, 1975) 47 in order to establish critical points in the method where a minor variation in procedure or a 48 change in source of material or reagent could profoundly affect the test result. 49 Analytical methods may be required in regulatory control programmes for the 50 detection of residues of substances for which the toxicological data available do not allow an

6 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 ADI or MRL to be established; the presence in food of the drug chloramphenicol would be an 2 example in this category. For such substances, determination of the lowest detectable 3 concentrations and confirmation of identity are the primary focuses for method validation. 4 Performance characteristics related to quantification may be less critical for such substances, 5 unless the results are to be used for exposure assessment. 6 Analytical performance characteristics are defined in the Procedural Manual, and 7 additional guidance is provided in other Codex texts from CCMAS, CCPR and CCRVDF. 8 9 3.5 General considerations for use of analytical methods in a regulatory 10 programme for foods 11 Analytical data are collected for a variety of purposes, including: 12 13 • national legal requirements to ensure the quality and safety of foods that are produced 14 domestically or imported or exported; 15 • monitoring to determine compliance with existing standards; 16 • surveillance, particularly for exposure assessment or to gather data for proposed standards; 17 and 18 • research for product development, including the development of specifications. 19 20 These purposes can have different analytical requirements, particularly in respect of 21 performance characteristics. 22 Uncertainties in analytical measurements, particularly for exposure assessment, can 23 contribute to the uncertainties in the safety and risk assessments. The fitness for purpose of 24 the analytical data for use in the safety and risk assessments should be determined on a case- 25 by-case basis, and any analytical or sampling uncertainties should be communicated as part 26 of the evaluation. 27 Quality control and quality assurance principles are essential components of chemical 28 analysis. They provide the basis for ensuring optimum method performance for all methods, 29 regardless of method attributes. Quality control monitors those factors associated with the 30 analysis of a sample by an analyst, whereas quality assurance provides the oversight by 31 independent reviewers to ensure that the analytical programme is performing in an acceptable 32 manner. Quality control and quality assurance programmes are invaluable in supporting 33 decision-making for risk assessment managers and enforcement agencies, improving the 34 reliability of analytical results and providing quality data for residue control programmes. 35 The establishment of quality measures consistent with the principles published by IUPAC is 36 important for regulatory control laboratories (Thompson and Wood, 1995). 37 38 3.6 Food additives—specifications 39 3.6.1 General considerations 40 Specifications of identity and purity are necessary products of JECFA safety evaluations for 41 food additives. Evaluations of food additives by JECFA depend on studies performed with a 42 chemical substance or product of defined identity, purity and physical form. The ADI is valid 43 only for products that do not differ significantly in identity and quality profile from the 44 material used to generate the data used in the evaluation. [For an overview of the purpose, 45 function and format of JECFA food additive specifications, and the interaction of JECFA and 46 Codex, see the Introduction, Combined Compendium of Food Additive Specifications, 47 JECFA Monographs 1 (FAO, 2005/2006).] 48 The specifications of identity and purity established by JECFA are intended to ensure 49 that the Committee’s safety evaluations apply, with a high degree of confidence, to all 50 products manufactured to comply with those specifications. The first Joint FAO/WHO

7 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 Conference on Food Additives (FAO/WHO, 1956) was asked to formulate general principles 2 governing the use of food additives and to recommend suitable methods for the chemical, 3 physical, pharmacological, toxicological and other properties of individual food additives. 4 The first two meetings of the Joint Expert Committee prepared reports on “General 5 Principles Governing the Use of Food Additives” (FAO/WHO, 1957) and “Procedures for the 6 Testing of Intentional Food Additives to Establish Their Safety for Use” (FAO/WHO, 1958), 7 and recommended the need for specifications; since then, specifications have been an 8 important part of JECFA evaluations for food additives. JECFA specifications have three 9 purposes: 10 11 1) to identify the substance that has been tested biologically; 12 2) to ensure that the substance is of the quality required for safe use in food; and 13 3) to require good manufacturing practice and maintain the quality of additives on the 14 market. 15 16 Since 1956, the meetings of JECFA have designated specifications as either full or 17 tentative. Until the twenty-third meeting of JECFA, specifications were designated as 18 tentative either because the chemistry data were inadequate or because a temporary ADI was 19 assigned to the additive. At and since the twenty-third meeting of the JECFA, a tentative 20 specification has been assigned only when the data were inadequate for preparing full 21 specifications. 22 A food additive may be a single chemical substance, a manufactured chemical 23 mixture or a natural product. Complete information on chemical composition—including 24 description, methods of manufacture and raw materials, and analyses for impurities—is 25 equally important for each type of additive. However, implementation of the requirement for 26 chemical composition data may vary depending on the type of substance. 27 For additives that are single-chemical substances, it is virtually impossible to remove 28 all impurities arising from their commercial production; therefore, analyses are generally 29 performed on the major component and predicted impurities, especially those with potential 30 toxicity. 31 For commercially manufactured complex mixtures, such as mono- and diglycerides, 32 information is needed on the range of substances produced, with emphasis on descriptions of 33 manufacturing processes, supported by analytical data on the components of the different 34 commercial products. 35 Natural products present particularly difficult problems due to their biological 36 variability and because the chemical constituents are too numerous for regular analytical 37 determinations. For additives derived from natural products, it is vital that the sources and 38 methods of manufacture are defined precisely. Chemical composition data should include 39 analyses for general chemical characteristics, such as proximate analyses for protein, fat, 40 moisture, carbohydrate and mineral content. Analyses should be undertaken for specific toxic 41 impurities carried over from raw materials or chemicals used in the manufacture of the 42 product. Further information necessary for the evaluation of “novel foods”, which are usually 43 substances derived from natural products, is provided in section 3.10. 44 JECFA policy has been to prepare specifications whenever constituents of the 45 substance added to food had the potential to be present in the finished food. Initially, 46 specifications were prepared only for intentional food additives—that is, those that are added 47 directly to a food to accomplish a technical effect (e.g. a preservative or colour). The 48 fourteenth JECFA (FAO/WHO, 1971) prepared specifications for extraction solvents because 49 although these “processing aids” are largely removed from food, evaluation of their safety in 50 use depends on their identity and purity. Since then, specifications have been prepared for all

8 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 processing aids (e.g. antifoaming or clarifying agents, enzyme preparations, filtering aids, 2 packing gases, release agents, and others) used in conjunction with food manufacture. 3 The twenty-seventh JECFA (FAO/WHO, 1983a) decided that chemical reagents used 4 in the preparation of food additives or processing aids (e.g. glutaraldehyde in the preparation 5 of immobilized enzyme preparations or acetic anhydride in the manufacture of modified 6 starches) do not usually need specifications. Carryover of these reagents or their 7 contaminants into food may be controlled by the specifications for purity of the specific 8 additive or processing aid. 9 Many food additive specifications have identical analytical methods and/or test 10 procedures. To avoid repetition in each individual specification, these methods and test 11 procedures were assembled in a volume entitled “Guide to Specifications” (FAO, 1978), and 12 subsequent specifications referred to that volume when appropriate. The volume was revised 13 and updated in 1983 (FAO, 1983) and 1991 (FAO, 1991a). In 2006, the information 14 contained in the volume was completely revised and rewritten and has been published as 15 Volume 4 of the Combined Compendium of Food Additive Specifications, FAO JECFA 16 Monographs 1 (FAO, 2005/2006). 17 Food additives may be marketed as formulated preparations, such as a mixture of a 18 main ingredient with a solvent vehicle and emulsifier. Specifications refer to each ingredient 19 in the formulated preparation as individual commercially manufactured food additive 20 substances. Mixtures should not be formulated in such a way that the absorption or 21 metabolism of any ingredient is altered so that the biological data, derived using the 22 individual component, are invalidated (FAO/WHO, 1966, 1972). Added substances such as 23 anticaking agents, antioxidants and stabilizers may influence the results of analytical tests 24 given in specifications. Therefore, in its nineteenth report, JECFA recommended that 25 manufacturers of food additives should indicate the presence of such added substances 26 (FAO/WHO, 1975a). 27 28 3.6.2 Formulation of specifications and information requirements 29 The formulation of satisfactory specifications requires detailed information to be made 30 available to JECFA on the method of manufacture of the additive, including information on 31 raw materials and on its chemical characterization. The Committee requires such information 32 to be provided as part of the total data package whenever an additive is submitted for risk 33 assessment; all such information is regarded as suitable for being made available publicly 34 unless requested otherwise and agreed by the JECFA Secretariat. Those submitting data for a 35 JECFA evaluation are advised to consult existing specifications for further guidance, which is 36 available in the Combined Compendium of Food Additive Specifications (FAO, 2005/2006), 37 where the individual criteria used in the elaboration of JECFA specifications are described. 38 The same criteria are used for most additives, but because of their particular characteristics, 39 separate criteria have been developed for enzyme preparations and also for flavouring 40 substances. 41 Specifications may be revised where there is new information available on methods of 42 manufacture or on the characteristics of the product, or where changes or revisions in 43 analytical methods are needed. Such specification changes may trigger a review of the safety 44 evaluation; conversely, a review of the specifications may be needed if the safety is re- 45 evaluated. 46 Although all the individual criteria in specifications monographs must be met, 47 additives are mainly defined by a combination of 1) a description of their manufacture, 2) a 48 minimum requirement for the content of the principal functional component(s) of the additive 49 and 3) maximum limits for undesirable impurities. The relative importance of these criteria 50 depends on the nature of the additive; for example, additives composed largely of single

9 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 components are mainly defined in terms of their chemical purity, whereas the definition of 2 more complex materials relies more on a description of the raw materials and the method of 3 manufacture. 4 5 3.6.3 Stability and fate of additives in food 6 Specifications are intended to apply to the additive as marketed and supplied for food use. In 7 considering whether specifications apply to food additive quality as manufactured or as added 8 to food, JECFA has decided to prepare specifications to cover the normal shelf-life of the 9 additive. Limits are set for decomposition products that may form during normal storage. 10 Manufacturers and users of food additives should ensure good packaging and storage 11 conditions and use good handling practices to minimize deleterious changes in quality and 12 purity (FAO/WHO, 1975a). Information on changes in the composition of food additives 13 during storage should be submitted for evaluation by the Committee. 14 Processing aids are substances that come into contact with food during processing and 15 may unintentionally become part of food because of their incomplete removal. Committees 16 have evaluated a number of processing aids, such as extraction solvents and enzyme 17 preparations, for their safety in use. When evaluating a processing aid, information should be 18 provided on its use and either analytical data on or a computed estimate of the amount of the 19 processing aid carried over into food. Particular attention should focus on any component of 20 the processing aid that may have the potential for biological effects, such as ethylenimine 21 leaching from polyethylenimine, an immobilizing agent used in the preparation of 22 immobilized enzyme preparations. 23 Certain food additives perform their functional effect by reaction with undesirable 24 food constituents (e.g. antioxidants react with oxygen in food, and ethylenediaminetetraacetic 25 acid [EDTA] reacts with trace metals) or by reactions that modify food constituents (e.g. 26 flour improvers). Food additives may also degrade under certain conditions of food 27 processing, even though such degradation is detrimental to their functional effect. For 28 example, the sweetener aspartame is transformed to a diketo-piperazine derivative at rates 29 that vary with the acidity and the temperature of the food. For such additives, the Committee 30 has evaluated analyses for additive reaction products in food as consumed and biological 31 testing data on either specific reaction products or samples of food containing the reaction 32 products as consumers would ingest them. 33 In order to ensure that test data are relevant to the way the additive is used in food, the 34 Committee requires information on potential reactivity to be provided as part of submissions 35 for the safety evaluation of all intentional food additives (FAO/WHO, 1981a). Four types of 36 data related to reactivity are required: 37 38 1) the general chemical reactivity of the additive; 39 2) stability of the additive during storage and reactions in model systems; 40 3) reactions of the additive in actual food systems; and 41 4) the metabolism of the additive in living organisms. 42 43 These data are important for relating toxicological data to the actual use of the additive in 44 food. 45 46 3.6.4 Methods of analysis 47 Information submitted to JECFA on the identity and purity of food additives should always 48 include details of the analytical methods that can be used to verify the information. 49 Information on the potential compositional variability of the substance should also be given, 50 together with details of any sampling protocols used to assess this. Insufficient information

10 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 on analytical methodology is one reason why JECFA may be unable to elaborate suitable 2 specifications, or why it may decide that it is only be able to assign a “tentative specification” 3 pending receipt of the further information required. 4 JECFA specifications incorporate guidance on the analytical techniques that should be 5 used to verify the information. Wherever possible, this should be done by reference to 6 Volume 4 of the Combined Compendium of Food Additive Specifications, FAO JECFA 7 Monographs 1 (FAO, 2005/2006). If this is not possible, details of the test procedures are set 8 out in the individual specifications monographs. 9 Because JECFA specifications are elaborated for worldwide use, the Committee 10 prefers to quote methods that require the use of apparatus and equipment that are available in 11 most laboratories, provided that such methods give results appropriate to the specified criteria. 12 Methods involving more recently developed techniques or equipment will therefore not 13 normally be quoted until such techniques are accepted internationally and are generally 14 available at reasonable cost. However, reference to specific methods of analysis should not be 15 taken as precluding the use of other methods, provided that these are validated as giving 16 results of at least equivalent accuracy and specificity to those quoted. 17 Changes to analytical methods are reviewed from time to time as part of JECFA’s 18 ongoing work. Changes may also be considered when substances are evaluated for the first 19 time or when new information becomes available on substances that have been previously 20 considered. Changes in analytical methodology may also trigger further consideration of 21 specifications—for example, where these changes reveal the possible presence of hitherto 22 unsuspected contaminants. 23 24 3.7 Pesticides—specifications 25 3.7.1 General considerations 26 When an active ingredient is evaluated by JMPR for the first time or during a periodic review, 27 it is characterized by its systematic and common names, its empirical and structural formulae 28 and its Chemical Abstracts Service (CAS) and CIPAC (Collaborative International Pesticides 29 Analytical Council) numbers. Details of the chemical and physical characteristics of the 30 active chemical are required, together with information on the proportions of different 31 components when the compound is a mixture (e.g. of stereoisomers). The need for accurate 32 specifications for pesticides was stressed by the 1968 JMPR (FAO/WHO, 1969) during its 33 deliberations on toxaphene and on technical grades of benzene hexachloride. Because of the 34 unknown or variable composition of these compounds, JMPR was unable to relate the 35 existing toxicological data to the products in agricultural use, and ADIs could not be 36 allocated. Attention was also drawn to the possibility of variability between nominally similar 37 chemicals produced by different manufacturers. 38 An additional issue for pesticides is the need for chemical characterization of, and 39 analytical methods for, active metabolites or breakdown products that may be present in or on 40 the food after treatment according to GAP. 41 In addition to characterization of the active substance(s) (see section 3.1), the “levels 42 of impurities that, according to current knowledge, are considered to be toxicologically 43 significant … must appear in the specifications” (WHO, 1967). Additional considerations 44 regarding the significance of impurities in pesticide preparations are given in Ambrus et al. 45 (2003). 46 The technical-grade pesticide is characterized by its minimum purity, isomer 47 composition and the limits for content of impurities that might impact on the hazard 48 assessment. The possible influence of known or unknown impurities on the toxicity of 49 technical-grade chemicals and of residues resulting from their use was discussed by the 1974 50 JMPR (FAO/WHO, 1975b). This JMPR noted that toxicity studies are generally performed

11 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 on technical-grade materials produced by commercial-scale processes and that the resulting 2 toxicological data normally, therefore, take into account the presence of impurities (provided 3 that the manufacturing process remains the same). However, it noted the problems 4 encountered with trace amounts of biologically active materials (e.g. 2,3,7,8-tetrachloro- 5 dibenzo-p-dioxin in 2,4,5-trichlorophenoxyacetic acid). It further noted that “specifications 6 such as those issued by FAO/WHO are seldom designed to take note of trace-level impurities, 7 unless the importance of such impurities has already been revealed by biological studies”. 8 The 1977 JMPR (FAO/WHO, 1978) noted that data on the nature and level of 9 impurities, intermediates and by-products in technical pesticides were often available, but, 10 because such data could provide valuable information to competitors, they were normally 11 considered to be a “trade secret”. The Joint Meeting therefore agreed that such data would not 12 normally be published in the JMPR reports or monographs. 13 In considering the applicability of recommendations to pesticides from different 14 manufacturers, the 1978 JMPR (FAO/WHO, 1979) indicated that evaluations and 15 recommendations are valid only for the specific technical grade of pesticide examined. 16 Considerable care and knowledge of the detailed specifications are required to extrapolate 17 evaluations and recommendations to products of different quality or composition. 18 Subsequent Joint Meetings (FAO/WHO, 1980, 1981b, 1985) stressed the importance 19 of information on the presence of impurities in technical pesticide products (e.g. the presence 20 of hexachlorobenzene in various pesticides, impurities in phenthoate, and dimethylhydrazine 21 in maleic hydrazide). The need for technical-grade pesticides to meet FAO specifications has 22 also been stressed. It was noted by the 1984 JMPR that occasionally data have been rejected 23 because the test material failed to comply with these specifications (FAO/WHO, 1985). 24 The 1987 JMPR (FAO/WHO, 1987) noted that ADIs based on studies using 25 compounds of specific purity can be relevant to products of different origin or purity, but that 26 there are examples where changes in the amount or type of impurity in the technical material 27 can markedly influence the toxicity of a compound. 28 Toxicity tests should normally be performed on the technical grade of the pesticide. 29 An exception to this is acute toxicity, for which both formulations and technical materials 30 must be tested in order to assess the risk to the applicator. However, the percentage of active 31 ingredient and impurities in the technical-grade material may vary among production batches 32 and may differ at various stages of product development. Furthermore, some toxicity testing 33 is likely to be performed with the product in the early stages of development. Preliminary 34 studies may be performed on batches of material produced within the laboratory in order to 35 assess the potential acute risks to individuals who will be involved in the development of the 36 material. Subsequent studies may be performed on material produced in a pilot plant, while 37 other toxicity studies may be performed on the marketed product, which will be produced in a 38 full-scale manufacturing plant. At each step in this sequence, there is a potential for change in 39 the percentage of the active ingredient in the “technical-grade” material and a potential for 40 change in the quantity and identity of the impurities that constitute the remainder of the 41 “technical-grade” product. It is therefore essential that detailed specifications should be 42 provided for the test material utilized in each study. 43 The active ingredient of the pesticide may exist in two or more forms (e.g. as a 44 diastereoisomeric mixture); this is normally the case for synthetic pyrethroids. Under such 45 conditions, the ratio of isomers in the test material must be clearly specified, because it has 46 been documented that different isomers frequently have different toxicological activities 47 (Green, 1978; FAO/WHO, 1980). For example, an ADI for permethrin (40% cis : 60% trans) 48 was allocated in 1982 (FAO/WHO, 1983b), whereas an ADI for permethrin (25% cis : 75% 49 trans) was not allocated until 1987 (FAO/WHO, 1987).

12 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 Data on the stability of the test material are also of importance. The percentage of the 2 active material will decrease and that of breakdown products will increase with time if a test 3 compound is unstable under the conditions of storage. This will be of major importance in the 4 evaluation of the results of studies where a single batch of technical material is utilized for a 5 long-term study or a multigeneration study. Variation in the amount of degradation occurring 6 in different batches (i.e. batches of different post-manufacturing age) may complicate the 7 interpretation of a study. Finally, reaction of the active compound with components of the test 8 diet will result in a decreased concentration of active compound and may result in the 9 production of toxic reaction products in the diet. In cases where the percentage of active 10 parent compound decreases and/or the breakdown products are more toxic than the parent 11 compound, NOAELs derived from the toxicity tests may not be representative of the product 12 as used. 13 To date, JMPR has evaluated only the active ingredients of pesticide formulations. 14 The toxicity of essentially inert ingredients—such as solvents, emulsifiers and 15 preservatives—that may occur as residues in food has not been considered. 16 In 1999, FAO, in cooperation with WHO, introduced a revised procedure for 17 evaluating data to establish specifications for pesticides (FAO, 1999). Specifications would 18 include all relevant impurities, defined as “those by-products of the manufacture or storage of 19 a pesticide which, when compared with the active ingredient, are toxicologically significant 20 to health or the environment, are phytotoxic to treated plants, cause taint in food crops, affect 21 the stability of the pesticide, or cause any other adverse effect.” The long-term aim was for 22 FAO/WHO specifications for technical material to be developed before the establishment of 23 an ADI or an ARfD. 24 Under the revised procedure of 1999, FAO/WHO specifications apply only to the 25 products of manufacturers whose technical materials (defined as active ingredients isolated— 26 as far as is practicable—from starting materials, solvents, etc.) have been evaluated as 27 satisfactory by the Joint FAO/WHO Meeting on Pesticide Specifications (JMPS) (FAO/WHO, 28 2006). 29 Data required to support the development of pesticide specifications by JMPS include 30 identity of active ingredient, physicochemical properties, route of manufacture, minimum 31 active ingredient content, maximum limits for impurities present above 1 g/kg, maximum 32 limits for impurities proposed as relevant at <1 g/kg, identity and nominal content of 33 compounds intentionally added to the technical material, toxicological and ecotoxicological 34 summaries, properties of formulations, and methods for the analysis and testing of technical 35 material and formulations (includes methods for relevant impurities). 36 In 2005, JMPR reiterated the previous conclusions that specifications for the technical 37 material should be developed for a pesticide before it is evaluated within the periodic review 38 programme of the CCPR and for new pesticides, but that this should not delay evaluation of 39 pesticides by JMPR (FAO, 2005a). 40 41 3.7.2 Identity and purity 42 Guidance on the development and use of specifications for pesticides evaluated by JMPR was 43 elaborated in 2002 by the first meeting of JMPS (FAO/WHO, 2002). 44 For the purposes of the characterization: 45 46 • A detailed specification of the test material used in each individual study must be 47 provided. 48 • Where isomeric mixtures exist, the ratio of isomers in the test material must be clearly 49 specified. 50

13 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 For purity considerations: 2 3 • The percentage of the active ingredient in any technical material used in a toxicity test or 4 proposed for marketing must be specified. 5 • Percentages of all identifiable impurities should be specified. 6 • Data on manufacturing processes may be required to permit determination of potential 7 impurities; however, because of confidentiality, such data will not be published in JMPR 8 monographs. 9 10 3.7.3 Stability 11 • Stability of the test material during storage and in the diet must be adequately investigated 12 and reported. 13 • Where instability in diets is observed, possible reaction products and the nutritional 14 quality of the diet should be investigated. 15 16 3.7.4 Physical and chemical properties 17 Data submitted on physical and chemical properties of the pure active ingredient are 18 evaluated in order to recognize the influence of these properties on the behaviour of the 19 pesticide during and after its application on crops or animals. JMPR receives data on 20 pesticide physical appearance, solubility in water (including pH effects) and in organic 21 solvents, vapour pressure, dissociation constant, n-octanol–water partition coefficient, 22 hydrolysis and photolysis. 23 The volatility of the compound, its stability in water and sensitivity to irradiation with 24 ultraviolet light may considerably affect its disappearance after application. 25 Epimerization may sometimes be observed during hydrolysis studies. For example, 26 esfenvalerate (2S,αS) was epimerized to the 2S,αR isomer more quickly than it was 27 hydrolysed under experimental conditions (FAO, 2002a). The proportion of epimers may 28 influence the toxicity. 29 The solubility of the pesticide is of great importance, because the ability of the 30 compound to penetrate plant and animal tissues is dependent on its solubility in water and 31 organic materials. 32 JMPR (FAO, 1991b) chose the octanol–water partition coefficient (Pow) of a pesticide 33 as the physical property to represent solubility in fat. In general, the compound would be 34 designated fat soluble when log Pow exceeded 4, but not when log Pow was less than 3. 35 Subsequently, JMPR (FAO, 2005b) examined the available data and concluded that 36 partitioning in meat between fat and muscle is essentially independent of log Pow for 37 compounds with values greater than 3. In consequence, and when no evidence is available to 38 the contrary, the compound is designated fat soluble when log Pow exceeds 3, but not when 39 log Pow is less than 3. Although log Pow of an individual component of a residue is an initial 40 indicator, it is not the only or prime factor used to assess fat solubility. The distribution of the 41 residue (as described in the residue definition) between muscle and fat obtained from 42 livestock metabolism and feeding studies should be the prime indicator of fat solubility. 43 44 3.7.5 Analytical methods 45 Pesticides are very diverse chemical compounds with a wide range of physical and chemical 46 properties. Analytical chemists have devised methods for analysis of pesticide residues, 47 including their transformation products, in a wide range of situations. 48 Methods should be validated to provide the supporting information on accuracy, 49 selectivity and reliability of the data generated by the method. Hill and Reynolds (1999)

14 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 explained the practicalities and compromises in validating analytical methods for pesticide 2 residues in food and animal feeds. 3 Analytical methods should be suitable for the required purpose, which usually falls 4 into one of three areas of residue analysis: 5 6 • data generation for registration; 7 • MRL enforcement and surveillance; and 8 • total diet studies. 9 10 JMPR evaluates the analytical methods used for generation of residue data to check 11 that the methods are suitable for the relevant analytes and sample types. The methods should 12 be supported by adequate validation data, especially on analytical recoveries, LOQ and 13 selectivity. 14 JMPR also reports information on methods that are suitable for MRL enforcement and 15 whether particular compounds are suitable for analysis by multiresidue methods. 16 Most analytical methods for residues of simple organic compounds in a food 17 commodity matrix consist of three main steps: 1) extraction, 2) cleanup and 3) determination 18 or measurement usually involving GC or LC. However, some analytes require other 19 approaches. For example, a chemical reaction may be needed to release an analyte from the 20 residue, or a derivative of the analyte may have to be prepared for the chromatography step 21 (e.g. the analytical method for residues of dithiocarbamates is non-specific and measures 22 carbon disulfide released by treatment with acid). 23 JMPR evaluates methods used for generating preregistration residue data that are 24 needed for analysis of samples from: 25 26 • supervised residue trials; 27 • food processing studies; 28 • livestock feeding studies and direct animal treatment; and 29 • sample storage stability studies. 30 31 Analytes include compounds to be specified in the residue definitions (i.e. the MRL 32 enforcement residue definition and the dietary intake risk residue definition). This substance 33 would, in the majority of cases, be the parent compound, with inclusion of one or more 34 metabolites or other transformation products when appropriate, based on the metabolism of 35 the pesticide in plants and animals. 36 The LOQ of the analytical method for residue trials would be typically 0.01–0.05 37 mg/kg. Lower LOQs may be needed in some circumstances. For example, dietary intake 38 calculations for a pesticide with a low ADI or ARfD might suggest that residues need to be 39 measured at levels less than 0.01 mg/kg, necessitating a method with a lower LOQ. Total diet 40 studies may need especially low LOQs for some analytes. 41 The FAO Panel of JMPR defines the LOQ of an analytical method for residues in 42 specified commodities as being the lowest level where satisfactory recoveries were achieved. 43 The LOQ is the smallest concentration of the analyte that can be quantified. It is commonly 44 defined as the minimum concentration of analyte in the test sample that can be determined 45 with acceptable precision (repeatability) and accuracy under the stated conditions of the test. 46 (FAO, 2002b). 47 Analytical recovery data support JMPR decisions on the acceptability or non- 48 acceptability of the associated residue data. Recoveries in the 70–120% range are considered 49 satisfactory. JMPR does not normally adjust or correct residue data using analytical recovery 50 data.

15 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 2 3.7.6 Analytical problems and challenges 3 Residue methods should normally be tested and validated on representative commodities 4 (chosen because of expected residue occurrence), such as: 5 6 • high moisture content plant material (e.g. lettuce, tomatoes); 7 • high oil and protein content (e.g. soybeans, peanuts, avocado); 8 • high starch or sugar content (e.g. cereal grains, potatoes); 9 • acidic commodities (e.g. citrus fruits); 10 • low-moisture feed materials (e.g. maize fodder); 11 • animal tissues (e.g. beef muscle, fat, liver, kidney); and 12 • milk and eggs. 13 14 Some matrices may cause particular problems (e.g. poor recoveries or interferences). 15 For example, onions, broccoli and cabbage release carbon disulfide from endogenous 16 precursors when treated with acid, which interferes with the measurement of dithiocarbamate 17 residues (FAO, 1993a). In another example, recoveries of approximately 50% were obtained 18 when racemic glufosinate was spiked into transgenic glufosinate-tolerant soybean plants, 19 because the transgenic plant material very rapidly metabolized the L-enantiomer, leaving 20 only the D-enantiomer for measurement (FAO, 1998b). 21 Analysts should be aware of interferences from the matrix that could apparently add 22 to the measured residue or cause losses during the procedure. For example, the 23 chromatographic response to indoxacarb residues was enhanced by the crop extract, 24 necessitating the preparation of standard solutions in crop extract (FAO, 2005c). 25 The analysis of ethylenethiourea residues in the presence of parent 26 ethylenebisdithiocarbamate (mancozeb) presents special problems that may not be covered by 27 normal validation testing. Mancozeb residues may be converted to ethylenethiourea under 28 some conditions during the analytical procedure (estimated conversion rates 0.22–8.5%). In 29 samples where mancozeb is present at concentrations up to 1 mg/kg, it is possible that 30 ethylenethiourea residues close to but above the LOQ (0.02 mg/kg) may have been produced 31 during the analytical procedure (FAO, 1993b). 32 The extraction efficiency for residues bound within the matrix cannot be tested by 33 spiking samples shortly before analysis, but bound 14C-labelled residues from metabolism 34 studies may be used to check extractability. Samples of plant and animal tissue from the 35 radiolabelled metabolism studies containing bound 14C residue levels may subsequently be 36 analysed by the routine residue method (or, at least, the extraction procedure of the routine 37 method) in order to define the extractability of the bound 14C residues. 38 The 1998 JMPR (FAO, 1998a) recommended that “Comparative extraction efficiency 39 studies including the frequently used extraction solvents, such as acetone/water, ethyl acetate 40 and acetonitrile/water should be carried out on samples from metabolism studies for the 41 compounds which are expected to be included in the residue definition(s)”. A IUPAC report 42 (Skidmore et al., 1998) stated that “The extraction procedures used in residue analytical 43 methods should be validated using samples from radiolabelled studies where the chemical has 44 been applied in a manner consistent with the label and Good Agricultural Practices.” 45 In analytical chemistry, the term “common moiety” means that structural portion of 46 different compounds that is the same and that tends to remain intact during chemical 47 reactions. A common moiety analytical method relies on this feature to measure the 48 concentration of a group of related compounds all together. Such a method may be useful 49 when a number of metabolites with the common moiety need to be included in the estimates 50 of dietary intake or when the composition of the residue is quite variable and the common

16 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 moiety is easier to measure than a specific component. An example of this is the analysis of 2 dithiocarbamate pesticides using acid-release carbon disulfide as the final analyte. 3 An analytical method used for testing the stability of residues during frozen storage 4 needs to be reproducible for the duration of the test (perhaps 2 years), and it should 5 distinguish the starting compound from degradation products. If analytical recoveries are too 6 variable, the variability will obscure conclusions about stability, and only large losses during 7 storage will be observable. 8 9 3.8 Veterinary drug residues 10 3.8.1 General considerations 11 The basic data requirements were established by the thirty-second meeting of JECFA 12 (FAO/WHO, 1988). The Committee must be assured that any veterinary drug it evaluates is 13 well characterized, with details of the chemical and physical properties of the drug and the 14 identity and concentrations of any major impurities. In addition, the manufacturing process 15 should be described and the consistency and quality of the final products demonstrated. This 16 information must be included in the dossier submitted for review by the Committee and is 17 used to define the substance used in the studies that lead to the establishment of the MRLVDs 18 (MRLs for a veterinary drug) and the ADI. 19 Veterinary drugs cover a broad range of chemical structures and usually undergo 20 metabolism after administration to an animal. Modes of administration include injection, 21 implantation, dermal application by spray or pour-on, and inclusion in feed or water, all of 22 which may have different rates of adsorption, with possible differences in the tissue 23 distribution and nature of the residues. The form and the distribution of the residues that 24 result from each authorized mode of application in each species must be determined, and the 25 depletion of the residues from edible tissues or animal-derived foods must be studied. A 26 marker residue must be identified, which is usually the form of the drug (parent compound or 27 metabolite) that is found at the highest concentration for the longest period in the target food. 28 The relationship of this marker residue to the total residue of the drug must be determined, 29 usually through treatment of experimental animals with an isotope-labelled form of the drug. 30 The tissue in which the highest residues are found is usually designated as a “target tissue” 31 for routine monitoring purposes. 32 Validation requirements for analytical methodology, whether intended for use in 33 pharmacokinetic and metabolism studies, residue depletion studies or in a regulatory control 34 programs for residues of veterinary drugs, are essentially the same. Performance 35 characteristics to be determined include linearity, specificity, accuracy, precision, limit of 36 detection (LOD), LOQ, practicability and applicability under normal laboratory conditions 37 and susceptibility to interference. Validation thus addresses all aspects of performance 38 characteristics of the analytical methods. Target values for method precision and recovery 39 have been established by the CCRVDF for the concentrations typically required to support 40 MRLVDs (FAO, 1993c). 41 42 3.8.2 Analytical methods 43 The first meeting of the Committee devoted exclusively to the evaluation of veterinary drugs 44 identified criteria to be applied in the assessment of analytical methods, which included 45 accuracy, precision, specificity, sensitivity, reproducibility, reliability and cost-effectiveness 46 (FAO/WHO, 1988). The Committee recognized that analytical methods are required to 47 “detect, quantify and positively identify residues of veterinary drugs; support toxicological, 48 drug metabolism, and pharmacokinetic studies; support residues studies of compounds to be 49 evaluated by the Committee; and satisfy the needs of public health agencies”. The initial 50 focus of JECFA was to ensure that methods used in the pharmacokinetic and residue

17 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 depletion studies evaluated by the Committee had been suitably described and appropriately 2 validated. The ninth meeting of the CCRVDF decided that no MRLVD should advance to 3 Step 8 of the process without a suitable method being identified to support the MRLVD. This 4 decision added emphasis to the role of JECFA in identifying analytical methods suitable for 5 regulatory use as part of their review (FAO/WHO, 1997). The eleventh meeting of the 6 CCRVDF determined that JECFA would have primary responsibility for review of methods 7 for compounds considered by the fiftieth and subsequent meetings of JECFA (FAO/WHO, 8 1999). The fifty-second meeting of JECFA approved a guidance document “JECFA 9 Requirements for Validation of Analytical Methods”, published with the residue monographs 10 from the fifty-eighth meeting of JECFA (FAO, 2002c), which updated the criteria established 11 at the thirty-second meeting of JECFA. 12 During JECFA review, the primary requirement for methods used in pharmacokinetic 13 and residue depletion studies is that the method has been shown to have performed reliably in 14 the hands of the analyst or analysts involved in that specific study. The dossier reviewed by 15 JECFA usually includes a complete validation report for the method, particularly if the 16 method has not been published in the peer-reviewed scientific literature. In the absence of 17 such a method, temporary MRLVDs may be established subject to provision of a validated 18 method to a subsequent meeting of JECFA. 19 For some compounds evaluated by JECFA, no residues are detected in one or more of 20 the four edible target tissues from any of the animals to which the drug has been administered 21 either at any time of sampling or at the time-point on the depletion curve at which MRLVDs 22 are established consistent with the ADI. In such cases, CCRVDF has requested that JECFA 23 establish MRLVDs for these tissues in which no residues have been detected, based on the 24 LOQ of the available residue control method, provided that such MRLVDs are consistent 25 with adequate health protection. 26 27 3.8.3 Analytical problems and challenges 28 JECFA and CCRVDF have not in the past usually recommended analytical methods for 29 residues of substances for which no ADI or MRLVD has been established. There are some 30 apparent exceptions, but these relate to compounds for which an ADI and MRLVDs have 31 been established and later withdrawn, or to compounds for which there has been sufficient 32 information available that a suitable monitoring method could be identified to assist national 33 authorities. 34 Experiments suitable for providing statistically based evidence of performance of 35 screening tests are similar to those described in a performance verification programme that 36 was established to demonstrate the suitability of test kits (AOAC, 1994). For validation of 37 confirmatory methods based on MS, requirements have been elaborated in documents 38 published by the European Communities (EC, 2002), the United States Food and Drug 39 Administration (FDA, 2003) and the American Society for Mass Spectrometry (Bethem et al., 40 2003). These require the presence of a match with the chromatographic retention time of a 41 standard, the presence of a minimum number of characteristic (structurally significant) 42 fragment ions and an agreement within specified limits between the ratios of the ions present 43 in a standard with those present in the sample. 44 45 3.9 Contaminants 46 The data required for the characterization of a contaminant should include its concentrations 47 in foods and the total diet, from as many countries as possible. The sixty-fourth meeting of 48 JECFA recommended that the data should be formatted using the Global Environment 49 Monitoring System – Food Contamination Monitoring and Assessment Programme 50 (GEMS/Food) to facilitate the collation and quality control of the data. The data should be

18 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 accompanied by additional details on sampling plans and analytical methods used to generate 2 the data. 3 4 3.9.1 Sampling plans 5 The generation of meaningful surveillance data requires the collection of representative 6 samples from carefully selected batches of food, which, in turn, are representative of clearly 7 defined locations (e.g. country, region within a country). Figure 3.1 is a simple schematic 8 diagram that demonstrates the relationship between analytical samples and the GEMS/Food 9 Region under evaluation (details of the GEMS/Food database are given in chapter 6). It is 10 clear that the representative nature of regions, countries, samples, subsamples and analytical 11 samples must be maintained throughout the survey, if the final analytical result is to be 12 accurately related to the GEMS/Food Region under evaluation. For example, if differing 13 agroclimatic regions within a country are likely to result in differing levels of mycotoxin 14 contamination, it is essential that representative batches of food are carefully selected from 15 each region. 16

17 18 19 Figure 3.1. The relationship between analytical samples and the GEMS/Food Region under 20 evaluation. 21 22 Once a batch has been selected, it is equally important that the sample is collected 23 using a clearly defined sampling plan, which aims to produce a representative sample. 24 Although sampling variability is unavoidable, it is essential that the precision of the sampling 25 plan is defined and considered to be acceptable. In general: 26 27 • A minimum of 30 batches of food should be sampled from each region (i.e. country, 28 agroclimatic region within a country, etc.). 29 • “Samples” = initial samples taken from the batch. These may be large bulk samples or 30 smaller composite samples.

19 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 • “Subsamples” = samples produced by riffle division of the unground bulk samples OR 2 samples produced by comminuting and subdividing the smaller composite samples. 3 • “Analytical samples” = samples subjected to analysis. These will usually be small 4 portions withdrawn from subsamples. 5 • The precision (coefficient of variation between different analytical samples) of the 6 sampling plan should be no more than 30%. 7 • The precision (coefficient of variation) of the complete analytical method (extraction, 8 cleanup, quantification) should be no more than 10%. 9 10 Contaminants in food commodities may result from environmental contamination by 11 persistent compounds formerly used as pesticides (e.g. persistent organochlorine pesticides). 12 The JMPR proposes limits (extraneous maximum residue limits, EMRLs) for such 13 contaminants when they originate from environmental sources and not from direct or indirect 14 uses on the crop or farm animal. 15 In 1990 (FAO, 1990), JMPR explained that EMRL assessments rely on monitoring 16 data and supporting information, including: 17 18 • country; 19 • year; 20 • commodity and portion analysed; 21 • pesticide and residue definition; 22 • sample classification as import, export or domestic production and consumption; and 23 • sampling plan described as random monitoring or target sampling. 24 25 Ideally, for reasonable EMRL estimates to cover international trade, JMPR should 26 have current and geographically representative data (FAO, 1995), but typically data are 27 available from only three or four (usually developed) countries. JMPR requests the 28 submission of all relevant data, including nil results. Because residues gradually decrease, 29 new data should be assessed every few years with a view to EMRL revision. 30 31 3.9.2 Analytical methods 32 The LOQs of the analytical methods to measure the concentrations of contaminants in foods 33 (on raw basis or as consumed basis) should be as low as reasonably possible (usually much 34 lower than the regulation limit). This consideration is of critical importance in exposure 35 estimations, because low levels of contaminants are frequently present in foods, and the 36 censored data (data points with non-quantified results) represent a bias source in calculations 37 of exposure. If the LOQ is not sufficiently low, then there is a risk of underestimation if all 38 non-detects are taken as zero or overestimation if all non-detects are taken as the LOQ. To 39 minimize this bias, it is recommended that the censored data should be treated following the 40 statistical approach discussed in chapter 6. 41 42 3.9.3 Reactions and fate of contaminants in food 43 Contaminants in food that have been evaluated by JECFA include environmental 44 contaminants and substances migrating from food packaging; metals are the environmental 45 contaminants that have been considered most often. Committees request information on the 46 chemical forms of metals in the food supply (e.g. ionic form and/or covalently bonded 47 chemical forms) and their concentration distribution in the food supply, as determined by 48 analyses of food or experimental models for carryover from environmental sources. For 49 contaminants derived from food packaging, data are required on the identification of

20 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 chemicals migrating from the packaging material and concentrations in food (analysed or 2 estimated from migration modelling studies). 3 4 3.10 Substances consumed in large amounts 5 Thorough chemical analysis should be performed on high-consumption substances, such as 6 bulk additives, to measure potential impurities and to provide information on nutritional 7 adequacy, especially when such substances replace traditional food. 8 It is not possible to provide a checklist of necessary chemical studies to cover all 9 high-consumption compounds. The substance should be subjected to a full analysis, and 10 particular attention should be paid to the points discussed in the following paragraphs. 11 Because the exposure to undesirable impurities concomitant with the intake of high- 12 consumption materials is potentially high, special effort should be made to identify the 13 impurities. Information on the production process, including the materials and procedures 14 involved, will point to the types of contaminants for which limits may need to be specified. 15 The specifications should be accompanied by details of product variability and of the 16 analytical methods used to check the specifications and details of the sampling protocols. If 17 the substance is so complex that comprehensive product specifications on chemical 18 composition are impracticable (as it might be for a microbial protein), the description of the 19 substance in the specifications may include relevant aspects of its manufacturing process. If 20 manufacturing data are based on production on a pilot scale, the manufacturer should 21 demonstrate that, when produced in a large-scale plant, the substance will meet the 22 specifications established on the basis of pilot data. 23 The permissible limits for impurities may in some cases correspond to the levels 24 accepted for natural foods that have similar structure or function or that are intended to be 25 replaced by the new material. If the substance is prepared by a biological process, special 26 attention should be paid to the possible occurrence of natural toxins (e.g. mycotoxins). 27 The substance should be analysed for the presence of toxic metals. Depending on the 28 intended use, analysis for metals of nutritional significance may also be appropriate. 29 If the nature of the substance or manufacturing process indicates the possible presence 30 of naturally occurring or adventitious anti-nutritional factors (phytate, trypsin inhibitors, etc.) 31 or toxins (haemagglutinins, mycotoxins, nicotine, etc.), the product should be analysed for 32 them specifically. Biological tests, either as part of the nutritional evaluation in the case of 33 enzyme inhibitors or more specifically as part of a mycotoxin screening programme, will 34 provide useful backup evidence concerning the presence or absence of these contaminants. 35 Finally, if under the intended conditions of use the substance may be unstable or is 36 likely to interact chemically with other food components (e.g. degradation or rearrangement 37 of the substance during heat processing), data should be provided on its stability and 38 reactivity. The various tests should be conducted under conditions relevant to the use of the 39 substance (e.g. at the acidity and temperature of the environment and in the presence of other 40 compounds that may react). 41 42 3.11 References 43 Alder, L., Holland, P.T., Lantos, J., Lee, M., MacNeil, J.D., O’Rangers, J., van Zoonen, P. and Ambrus, A. 44 (2000). Guidelines for Single-Laboratory Validation of Analytical Methods for Trace-level Concentrations 45 of Organic Chemicals in Principles and Practices of Method Validation, Fajgelj, A., & Ambrus, A. 46 (ed.).ISBN 0-85404-783-2. The Royal Society of Chemistry, Cambridge, UK, pp. 179-248. 47 Ambrus, A., Hamilton, D.J., Kuiper, H.A., and Racke, K.D. (2003). Significance of Impurities in the Safety 48 Evaluation of crop protection products (IUPAC Technical Report) Pure Appl. Chem., Vol. 75, No. 7: 937– 49 973. 50 AOAC (1994). Test Kit Performance Testing Program: Policies and Procedures. AOAC Research Institute, 51 Gaithersburg, MD, USA.

21 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 Bethem, R, Boison, J.O., Gale, J., Heller, D., Lehotay, S., Loo, J., Musser, S., Price, P., and Stein, S. (2003). 2 Establishing the Fitness for Purpose of Mass Spectrometric methods. Journal of the American Society for 3 Mass Spectrometry 14: 528-541. 4 EC (2002). European Commission Decision 2002/657/EC, implementing Council Directive 96/23/EC 5 concerning the performance of analytical methods and the interpretation of results, Official Journal of the 6 European Communities, L221: 8-36. 7 EURACHEM. (1998). The Fitness for Purpose of Analytical Methods A Laboratory Guide to Method 8 Validation and Related Topics, EURACHEM. http://www.eurachem.ul.pt/guides/valid.pdf (Accessed 5 9 February 2007). 10 Fajgelj, A. and Ambrus, Á. (2000). Principles and Practices of Method Validation. Proceedings of the Joint 11 AOAC/FAO/IAEA/IUPAC International Workshop on the Principles and Practices of Method Validation, 12 Budapest, Hungary, November 1999. The Royal Society of Chemistry, Cambridge, UK. (Report accessed at 13 http://www.iaea.org/trc/pest-qa_val.htm 5 February 2007). 14 FAO (1978). Guide to specifications: general notices, general methods, identification tests, test solutions, and 15 other reference materials. Food and Agriculture Organization of the United Nations, Rome (FAO Food and 16 Nutrition Paper 5). 17 FAO (1983). Guide to specifications: general notices, general methods, identification tests, test solutions, and 18 other reference materials. Food and Agriculture Organization of the United Nations, Rome (FAO Food and 19 Nutrition Paper 5, Rev. 1). 20 FAO (1990) Data requirements for extraneous residue limit estimations. Pesticide Residues in Food Report 1990. 21 Food and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection 22 Paper 102: 4-5). 23 FAO (1991a). Guide to specifications: general notices, general methods, identification tests, test solutions, and 24 other reference materials. Food and Agriculture Organization of the United Nations, Rome (FAO Food and 25 Nutrition Paper 5, Rev. 2). 26 FAO (1991b). Fat-soluble pesticides. Pesticide Residues in Food Report 1991. Food and Agriculture 27 Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 111: 15-16). 28 FAO (1993a). Maneb. Pesticide Residues in Food 1993. Evaluations. Part I. Residues. Food and Agriculture 29 Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 124: 705). 30 FAO (1993b). Mancozeb. Pesticide Residues in Food 1993. Evaluations. Part I. Residues. Food and Agriculture 31 Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 124: 673). 32 FAO (1993c). Codex Alimentarius Commission , Volume 3; Residues of Veterinary Drugs in Foods, 2nd ed. 33 FAO (1995) JMPR approach to estimating extraneous residue limits (ERLs). Pesticide Residues in Food Report 34 1995. Food and Agriculture Organization of the United Nations, Rome (FAO Plant Production and 35 Protection Paper 133: 21-23). 36 FAO (1998a). 2.8. Data requirements for the validation of analytical procedures. Report on Pesticide Residues 37 in Food. Food and Agriculture Organization of the United Nations, Rome (FAO Plant Production and 38 Protection Paper 148: 11). 39 FAO (1998b). 4.15 Glufosinate-ammonium. Pesticide Residues in Food Report 1998. Food and Agriculture 40 Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 148). 41 FAO (1999). Relevance of pesticide specifications for JMPR evaluation. Pesticide Residues in Food Report 42 1999. Food and Agriculture Organization of the United Nations, Rome (FAO Plant Production and 43 Protection Paper 153: 16-17). 44 FAO (2002a). Esfenvalerate. Pesticide Residues in Food 2002. Evaluations. Part I. Residues. Food and 45 Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 175: 46 600-601). 47 FAO (2002b). Submission and evaluation of pesticide residues data for the estimation of maximum residue 48 levels in food and feed. Food and Agriculture Organization of the United Nations, Rome (FAO Plant 49 Production and Protection Paper 170). 50 FAO (2002c). Residues of some veterinary drugs in animals and foods. Food and Agriculture Organization of 51 the United Nations, Rome (FAO Food and Nutrition Paper 41/14). 52 FAO (2005a). Joint FAO/WHO Meeting on Pesticide Specifications. Pesticide Residues in Food Report 2005. 53 Food and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection 54 Paper 183: 33-34). 55 FAO (2005b). Definition of fat-soluble pesticides in meat and fat. Pesticide Residues in Food Report 2005. Food 56 and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 57 183: 27-31). 58 FAO (2005c). Indoxacarb. Pesticide Residues in Food 2005. Evaluations. Part I. Residues. Food and Agriculture 59 Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 184: 518).

22 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 FAO (2005/2006). Combined Compendium of Food Additive Specifications, Volume 1 – 4. Food and 2 Agriculture Organization of the United Nations, Rome (FAO JECFA Monographs 1). 3 FAO/IAEA (1998). Report of the Joint FAO/IAEA Expert Consultation on Validation of Analytical Methods for 4 Food Control, 2-4 December 1997, Vienna. Food and Agriculture Organization of the United Nations, 5 Rome. 6 FAO/WHO (1956). Joint FAO/WHO Conference on Food Additives (FAO Nutrition Meetings Report Series No. 7 11; WHO Technical Report Series No. 107). 8 FAO/WHO (1957). General principles governing the use of food additives. First report of the Joint FAO/WHO 9 Expert Committee on Food Additives (FAO Nutrition Meetings Report Series No. 15; WHO Technical 10 Report Series No. 129). 11 FAO/WHO (1958). Procedures for the testing of intentional food additives to establish their safety for use. 12 Second report of the Joint FAO/WHO Expert Committee on Food Additives (FAO Nutrition Meetings 13 Report Series No. 17; WHO Technical Report Series No. 144. 14 FAO/WHO (1966). Specifications for the identity and purity of food additives and their toxicological evaluation: 15 some antimicrobials, antioxidants, emulsifiers, stabilizers, flour-treatment agents, acids, and bases. Ninth 16 report of the Joint FAO/WHO Expert Committee on Food Additives (FAO Nutrition Meetings Report 17 Series No. 40; WHO Technical Report Series No. 339). 18 FAO/WHO (1969). Pesticide residues in food. Report of the 1968 Joint Meeting of the FAO Working Party of 19 Experts on Pesticide Residues and the WHO Expert Committee on Pesticide Residues, Geneva, World 20 Health Organization (FAO Agricultural Studies No. 78; WHO Technical Report Series, No. 417). 21 FAO/WHO (1971). Evaluation of food additives specifications for the identity and purity of food additives and 22 their toxicological evaluation: some extraction solvents and certain other substances; and a review of the 23 technological efficacy of some antimicrobial agents. Fourteenth report of the Joint FAO/WHO Expert 24 Committee on Food Additives (FAO Nutrition Meetings Report Series No. 48; WHO Technical Report 25 Series No. 462). 26 FAO/WHO (1972). Evaluation of food additives: some enzymes, modified starches, and certain other 27 substances: toxicological evaluations and specifications and a review of the technological efficacy of some 28 antioxidants. Fifteenth report of the Joint FAO/WHO Expert Committee on Food Additives (FAO Nutrition 29 Meetings Report Series No. 50; WHO Technical Report Series No. 488). 30 FAO/WHO (1975a). Evaluation of certain food additives: some food colours, thickening agents, smoke 31 condensates, and certain other substances. Nineteenth report of the Joint FAO/WHO Expert Committee on 32 Food Additives (FAO Nutrition Meetings Report Series No. 55; WHO Technical Report Series No. 576). 33 FAO/WHO (1975b). Pesticide residues in food. Report of the 1974 Joint Meeting of the FAO Working Party of 34 Experts on Pesticide Residues and the WHO Expert Committee on Pesticide Residues, Geneva, World 35 Health Organization (FAO Agricultural Studies No. 97; WHO Technical Report Series, No. 574). 36 FAO/WHO (1978). Pesticide residues in food 1977. Report of the Joint Meeting of the FAO Panel of Experts on 37 Pesticide Residues and Environment and the WHO Expert Committee on Pesticide Residues, Food and 38 Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 10 39 Rev.). 40 FAO/WHO (1979). Pesticide residues in food 1978. Report of the Joint Meeting of the FAO Panel of Experts on 41 Pesticide Residues in Food and Environment and the WHO Expert Group on Pesticide Residues, Rome, 42 Food and Agriculture Organization of the United Nations (FAO Plant Production and Protection Paper 15). 43 FAO/WHO (1980). Pesticide residues in food 1979. Report of the Joint Meeting of the FAO Panel of Experts on 44 Pesticide Residues in Food and the Environment and the WHO Expert Group on Pesticide Residues. Food 45 and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 20). 46 FAO/WHO (1981a). Evaluation of certain food additives. Twenty-fifth report of the Joint FAO/WHO Expert 47 Committee on Food Additives (WHO Technical Report Series No. 669). 48 FAO/WHO (1981b). Pesticide residues in food 1980. Report of the Joint Meeting of the FAO Panel of Experts 49 on Pesticide Residues in Food and the Environment and the WHO Expert Group on Pesticide Residues. 50 Food and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection 51 Paper 26). 52 FAO/WHO (1983a). Evaluation of certain food additives and contaminants. Twenty-seventh report of the Joint 53 FAO/WHO Expert Committee on Food Additives (WHO Technical Report Series No. 696 and corrigenda). 54 FAO/WHO (1983b). Pesticide residues in food 1982. Report of the Joint Meeting of the FAO Panel of Experts 55 on Pesticide Residues in Food and the Environment and the WHO Expert Group on Pesticide Residues. 56 Food and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection 57 Paper 46). 58 FAO/WHO (1985). Pesticide residues in food 1984. Report of the Joint Meeting on Pesticide Residues. Food 59 and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 62).

23 Update Project Chapter 3: Analytics and Specifications Draft May 2008

1 FAO/WHO (1987). Pesticide residues in food 1987. Report of the Joint Meeting of the FAO Panel of Experts on 2 Pesticide Residues in Food and the Environment and a WHO Expert Group on Pesticide Residues. Food 3 and Agriculture Organization of the United Nations, Rome (FAO Plant Production and Protection Paper 84). 4 FAO/WHO (1988). Evaluation of certain veterinary drug residues in food. Thirty-second Report of the Joint 5 FAO/WHO Expert Committee on Food Additives. (WHO Technical Report Series 763). 6 FAO/WHO (1997a). Guidelines for the assessment of the competence of testing laboratories involved in the 7 import and export control of food. CAC/GL 27-1997, Codex Alimentarius Commission, Rome. 8 FAO/WHO (1997b). Report of the Ninth Session of the Codex Committee on Residues of Veterinary Drugs in 9 Foods. Alinorm 97/31. Codex Alimentarius Commission, Rome. 10 FAO/WHO (1997c). Report of the Tenth Session of the Codex Committee on Residues of Veterinary Drugs in 11 Foods. Alinorm 97/31A. Codex Alimentarius Commission, Rome. 12 FAO/WHO (1999). Report of the Eleventh Session of the Codex Committee on Residues of Veterinary Drugs in 13 Foods. Alinorm 99/31. Codex Alimentarius Commission, Rome. 14 FAO/WHO (2002). Manual on development and use of FAO and WHO specifications for pesticides, 1st edition. 15 FAO Plant production and protection Paper 173. FAO, Rome 2002. (Available at 16 http://whqlibdoc.who.int/publications/2002/9251048576.pdf 17 FAO/WHO (2003). Codex Alimentarius Commission Procedural Manual, 13th Ed., Food and Agriculture 18 Organization of the United Nations, Rome. 19 FAO/WHO (2005). Codex Alimentarius Commission Procedural Manual, 15th Ed., Food and Agriculture 20 Organization of the United Nations, Rome. 21 FAO/WHO (2006). Manual on development and use of FAO and WHO specifications for pesticides. March 22 2006 revision of the First Edition (available at http://www.fao.org/ag/AGP/AGPP/Pesticid/Default.htm , 23 accessed 7 February 2007). 24 FDA (2003). The United States of America Food & Drug Administration. Guidance for Industry - Mass 25 Spectrometry for Confirmation of the Identity of Animal Drug Residues. 26 (at www.fda.gov/cvm/guidance/guide118.doc , accessed 7 February 2007). 27 Green, W.I. (1978). Mechanisms of action of anti-thyroid compounds. In: Werner, S.C. & Ingbar, S.H., ed. The 28 thyroid, New York, Harper and Row, pp. 77-87. 29 Hill, A.R.C. and Reynolds, S.L. (1999). Guidelines for in-house validation of analytical methods for pesticide 30 residues in food and animal feeds. Analyst 124: 953-958. 31 ISO/IEC-17025:2005 (2005). General requirements for the competence of calibration and testing laboratories. 32 International Organization for Standardization, Geneva. 33 (at http://domino.iec.ch/preview/info_isoiec17025{ed2.0}en.pdf , accessed 5 February 2007). 34 Skidmore, M.W., Paulson, G., Kuiper, H.A., Ohlin, B. and Reynolds, S. (1998). Bound xenobiotic residues in 35 food commodities of plant and animal origin. Pure and Applied Chemistry 70: 1423-1447. 36 Thompson, M. and Wood, R. (1993). International Harmonized Protocol for Proficiency Testing of (Chemical) 37 Analytical Laboratories. Pure & Appl. Chem. 65: 2132-2144. 38 Thompson, M. and Wood, R. (1995). Harmonized Guidelines for Internal Quality Control in Analytical 39 Chemistry Laboratories. Pure & Appl. Chem. 67: 649-666. 40 Thompson, M., Ellison, S.L.R. and Wood, R. (2002). Pure and Applied Chemistry, 74: 835-852. 41 WHO (1967). Procedures for investigating intentional and unintentional food additives. Report of a WHO 42 Scientific Group. World Health Organization, Geneva, (WHO Technical Report Series, No. 348). 43 Youden, W.J. and Steiner, E.H. (1975). Statistical Manual of the Association of Official Analytical Chemists, 44 AOAC International, Gaithersburg, VA, USA.

24