2. General considerations

2.1 Work sharing

General considerations regarding work sharing

The Meeting discussed the advantages and limitations of work sharing in the context of the FAO/WHO/Organization for Economic Co-operation and Development (OECD) pilot project on work sharing conducted in 2004 with the pilot study on trifloxystrobin and the work-sharing project for 2006 on quinoxyfen, as proposed by the CCPR. The purpose of this project was to test whether national and regional evaluations of toxicology and pesticide residues could be used as a basis for the JMPR evaluations.

Overall, work sharing, as understood by the JMPR, should represent an independent expert peer review of critical data and existing evaluations; where possible, appropriate text from existing national/regional evaluations should be used by the JMPR experts. The Meeting noted that work sharing can be useful and save time. However, there are clear limitations to work sharing, as noted in the 2004 JMPR report. The Meeting emphasized that it is critical that JMPR continues to perform an independent evaluation and expert review of the evaluation that ensures consistency, and results in an international consensus evaluation. In this context, the JMPR monographs can be described in three parts: (1) the description of actual studies; (2) the interpretation and evaluation of the studies; and (3) the final evaluation/appraisal of the compound. Part 1 is most accessible to work sharing, provided that there is sufficient harmonization between monograph formats used by different authorities. By using study descriptions and data tables from existing evaluations, the JMPR expert may be able to save time in the preparation of the JMPR monograph. Part 2 could be taken directly or modified or rewritten from existing national/regional evaluations after a review by the JMPR experts. Part 3 should represent an independent JMPR evaluation and review.

Considerations regarding toxicological evaluations

Regarding the need to develop a JMPR toxicological monograph for compounds for which agreement exists in the evaluations made by several authorities, the Meeting pointed out that JMPR toxicological monographs are rather unique in their complete, detailed, and transparent evaluations, and are more readily accessible than other monographs. This had also been stated at the recent OECD workshop on work sharing.

For the proposed work-sharing project for 2006, a different approach from that used in the 2004 pilot project was proposed for the toxicological evaluation. Instead of using one national/regional evaluation as a "template" for the JMPR toxicological monograph and comparing this with other evaluations, the Meeting recommended that the Temporary Adviser should judge which relevant parts of different national/regional evaluations should be used in preparing the JMPR toxicological monograph.

In the preparation of the present Meeting, the United States Environmental Protection Agency (EPA) made available to the toxicology experts tables of four of the five scheduled new compounds (fenhexamid, novaluron, dimethenamide-P, indoxacarb), summarizing the toxicological end-points/studies and uncertainty factors used in the evaluations made by several national/regional authorities. The WHO Core Assessment Group generally agreed that such comparison tables are a useful resource in both the preparation of the working papers for the Meeting and in discussions during the meeting, mainly for identifying differences in the selection of end-points. It was also agreed that such a comparison table would be useful in the preparation of working papers for the 2006 JMPR. The Meeting recommended, however, that less detail was needed and that a summary of end-points, critical studies, no-observed-adverse-effect levels (NOAELs) and lowest-observed-adverse-effect levels (LOAELs), and uncertainty factors would be sufficient. Also, it was recommended that national/regional assessments should be clearly identified as being in a draft or final form.

Criteria for a work-sharing project for residue evaluations

Experience gained during the evaluation of trifloxystrobin by the 2004 JMPR indicated the need to select criteria for the acceptance of a compound to be evaluated via work sharing, in order to increase efficiency and thus reduce the workload of the JMPR and support the acceleration of the process by which MRLs are accepted.

The following criteria for the selection of a compound to be evaluated using the work-sharing process relevant to residue evaluations were recommended:

• The compound must be validated at national, regional and international levels, covering all aspects of the residue evaluation (including data from supervised trials).

• Summaries of data validated at national, regional and international levels, covering all aspects of the residue evaluation (including data from supervised trials) must be available.

• The data should be available in a standard format, harmonized at the international level (OECD, FAO). This would allow exchange of a valid database, thus saving time and potentially reducing the workload.

• In the national/regional documents, factual information should be separated from interpretations.

• The definition of the residue should preferably be identical in the different national and regional evaluations.

Conclusions

The main criterion for the selection of a new pesticide to be evaluated via work sharing (toxicological and residue evaluations) is that it has been reviewed by at least three national/regional agencies. In the event that the findings are similar, relevant parts of national/regional reviews should be used in the preparation of JMPR documents. An independent appraisal should be prepared and approved by JMPR that represents international consensus.

2.2 Development of OECD Test Guidelines and Guidance Documents for Pesticide Residue Chemistry

The Meeting was informed that OECD has developed five test guidelines and two guidance documents on residue chemistry based on guidelines currently used in Australia, Canada, Japan, European Union (EU), FAO, and the United States of America (USA). The documents primarily provide guidance for the generation of data but are also useful for assessing these data. Guidelines, once accepted, are mandatory among the OECD Member States, while guidance documents are not. The documents are currently out for comments by OECD Member States and other stakeholders. An OECD ad hoc Expert Group was responsible for drafting the guidelines and guidance documents. The following countries and organizations are represented in the Expert Group: Australia, Canada, Germany, Italy, Japan, Netherlands, the United Kingdom (UK), USA, the European Commission, FAO and CropLife International/BIAC (industry). The Expert Group is chaired by the USA.

Up to this point, OECD Harmonized Residue Guidelines were developed for: 1) Metabolism in Crops; 2) Metabolism in Livestock; 3) Residues in Livestock; 4) Metabolism in Rotational Crops; and 5) Residues in Rotational Crops (Limited Field Studies). The two guidance documents developed are: (1) Overview for Residue Chemistry Studies (including Livestock Feedstuff tables, tables of Raw Agricultural Commodities, and Glossary of Terms); and (2) Definition of the Residue. In the near future, three additional guidelines will be drafted, on (1) storage stability; (2) Processing studies - nature of the residue; (3) Processing studies - Magnitude of the residue. A guidance document will be developed on analytical methods. In the final phase of the project, a guideline on crop field trials will be drafted. In parallel, templates for summarizing data contained in a study report are being developed.

The Meeting agreed that harmonization of guidelines for determination of pesticide residues provides a foundation for work sharing of residue chemistry reviews among countries and that harmonization will also lead to mutual acceptance of regulatory results thus minimizing trade barriers. Therefore the Meeting welcomed the development of the above guidelines and guidance documents and looks forward to further developments.

2.3 Statistical approach to MRL estimation

The Meeting considered a statistically-based procedure used routinely in the North American Free Trade Agreement (NAFTA) countries for the estimation of maximum residue limits (tolerances). The supervised field trial residue data resulting from the use of a specific pesticide on a specific commodity (or commodity group) are entered in any order into a spreadsheet which then tests the data set for fit to a log normal distribution. If the data pass the log normality test, as evidenced both by a mathematical calculation and a probability plot, the 99^th percentile value and the 95% upper confidence limit of the 95^th percentile value are calculated (LN99, LN95).

For a log normal distribution where the number of residue data points are small (n < 15), an alternate calculation of 3.9 times the upper confidence interval of the median is made (UPLMed95). For the situation of the mean equal to the standard deviation, this represents the 95^th percentile.

Finally, if the data set does not appear to be log normal, the mean plus three standard deviations is determined (California method), which represents at least 89% of data for a distribution-free situation.

A visual plot of the log normal test and the various possible values are provided. The scientist is thus presented with a series of suggested values from which he/she may select or reject, based on expert judgment. The decision-tree sequence is illustrated in Figure 1. Note that the MLE (maximum likelihood estimate), for values below the limit of quantification (LOQ) and the rounding function are options that may be toggled on or off.

Figure 1: Options for estimation of MRL from field trial data sets

The Meeting used the spreadsheet to obtain maximum residue level estimates for several of the compounds considered. The findings were compared with the estimates made independently by the evaluators. The comparison is summarized in Table 1. Generally, the recommendations of the spreadsheet compare favourably with those determined by the evaluator.

Table 1: Comparison of some MRLs as determined by JMPR and by spreadsheet

¹ Rejected spreadsheet finding of a log normal situation (29 mg/kg).
² Rejected spreadsheet finding of a log normal situation (72 mg/kg).
³ Rejected spreadsheet finding of a log normal situation (96 mg/kg).
⁴ Rejected spreadsheet finding of a log normal situation (3.59 mg/kg).
⁵ Rejected spreadsheet finding of a log normal situation (97 mg/kg).

Estimation of the maximum residue level for glyphosate on soya beans is an example of how the spreadsheet may assist the evaluator. The data set contains 36 residue values ranging from 0.27 to 17 mg/kg with a median value of 1.85 mg/kg and an average of 2.76 mg/kg. The evaluator would most likely consider 20 or 30 mg/kg, based on the JMPR rounding system. The spreadsheet calculations (Figure 2) indicate that the data appear to be log normally distributed, as indicated from both the Shapiro-Francia Normality Test and a visual inspection of the probability plot. The 99^th percentile value is 19.27 mg/kg and the 95% upper confidence limit of the 95^th percentile is 16.11mg/kg. Under JMPR rounding procedures of these estimates, the appropriate estimate would be 20 mg/kg. The spreadsheet proves useful in helping the evaluator decide between the 20 mg/kg and 30 mg/kg values and provides a good statistical rationale for the choice.

Lognormal probability Plot

(Values in parentheses are confidence limits on the 95^th or 99^th percentiles)

Figure 2: Log normality test and possible MRLs for glyphosate on soya bean.

However, it is also evident from the worked examples that the evaluator must not simply default to the output of the spreadsheet but must always use scientific judgment. Particular caution is required with small data sets. Consider the supervised field trial results for fenhexamid on lettuce. There are only 8 sample points, with a median of 11.5 mg/kg, an average of 10.4 mg/kg, and a maximum of 19 mg/kg. The Meeting agreed that 30 mg/kg was a reasonable estimate of the maximum residue level. The spreadsheet determined that the distribution was log normal, with a 99^th percentile value of 72.17, or a maximum residue level estimate of 75 mg/kg. This value is unreasonably high when compared with the maximum value, in fact, more than three times the maximum. Also, the finding of log normality seems not so evident upon inspection of the probability plot (Figure 3). If log normality is rejected, the nonparametric estimate is 29.78 mg/kg, in agreement with the 30 mg/kg Meeting estimate.

Lognormal probability Plot

(Values in parentheses are confidence limits on the 95^th or 99^th percentiles)

Figure 3: Log normality test and possible MRLs for fenhexamid on lettuce.

The Meeting agreed that the statistical spreadsheet is a useful tool that can assist the evaluator in deriving an appropriate maximum residue level estimate from the supervised field trial data. It was also recognized that the evaluator should not rely upon the spreadsheet to provide the proper value, but he/she must exercise good scientific judgment in selecting among the optional outputs or in proposing a value not provided by the spreadsheet. The greatest attribute of the new procedure is in providing a statistical analysis of the data and thereby a more defensible basis for the estimate. The Meeting noted that further development of the procedure may be necessary for small data sets.

The Meeting also considered that the widespread use of a standardized estimation procedure has the potential to advance consistency in the estimation of MRLs at the national and international levels.

The Meeting concluded that (1) evaluators may use the spreadsheet in the evaluation of verified field trial residue data as an aid in the estimation of the maximum residue level; and that (2) the procedure will be included in the next update of the FAO Manual.

2.4 Crop classification and harmonization

The Meeting noted the two well established crop classification systems: the Codex Crop Classification and the United States Crop Grouping Scheme. Both systems originated from the same source, the work of Dr R. Duggan, USDA, and were developed in synchronic timelines. They are fairly similar (approximately 70-80%) on the commodities identified.

Besides the similarities, some major differences also exist between the two systems and these are noted in the following table:

Given the above differences, it was recognized that the two classification systems were originally developed with different focuses. The Codex system was for MRL settings from agricultural and trade perspectives while the US system was for residue extrapolations from the point of view of residue exposure and pesticide use patterns.

Currently both systems are undergoing revisions led by the CCPR Delegation of the Netherlands for Codex, and IR-4 and US EPA for the USA. The US revision is a joint effort of the International Crop Grouping Consulting Committee, which is formed by representatives from Asia, Australia, EU, Middle East, and NAFTA countries, including representatives of the Netherlands, the Codex revision leading group.

The Meeting supports the collaboration initiatives being made by the two workgroups to bring the strengths of the two systems together in a harmonized classification system. It was recognized that such a system would facilitate the work of JMPR and CCPR, and would benefit participating countries in residue research, risk assessment, and MRL setting.

2.5 International speciality Crop Foundation initIative for Minor Use

Obstacles to the registration of pesticides for minor uses and specialty crops are observed in many countries, as the development of pesticide uses for these purposes is not economically attractive to manufacturers. Specialty crops are traded regionally and internationally and are often of high economic value; however, the establishment of MRLs for these crops remains scarce due to the lack of available residue data and/or registered uses.

The Meeting noted the activities of the USDA/IR 4-Project within NAFTA and its recent collaborations with EU member states and Asian countries. IR-4 is a special minor use programme funded by the US-government and the private sector. This programme has developed residue data to support registration of agrichemicals on specialty crops for over 40 years with over 8,300 food use clearances and 10,600 ornamental use clearances. Since 2001, IR-4's tolerance petitions comprised 50% of the total numbers approved by the US EPA (IR-4 website: http://ir4.rutgers.edu).

The Meeting was informed that IR-4 is willing to share its data and experiences with other countries, and is willing to support an international collaboration, proposed as the International Specialty Crop Foundation, to be housed and led by IR-4. This Foundation should be funded by various sources, i.e. governments, pesticide industry, growers associations, importers/exporters and others, and hopefully will have the participation of international organizations such as FAO and OECD.

Such an initiative would enhance the availability of residue data for minor uses and specialty crops being submitted to national registration authorities for their approval and subsequently to the JMPR. It would maximize efforts and save resources of participating countries, and would further strengthen the process of establishing international MRLs, benefiting both industrialized and developing countries.

The essential elements in this process are:

• Identifying needs and establishing a list of tasks to be included in the Foundation's programme;

• Elaborating harmonized protocols for residue trials to produce data for the registration of pesticides on minor or specialty crops;

• Organizing training courses for representatives of developing countries intending to participate in the programme, to facilitate good agricultural practice (GAP) trials resulting in acceptable residues.

The Meeting encourages governments and international organizations, such as FAO, to support and participate in this initiative.

2.6 Estimation of long-term intakes of pesticides in/on Dried chili peppers and short-term intake of mevinphos on spices

The 2004 JMPR estimated maximum residue levels for spices based on monitoring data and for dried chili peppers taking into account the MRLs on fresh peppers. The 37^th Session of CCPR advanced the proposed MRLs for dried chili peppers to Step 5 (Annex 1) and requested the JMPR to perform an overall long-term intake assessment for each compound where the TMDI for dried chili peppers was > 5% of the ADI (acephate, carbaryl, dicofol, dimethoate, methamidophos, oxamyl, profenofos, and vinclozolin).

The Committee did not advance the recommended limits for mevinphos for spices to step 8 with omission of step 6 and 7 due to acute intake concern.

Estimation of long-term intake for pesticides for which MRLs were recommended for dried chili peppers

The 2004 JMPR calculated TMDIs only for dried chili peppers, without considering the other uses of a compound. The long term intakes were reported without taking into account the assumption that 10% of fresh peppers are consumed as dry chili pepper. Therefore the calculations of IEDI and TMDI were carried out by the 2005 JMPR for all uses of the compounds including dried chili peppers and spices, where relevant, based on the most recent evaluation by JMPR. In general, IEDIs were calculated for compounds evaluated from 1999 to 2004 where STMRs were available. For compounds evaluated before 1999, only the TMDI could be calculated. In all cases, the contribution to long-term intakes from the use of the compound on dried chili peppers was also calculated.

Dried chili pepper consumption was estimated to be approximately 10% of the total consumption of fresh peppers and chili peppers, i.e. VO 0051, VO 0444, and VO 0445. Given that the water content of fresh peppers is about 90% and that 10% of the consumption of sweet peppers is derived from dried chili peppers, then dried chili pepper consumption can be estimated as approximately 1% of fresh peppers, based on dry weight vs fresh weight. The residues were derived by extrapolation from the MRL or STMR (where available) for fresh peppers by applying a concentration factor of 10.

Long-term intake estimates (total IEDIs) for compounds that initially showed a TMDI > 5% for chili peppers are summarized in Table 2.

Acephate, carbaryl, methamidophos, oxamyl, and profenofos

The IEDI for each of the five GEMS/Food regional diets were in the range of 2-70% of the ADI for acephate, carbaryl, methamidophos, and oxamyl when considering all of the uses of each compound, including the proposed use on dried chili peppers. The TMDIs for all regional diets were within the ADI for profenofos. Therefore, the long-term intake of residues of these five compounds from all uses that have been considered is unlikely to present a public health concern.

Dimethoate

The IEDI for the European diet was already about 120% of the ADI for dimethoate even without the addition of dried chili peppers. Based on the STMR estimated by the JMPR, the dry chili pepper contribution, to the ADI is 3 to 14%, resulting in an intake of 130% of the ADI in the European diet.

Table 2: Intake estimates (total IEDI) for compounds with TMDI > 5% of ADI

Dicofol

TMDI calculations (last evaluated in 1994) showed that even without dried chili peppers, intakes for all the five regional diets greatly exceeded the ADI (100-690% of ADI). In fact, the contribution of dried chili peppers was only from 2% to 9% of the ADI for the five regional diets.

Vinclozolin

TMDI for the European diet was 120% of the ADI. The dried chili peppers' contribution to the intakes of at the proposed MRL of 30 mg/kg ranged from 1% to 5%, the highest being for the European diet.

Other compounds

The IEDI or mixed calculation based on the available STMR and MRL values (Table 3) indicated that the contribution of the residues of other compounds in/on dried chili peppers (carbendazim, chlorpyrifos, diazinon, dinocap, ethephon, ethoprophos, imidacloprid, malathion, methomyl, methoxyfenozide, piperonyl butoxide, pyrethrins, spinosad. tebufenozide, tolyfluanid) to the ADI was < 1%.

Table 3: IEDI calculations for other compounds based on latest JMPR review, incorporating proposed MRLs for dried chili peppers

¹ Diazinon - This was a mixed assessment, using STMRs where available and MRLs in most cases.

² Ethoprophos - The 2004 JMPR recommended to withdraw CXL for sweet pepper of 0.02 mg/kg and replaced it with an MRL of 0.05 mg/kg. The 2005 CCPR agreed and advanced the MRL of 0.05 mg/kg to Step 5/8. Therefore, for dried chili peppers, the MRL is 0.5 mg/kg. The STMR-P for dried chili peppers was used in the calculation.

³ Imidacloprid - The 2005 CCPR session decided to change the MRL for imidacloprid on dried chili pepper to 10 mg/kg, as the MRL for peppers of 1 mg/kg was based on a fresh weight basis (para 185, ALINORM 05/28/24). The STMR extrapolated from fresh pepper was used in the calculation.

The TMDI calculations for all uses based on the current MRLs are summarized in Table 4.

Table 4: TMDI calculations for other compounds, based on current MRLs, incorporating proposed MRLs for dried chili peppers

¹ Dithiocarbamates - Based mainly on maneb, mancozeb, zineb, and metiram, which have a group ADI of 0.03 mg/kg bw/day.

TMDI calculations for all compounds in Table 4, except for chlorpyrifos-methyl, were all within the ADI for all the regional diets. Therefore, long-term intake of residues of these compounds from all uses including those for dried chili peppers is unlikely to present a public health concern.

Chlorpyrifos-methyl

TMDI calculations resulted in the ADI being slightly exceeded for the Middle-Eastern diet (120%), even without dried chili peppers. Dried chili peppers had no contribution to the intakes of any of the five regional diets.

International estimate of short-term dietary intake (IESTI)

The JMPR calculated the short-term intake based on the FAO/WHO consumption data taking into account the 1.75 and 1.67 g/kg bw/day consumption reported for adults and children in France, respectively. Some countries have lower consumption figures for spices: e.g. Australia 0.299 g/kg bw/day for adults; UK 0.1118 g/kg bw/day for adults, USA 97.5^th percentile consumption (0.045 g/kg bw/day for adults and 0.087 g/kg bw/day for children 1-6 years); Germany 15.4 g/child/day (0.93 g/kg bw/day); These consumption figures indicate that the short-term intake would not exceed the ARfD in the latter countries.

IESTI for Mevinphos on Spices:

The short-term intake can only be defined at the national level.

2.7 Consideration of alternative GAPs

The CCPR, at its 37^th Session in 2005 (ALINORM 05/28/24, paragraphs 105, 107, 108, 135 and 243) drew attention to the acute intake concerns for proposals for disulfoton, fenamiphos and aldicarb.

1/ CCPR. 2002. 34th Session. 5. Dietary exposure in relation to MRL setting. (A) acute dietary risk assessment. Prepared by WHO. Document CX/PR 02/03, March 2002

2/ IESTI for sweet peppers.

In evaluating supervised trials data to support an MRL, JMPR identifies the maximum GAP in each country, and evaluates trials data from that country and others with similar climate and cultural practices.

JMPR recommends a maximum residue level that relates to the highest residues from a national GAP where there are sufficient supervised trials data.

The CCPR has requested an extension to that procedure, which would become:

JMPR recommends a maximum residue level that relates to the highest residues from a national GAP where there are sufficient supervised trials data and where the residues do not result in an IESTI (international estimate of short-term dietary intake) that exceeds the acute reference dose.

The Meeting agreed that this would be a suitable procedure, because Codex MRLs are used as standards for food in trade, not for enforcement of GAP as in a national registration system where the MRL must be associated with the maximum registered use.

Aldicarb was evaluated by the JMPR in 2001, disulfoton in 1994 and 1998 and fenamiphos in 1999. Some of the registered uses recorded in those evaluations have quite likely been modified since then. The problems of basing an evaluation on obsolete GAP should be noted.

The Meeting sought advice from CCPR on the best way to proceed.

Retrospective approach

CCPR would refer the compound to JMPR and request reconsideration of GAP in specific cases. It would also request the manufacturer(s) and national governments to provide appropriate up-to-date GAP information to support the proposed evaluation.

The advantage of this approach is that reconsideration of GAP would be only at the request of CCPR for specific cases where all other avenues for refinement had been exhausted. The disadvantage is that, because of the elapsed time, the reported GAP may well have changed since the time of the evaluation. A new data submission would be needed.

Prospective approach

During a residue evaluation where the IESTI is exceeded, JMPR should draw attention to available information on alternative GAPs and associated supervised trials data where the IESTI would not appear to be exceeded.

The advantage of this approach is that no time is lost in referral between CCPR and JMPR and the adoption of a Codex MRL would likely occur more quickly. The disadvantage is that the acceptance of alternative GAP would likely become the first action instead of the final resort when other avenues of refinement are exhausted.