2. General considerations

2.1 The WHO classification of pesticides by hazard

"The WHO recommended classification of pesticides by hazard" (IPCS, 2002) is prepared for and refers to "the acute risk to health that might be encountered accidentally by any person handling the product".

The classification is generally based on the oral and/or dermal LD₅₀ in rats, but exceptions apply (for a complete discussion of the criteria of this classification see Part I of the document).

As such, the criteria guiding and the purpose of this classification are distinct from the considerations and procedures that are required when assessing acute dietary risk, including the derivation of the acute reference dose, when necessary. Consequently, potential confusion might occur between the "WHO classification of pesticides by hazard" and the performance of acute dietary risk assessment. For this reason, the Meeting considered that it was no longer appropriate to include this classification in future JMPR evaluations of the compound under consideration.

2.2 Setting the acute reference dose on the basis of haematological effects

The 2001 and 2002 Joint Meetings indicated that haematological effects (e.g. methaemoglobinaemia, haemolytic anaemia) could arise from a single exposure to a chemical and hence such effects may form the basis for an acute reference dose (RfD). The 2003 Meeting set acute RfDs for famoxadone, methoxyfenozide and tebufenozide on the basis of haemolytic anaemia which occurred after repeated dosing. The mechanism by which these compounds cause haemolysis is unknown. Anaemia was accompanied, when measured, by significantly increased (but by <10%) methaemoglobinaemia. There are no data available to indicate that haemolysis caused by these compounds occurred via the formation of methaemoglobin. However, increased methaemoglobin may be a marker of oxidative stress to the erythrocyte, which is a cause of haemolysis.

The Meeting reviewed three compounds, tebufenozide, methoxyfenozide and famoxadone all with low acute oral toxicity (LD₅₀ of >5000 mg/kg in rats). In dogs given a single dose of tebufenozide, no haematological effects were observed at 89 mg/kg bw, the highest dose tested. However, haematological effects occurred in dogs after repeated doses of 5 mg/kg bw per day for 2 weeks. Methoxyfenozide caused minimal haematological effects at a dose of about 200 mg/kg bw per day for 2 weeks, with a clear NOAEL of about 20 mg/kg bw per day. Famoxadone caused mild haematological effects after 30 days (but not after 16 days) of treatment with about 60 mg/kg bw per day.

Together, these data suggest that haematotoxicity would not necessarily be expected to occur after a single dose of these compounds. Therefore, the Meeting recognizes that the setting of acute RfDs for these three compounds is based on conservative assumptions and that refinements of the approach taken by the 2003 Meeting may be possible in the future.

For such refinements, general guidance or a framework on how to perform single dose studies to address endpoints such as methaemoglobinaemia, haematotoxicity and other acute alerts would be useful. Therefore, the 2003 JMPR recommended that WHO establishes a Working Group including scientists who have helped develop the concepts of the acute RfD at JMPR. The Working Group should develop further guidance on how to interpret existing databases, including single-dose studies and early data from repeated-dose studies, and how to perform single-dose studies in relation to establishing an acute RfD. This guidance will to be submitted for consideration by the 2004 JMPR.

2.3 Review of provision of scientific advice

In 2001, the Codex Alimentarius Commission requested that FAO and WHO review the status and procedures of the expert bodies, such as JMPR, and develop recommendations on additional ways to improve the quality, quantity and timeliness of scientific advice to the Commission. The evaluation of Codex Alimentarius, conducted in 2002 considered this to be a matter of urgency. FAO and WHO have initiated a full review of the provision of scientific advice to Codex and Member States via expert committees and ad hoc consultations. This review will take into account ongoing activities to improve working procedures. The evaluation of the JMPR, conducted in 2001, will prove to be an important source of information in this regard. The overall review will be performed through a consultative process that started with an FAO/WHO Planning Meeting in May 2003. A Joint Workshop will be held in January 2004. This Workshop will be preceded by an electronic forum debate on key issues identified by the Planning Meeting. An independent expert consultation will be held in mid 2004 to agree upon recommendations for consideration by FAO and WHO.

2.4 Project to Update the Principles and Methods for the Risk Assessment of Chemicals in Food

The Meeting recognized the importance of this Project, particularly in view of the evolution of the risk analysis paradigm and the processes used by the JMPR. The Meeting strongly urged the timely completion of this Project and recommended that the Project:

- give clear general principles for risk assessment procedures;

- include guidance for special toxicological considerations, such as the acute RfD;

- give general guidance for analytical methods, including fitness for purpose, sampling, and quality assurance;

- give clear guidance for exposure assessment, including assessment of acute exposures,

- include clear description, with uncertainties, of the process by which maximum residue limits (MRLs) are recommended;

- for residues of pesticides and veterinary drugs, where harmonization of certain risk assessment approaches may not be possible, give clear, transparent justifications for the differences.

2.5 Selective surveys to provide residue data for estimating MRLs for spices

In response to the request of the 34th Session of the CCPR, the 2002 JMPR considered the options for estimating maximum residue levels for spices based on monitoring data (JMPR Report 2002 section 2.7.) and provided guidance on the format for reporting such data. As the 35th Session of the CCPR decided to elaborate MRLs based on monitoring data (Paras 187-200, ALINORM 03/24A, 2003), the present Meeting gave further consideration to possible options for estimating maximum residue levels where sufficient monitoring data are not available and prepared guidelines for conducting selective surveys to generate pesticide residue data reflecting the field and post-harvest application of pesticides.

The information presented to the CCPR or the JMPR indicated that registered or permitted uses of pesticides on specific spices may not be generally available, and farmers may use any pesticides available on the market to protect their spices that they found effective to control pests and diseases on vegetables. In addition, the spices may be indirectly exposed to pesticides which are applied to the main crops.

Post-harvest treatment is usually made on a spice sourced from several fields. The crops on the fields may have been exposed to different pesticides, which may increase the number of pesticide residues derived from pre-harvest applications.

In view of the current practice in growing spice-producing plants it may be necessary to estimate maximum residue levels for pesticides based on general agricultural practice but without established GAP. Therefore information on existing national MRLs for spices, together with registered or authorized uses of pesticides on main crops of fields where spices are grown is important for a comprehensive evaluation of the residue data. National data requirements and methods of evaluation should be included in the data submission where available.

Spices are usually difficult substrates for the determination of trace organic contaminants. Reliable identification of pesticide residues and their quantitative determination in spice samples of unknown origin can be a very laborious and complicated task especially where access to GC-MS and LC-MS-MS techniques is limited. The analysts should, therefore, have as much information as possible on the actual or possible use of pesticides on the spices to be analysed.

In a selective field survey samples are taken from fields where the crop is grown, treated directly or indirectly with pesticides, and harvested according to the local agricultural practice. The essential feature of the selective field survey is that all pesticide applications, the growth stage of the crop and post-harvest treatment of spices are recorded, and the records are attached to the sampling report. Therefore the laboratory can look for all pesticides applied, in addition to organochlorine pesticides which may be taken up from soil.

Since the uses of pesticides are known, the selective survey is a better alternative than monitoring of pesticide residues in samples of unknown origin in cases where there are not enough residue data available from previous work.

Guidelines for conducting selective surveys

- A successful survey requires full co-operation with the growers who should understand that it is conducted for promoting their production and that the correct information is essential for success.

- Sites for survey should be selected to represent the growing conditions of the particular spice. The more information and residue data provided the more realistic the maximum residue level that can be estimated.

- The minimum number of fields surveyed and samples collected depends on the diversity of the growing conditions. As an initial step, the Meeting considered that a minimum of 10 reliable residue results representing the typical growing or processing conditions with supplementary information are required for each spice-pesticide combination.

- In the case of post-harvest application the minimum of 10 samples should be taken from lots treated independently, preferably at different processing plants.

The following details should be reported.

• Person and organization responsible for organizing, supervising and reporting on the selective field survey

• Commodity description with Codex code number if available or classification by the CCPR; (Para 199, ALINORM 03/24A, 2003)

• The typical agricultural practice concerning the use of pesticides, and if available, the registered or permitted uses of pesticides on the plant producing the spice (dosage, number and method of applications, timing of application and growth stage of plant)

• Where the plant is grown between rows of a major crop, the registered or permitted uses of pesticides on the major crop

• Description of growing conditions of the plant producing the spice (e.g. main or intermediate crop), the growth stage at harvest, date of harvest and harvested part of the plant

• The date and method of application, and dosage of pesticides actually applied, for treatments

  ¨ carried out on the fields where the samples are taken directly from the fields, or

  ¨ details of post-harvest application together with information on pre-harvest treatments where available

• Description of the processing of the spice and its storage conditions

• Storage conditions of samples until analysis

• Portion of sample analysed

• Performance parameters of the method (LOQ, mean recovery and its CV at various fortification levels), methods of detection and confirmation of residues

• Residues of ai and metabolites (mg/kg) found in the samples.

Results should be tabulated as shown below.

Crop/spice

¹ indicate whether the application was direct or indirect.

Since the estimation of maximum residue levels based on monitoring and survey data is a new approach it may be necessary to revise the requirements for residue data and reported information, taking into account the experience gained with the evaluation of the submissions.

2.6 Expression of MRLs for fat-soluble pesticides in milk and milk products

At the 35th (2003) Session of the CCPR, concern was expressed at the very low MRLs proposed for some fat-soluble pesticides (e.g. diphenylamine, chlorpropham) in milk. The JMPR was requested to take into account the LOQs achievable in practice. The specific issue of diphenylamine is considered separately in this Report.

The fat content of milk varies widely. In addition, there are many milk products with varying fat contents and it would be difficult to propose separate MRLs for each of them. It was therefore originally decided to estimate MRLs for fat-soluble compounds in milk and milk products on a fat basis, i.e. on the residue levels expressed as if wholly contained in the fat. This simple system was satisfactory for very fat-soluble persistent organochlorine pesticides because it allowed the MRLs for milk to be extrapolated to all milk products, irrespective of fat content.

Currently the JMPR follows the Codex convention of expressing the MRLs for fat-soluble compounds in milks on a calculated whole-product basis, assuming all milks to contain 4% fat (FAO/WHO, 2000)^[1]. The residue concentration is calculated for the whole product on the basis of the concentration measured in the fat. The MRL would be 1/25th of the residue concentration estimated for the milk fat. For a milk product with a fat content below 2%, the MRL is half that specified for milk. Fat-soluble pesticides to which these general provisions apply are indicated by the letter "F" in conjunction with the MRL specified for milk. MRLs for pesticides which are of low solubility in fat are estimated on a whole milk basis and do not have the suffix "F".

However, many pesticides are of intermediate fat-solubility and, even though regarded as fat-soluble, may be distributed between the fat and aqueous phases of the milk. For example, if the ratio of the residue concentration between the fat and aqueous phases is 15:1 in milk with 4% fat, the ratio of the total residue in the two phases is about 2:3, meaning that most of the residue remains in the aqueous phase. To comply with the Codex convention, MRLs for such less fat-soluble pesticides have the same basis as those for water-soluble pesticides and do not have the suffix "F".

The Meeting recognized that the current methods for deriving and expressing MRLs for milk may appear to be complex and cause problems for the CCPR and enforcement authorities. An example of such a problem was the decision of the 1995 CCPR to delete the suffix "F" from MRLs set at the LOQ (ALINORM 95/24A, paragraph 180). Contrary to the intention of the JMPR in recommending such MRLs, the CCPR decision implied that analysis for compliance testing must be conducted on whole milk and that the same MRL value is then applied to both milk and milk products.

The JMPR decides whether or not an MRL for a fat-soluble compound in milk should have the suffix "F" by considering the following characteristics of the compound(s) that are covered by the definition of the residue: (i) the log P_ow, (ii) the solubility in fatty animal tissues, and (iii) the distribution between the fat and non-fat fractions of the milk, where available.

The rationale for recommending an MRL for a fat-soluble pesticide with the suffix "F" is illustrated with fenvalerate/esfenvalerate as an example.

The log P_ow of 6.2 and the animal metabolism studies indicated that fenvalerate/esfenvalerate should be described as soluble in body fat, and almost all the residue was found to partition into the fat of milk. The JMPR therefore concluded that fenvalerate/esfenvalerate residues should be described as fat-soluble and recommended an MRL in milk of 0.1 mg/kg F.

The rationale for recommending an MRL for a fat-soluble pesticide without the suffix "F" can be illustrated by using spinosad as an example.

The log P_ow values of 4 and 4.5 for spinosyns A and D, together with the animal metabolism studies, indicated that spinosad was soluble in body fat. However spinosad residues were incompletely partitioned into the fat of milk. In a study of the treatment of dairy cows, the ratio of the residue level in cream to that in milk was 4.2 (mean of 119 observations). In a dairy cow feeding study, residue levels in cream were 3-5 times that in the milk. The JMPR concluded that spinosad should be described as fat-soluble for the purpose of measuring residues in meat but that it should not be treated as fat-soluble for the measurement of residues in milk, and therefore noted that the residue is fat-soluble but residues in milk should be measured on the whole milk. The MRL recommended for cattle milk was 1 mg/kg (without the suffix "F").

The Meeting drew attention to the requirements for testing compliance with MRLs. Where the suffix "F" is appended, the separated and extracted milk fat is analysed; where there is no suffix, the whole milk is analysed. For the determination of recoveries, milk fat or whole milk is fortified, depending on the presence or absence of the suffix "F".

To apply an MRL for milk which has the suffix "F" to milk products, the MRL for milk is multiplied by 25 and the resultant value applies to the fat extracted from the milk product. The residue in the milk fat is determined and neither the MRL nor the analytical result is adjusted for the fat content of the milk product. To apply an MRL for milk which does not have the suffix "F" to milk products, the MRL is used without change and the whole product is analysed to determine compliance with the MRL. The two procedures are not influenced by whether the MRL is at the LOQ or at some higher value.

2.7 Refinement of the estimations of maximum residue levels for processed commodities

During recent years, many relevant processing studies have been provided to the JMPR. These are normally conducted according to national authority requirements and simulate commercial practices. Whenever possible, a processing factor (PF) is calculated and residues in the processed commodity estimated by multiplying the estimates in the raw commodity by the PF.

The JMPR has been working to increase the transparency of the evaluation process, improve the quality of the estimations and harmonize the procedures at international level. The Meeting agreed to outline the procedures for estimating maximum residue levels, STMR-Ps and HR-Ps for processed commodities which have been applied previously and/or refined at this Meeting.

Maximum residue levels. Until now, a maximum residue level in a processed commodity was derived by applying the calculated processing factor to the maximum residue level estimated for the raw commodity, and the resulting number was rounded up to the closest value from the scale of numerical values adopted by the JMPR in 2001^[2]. The maximum residue level in the raw commodity is, however, already a rounded-up level.

The procedure of rounding up residue values twice to achieve an estimate is not scientifically justified, and the Meeting agreed that the maximum residue level in processed commodities should be derived directly from the highest residue found in the supervised trial.

Maximum residue levels will only be estimated for processed commodities in which the residue is concentrated during the processing procedure (PF >1) and for which there is an existing Codex commodity code. When the PF is <1, the MRL in the raw commodity applies also to the processed commodity (Maximum limits for processed or ready-to-eat foods and feed, ALINORM 03/24A, 2003). Estimates of STMR-Ps and HR-Ps will be made whenever necessary, regardless of whether the PF is higher or lower than 1.

Processing factor. The processing factor is defined as the ratio of the residue found in the processed commodity to the residue in the raw commodity before processing. In many cases, the processed commodity does not have quantifiable residues (<LOQ), and the calculated processing factor will be below a certain value (PF <0.01, for example). Although the real processing factor cannot be calculated, the Meeting generally estimates the residue in the processed commodity from the limiting value (PF = 0.01). In cases where residues are not detected in the raw commodity, a processing factor cannot be calculated and estimates in processed commodities cannot be made.

HR-P and STMR-P estimates. An HR-P will only be estimated when it is needed for the calculation of an International Estimated Short-Term Intake (IESTI), that is for commodities covered by Case 1 or Case 2 as defined in Section 3 of this Report.

HR-Ps and/or STMR-Ps for commodities for human consumption will be estimated regardless of the availability of consumption data.

2.8 Development of automated spreadsheet applications for the calculation of dietary intake

The Meeting agreed to adopt automated spreadsheet applications for the calculation of dietary intake in order to harmonize and facilitate the process. The spreadsheet applications were constructed by RIVM/SIR^[3], The Netherlands, in co-operation with WHO/GEMS/Food and incorporated all available consumption data^[4] in Excel spreadsheets. The data were linked as closely as possible to the Codex Commodities for which MRLs, HR(-P)s and STMR(-P)s have been estimated. The spreadsheets are used to calculate IEDIs and IESTIs using the formulae described in Section 3 of this Report. To use the spreadsheets, estimates made by the JMPR of ADIs, acute RfDs, STMR(-P)s, HR(-P)s, and when necessary MRLs, are entered according to the instructions attached to the templates. Calculations and the generation of a final table are then performed automatically. The present Meeting used the spreadsheet applications for the first time and welcomed the resulting reduction of the workload.

The spreadsheet applications are available on the internet at the address http://www.who.int/foodsafety/publication/chem/regionaldiets and will be updated when necessary.

2.9 Improving estimates of dietary intake

The Meeting considered the areas where the estimations of dietary intake could be improved. It was concluded that the calculations are greatly simplified by the automated spreadsheet applications elaborated by RIVM/SIR^[5], The Netherlands, in co-operation with WHO/GEMS/Food. However the calculated values cannot be better than the data-base or estimated factors used. For many Codex commodities for which maximum residue levels, STMR(-P)s and HR(-P)s are estimated, no dietary intake is available. Consequently, immediate refinements could be achieved by:

• improving the accuracy of consumption figures for long-term exposure by introducing the proposed 13 sub-regional diets which have been discussed in recent years by the CCPR and JMPR instead of the currently used five regional diets;

• the increased availability of large portion sizes and unit weights for the calculation of short-term exposure, especially those from developing countries.

Further improvement could be obtained by:

• evaluation of the studies submitted to the JMPR in the last decade representing typical commercial processing to investigate whether it would be possible to derive default processing factors and/or extrapolate processing data;

• refinement of generic and commodity-specific variability factors as used in the short-term intake calculations;

• elaboration of procedures for probabilistic modelling at the international level.

The last two points are discussed further below.

The 35th Session of the CCPR (ALINORM 03/24A, paras 20-31) discussed the paper "Discussion paper on the proposals for improvement methodology for point estimates" [of acute intake of pesticide residues] (CX/PR 03/3). The Report of the Session (para 28) states "The Chair summarized the discussion that: (1) the possibility of accepting limited exceedance should not be considered at present time; (2) the possibility of using a tiered approach could be considered in the future; and (3) JMPR should be asked to mention the probabilistic aspects in the point estimates, when the results exceed the acute RfD."

The Committee also requested the JMPR to consider this paper especially in relation to the use of probabilistic aspects of point estimates.

The Committee also agreed to establish a Working Group to prepare a paper on the possible adoption of probabilistic methodology for the purpose of setting Codex MRLs. This should include worked examples of probabilistic calculations for some compounds, using supervised trials data, where the IESTI exceeds the acute RfD. The Working Group should also discuss and propose parameters to be used in probabilistic calculations at the international level, and the paper should be considered by the next Session of the Committee.

In response, the Meeting agreed in principle to adopt a tiered (i.e. sequential) approach to estimating short-term dietary intake, in which the second tier could be probabilistic modelling. However it also recognized the lack of consumption data and the lack of an available model validated at the international level, which hamper the development of such a second tier. It observed that a possible solution would be to ask the country from which the large portion as used in the JMPR point estimate came, to provide the second tier. However, this would necessitate international consensus on the parameters used in the probabilistic model. The Meeting therefore welcomed the initiative of the CCPR in deciding to establish a Working Group on this subject. The Meeting noted that a probabilistic model useful for JMPR purposes is under development in The Netherlands (RIKILT, Institute of Food Safety) and agreed that it would consider this model when available.

The Meeting took note of the IUPAC report on short-term dietary risk assessment^[6] and on the basis of the evidence presented there agreed to use in future a new default variability factor of 3 in the calculation of residue levels in high-residue units used in point estimates of short-term intake. See also Item 2.10.

In the situation that the IESTI exceeds the acute RfD, the Meeting agreed to indicate in the section on Dietary Risk Assessment ways in which those parameters used in the dietary risk assessments which are based on conservative assumptions might be refined.

2.10 IESTI calculation: refining the variability factor for estimation of residue levels in high-residue units

Current JMPR procedures for estimating the short-term dietary intake of pesticide residues rely on the deterministic procedures proposed by the FAO/WHO Consultation in 1997^[7].

The Consultation proposed methods for calculating short-term intake (1) where the residue in a composite sample reflects the residue level in a meal-sized portion of the commodity, and (2) where the meal-sized portion such as a single unit of fruit might have a higher residue than the composite. The concept of the variability factor was introduced to calculate the residue level in that single unit, originally with the conservative assumption that all of the residue might be in one unit of the composite sample. For fruit such as apples the variability factor was 10 because 10 apples were expected in a typical composite sample.

The 1999 JMPR^[8] summarized the methods for calculating the short-term intake of residues and reported the results of such calculations for the first time. The 1999 Meeting, taking into account the available data, used a variability factor of 7 for most items (>25 g and <250 g), a value of 5 for large items (>250 g) and a value of 10 for granular soil treatments and leafy vegetables.

The 2002 JMPR, on the basis of new data, concluded that a variability factor of 3 would be suitable for residues in head lettuce and head cabbage.

At the 35th Session of the CCPR (Paras 20-31, ALINORM 03/24A, 2003) the delegation of The Netherlands introduced a discussion paper on improved methodology for point estimates of dietary exposure in relation to setting MRLs. An unpublished IUPAC report on short-term dietary risk assessment had been used in the preparation of the paper. The CCPR agreed to establish a Working Group to prepare a paper considering the adoption of probabilistic methodology for the purpose of setting Codex MRLs. The same Session also requested the JMPR to consider possible improvements in the point estimates.

The IUPAC report^[9], now submitted for publication, summarized and analysed the available data on residue level variability from unit to unit for a number of pesticides over a range of crops (apples, carrots, celery, grapes, kiwifruit, lettuce and potatoes).

In cases where the number of unit analyses was large enough to provide 95% assurance that at least one value exceeded the 97.5th percentile, the average variability factor (97.5th percentile value of residue ÷ mean) was 2.7 (range 1.5-7.2) for supervised trials involving approximately 8000 unit analyses. The average variability factor was 3.0 (range 2.4-3.5) for market-place monitoring data involving almost 3000 unit analyses. Most of the estimated variability factors were in the 2.0-3.0 range for the 30 sets of data representing trials and market-place monitoring.

The variability factor did not generally seem to be dependent on the pesticide or the crop. However in one trial with a post-harvest application of a mixture of 3 compounds on apples there appeared to be a separation of the compounds during the post-treatment drainage of the fruit, with one of the compounds producing substantially higher residues in fruit at the bottom of the stack. Variability factors for the three compounds were 7.2, 2.8 and 2.5. In this instance the variability factor was dependent on the pesticide in combination with the method of treatment.

The Meeting agreed to adopt a default variability factor of 3 for the estimation of residue levels in high-residue units in the IESTI calculations where unit weights exceed 25 g. A variability factor is not used in IESTI calculations where unit weights are below 25 g.

The current practice will continue of using specific unit variability factors in preference to the default value where the supporting data are available, valid and sufficient.

2.11 Revised data requirements for studies of environmental fate

The 2003 CCPR considered a proposal to reduce the range of environmental fate studies to be reviewed by the JMPR (CX/PR 03/18, Agenda Item 18). The Committee agreed that the JMPR "... should proceed with the consideration of environmental fate but should focus on those aspects that were most relevant to MRL setting" and that the current data requirements should be revised accordingly (ALINORM 03/24A, para 212). The matter was also referred to the Codex Alimentarius Commission (CAC). The CAC concurred with the CCPR recommendation. (CAC, July 2003, Rome).

The current aspects of the environmental fate of residues required for elucidation, as outlined in the FAO manual on the submission and evaluation of pesticide residues data for the estimation of maximum residue levels in food and feed, Food and Agriculture Organization of the United Nations, Rome, 2002, Second edition, are as follows.

Physical and chemical properties
Metabolism and degradation in soil
Persistence in soil under aerobic and anaerobic conditions
Mobility of the parent compound and its transformation products in soil
Adsorption by various soil types
Hydrolysis rate and products
Photolysis on soil and plant surfaces
Crop uptake and bioavailability of parent compound and its major degradation products
Residues in rotational crops
Soil dissipation
Residue degradation and disposition in water-sediment systems

The Meeting reviewed the various types of environmental fate studies as related to the process of estimating residues in commodities and concluded that some of the studies do not assist significantly in defining the residue of concern or estimating residue levels. The recommendations of the Meeting for future submissions of data on environmental fate are summarized in the Table below. It should be noted that the studies required are in some cases dependent upon the use pattern (soil, foliar, seed treatment) and that paddy rice presents a unique situation.

The Meeting adopted these revised criteria for its considerations during the 2003 Meeting and will use them in future deliberations. The Meeting proposed to incorporate these changes into the next revision of the FAO Manual.

Requirements for submission of data on environmental fate for the JMPR.

2.12 Pilot project on worksharing

The revised proposal for the FAO/WHO pilot project on worksharing for the JMPR was presented according to the recommendation of the OECD Working Group on Pesticides (WPG) Registration Steering Group (RSG) meeting in March 2003. The proposal was originally presented at the November 2002 RSG meeting with the intention of parallel review by the OECD countries and the JMPR.

The pilot project proposal was revised to facilitate the use of national and other international evaluations of pesticide residues and toxicology by the JMPR. The worksharing process will be used only for new compounds in the pilot phase.

For residue reviews, worksharing will include studies or information on identity, physical and chemical properties, metabolism, environmental fate in soil and water-sediment systems (see Section 2.11 analytical methods (specialized methods used in the analysis of samples and enforcement methods), stability of pesticide residues in stored analytical samples, extraction efficiency of residue analytical methods (using radiolabelled compounds), stability during cold storage (for diverse crop samples, and for animal tissues, milk, and eggs when MRLs are needed for animal commodities) definition of the residue and fate of residues during processing (if appropriate to the crops treated).

For toxicology reviews, all data evaluations that can be released by national governments will be considered in the worksharing process. The process will include comparison of similarities and differences between decisions on selection of endpoints, acceptability of studies, and degree of documentation in the study reviews.

Criteria for worksharing include the availability of the experts involved in the national and international reviews for consultation, and of complete toxicology and residue chemistry study reports.

The 2003 CCPR selected trifloxystrobin as the first compound for the worksharing pilot project, as it has been evaluated in Australia, Canada, the USA and the EC and is scheduled for evaluation by the 2004 JMPR.

The CCPR and the OECD-WGP/RSG will be regularly informed of the progress of the pilot project on worksharing. The project will be evaluated in terms of the harmonization and acceleration of pesticide residue evaluation at national and international levels.

The Meeting acknowledged that at present national assessments are already being extensively used; the worksharing project intends to facilitate and formalize this practice.

2.13 Implementation of the recommendations of the York workshop and the Zoning Report

The York Workshop of September 1999 on "Developing Minimum Data Requirements for Estimating MRLs and Import Tolerances" and the previous preparatory meetings in York in 1998 and 1999 aimed to facilitate international harmonization of data requirements for setting MRLs and import tolerances, which would also provide guidance to the JMPR in estimating maximum residue levels.

Four important areas were identified during the Workshop which required harmonization: (1) criteria for determining the minimum number of trials, (2) extrapolation of data on residues in one crop to support an MRL for a related crop, (3) processing studies, and d) global zoning. Among these only the zoning issue was pursued as it was considered the least harmonized among the issues mentioned.

At its 32nd Session the CCPR decided to refer the recommendations of the York Workshop to the 2002 JMPR. The JMPR had in fact already been using the recommendations of the York meeting whenever possible. However, some of the recommendations of the York report would need further information before their full utilization. For example, the number of trials was to be based on significance in the diet and in trade and on the number of zones in which the crop was grown, but the criteria for significance or insignificance in trade were not elaborated and the zones were not identified. FAO and the OECD undertook to address the impact of climatic zones on pesticide residues and produced the "Zoning Report", which concluded that the impact of climate on the residues of some pesticides applied to certain crops is negligible, and residue data derived from similar use patterns (GAP) and growing conditions may be compared regardless of the geographical location of the trials. In 2002 the JMPR expressed the hope that the York meeting report would be finalized and would be made available for consideration by the JMPR.

The OECD Working Group on Pesticides adopted the Zoning Report and the report of the York Workshop at its meeting in November 2002 (ENV/JM/PEST/2002/15). The European Commission indicated an interest in following up the recommendations of the Zoning Report as it could facilitate the establishment of MRLs.

The Meeting reviewed the Zoning Report and the York Report and agreed to start applying their recommendations. However, practical experience would be necessary to see how the recommendations could be implemented in the work of the JMPR, e.g. a change in the data requirements to allow the global comparison of residue data might be necessary. Hence the JMPR would like to test the practical applicability of the principles with one pesticide in 2004 and requested FAO to initiate the process and to identify a compound suitable for the pilot project.

2.14 Submission of additional data for evaluation of pesticide residues

The data and information required for JMPR evaluations are discussed in detail in the 3rd Chapter of the FAO Manual, with special sections on the reconsideration of previous recommendations (p. 32) and the evaluation of additional information (p. 45). In 2002 the Joint FAO Secretary to the JMPR issued a letter giving specific guidance for data submitters to facilitate the selection and submission of appropriate data.

The Manual emphasized that the submission of additional data and/or information should always be accompanied by a clear statement of the reason for it and for the suggestions or requests for changes.

The Meeting continues to receive supplementary data and information without an indication of the specific purpose for its provision. It is therefore re-emphasized that the submitter must explain clearly why the data or information was submitted, with reference to JMPR or CCPR Reports. This will be a pre-condition for scheduling the evaluation of the submitted material and will be used by the FAO Joint Secretary in presenting the rationale for the evaluation.

The Meeting reconfirmed that the evaluation of the results of additional metabolism studies, and of supervised trials revealing information on the proportions of the parent compound and significant metabolites can only be carried out at the time of a periodic review when all relevant information is available and taken into consideration in deciding on the definition of the residue.

The Meeting draws attention to the fact that trials conducted with analytical methods having lower LOQ values or reflecting the residues derived from new GAP, together with corresponding residue data, can be used for decreasing an existing MRL only under specific conditions. For instance if the critical GAP on which the current JMPR recommendation was based had been changed, or sufficient residue data reflecting the uses according to the current GAP in countries where the original trials were conducted are available with the more sensitive new method. Otherwise, the manufacturer or company supporting the use of the compound should keep the reports of these trials until the periodic review of the compound and then provide all relevant information for evaluation, regardless of whether it had been submitted earlier or not.

When the intention is to change a CXL the request should be addressed to the CCPR; other matters should be addressed to the FAO Joint Secretary to the JMPR.