Cover Image
close this bookFAO manual on the submission and evaluation of pesticide residues data for the estimation of maximum residue levels in food and feed (1997)
View the document(introduction...)
View the document6.1 Guidelines for predicting dietary intake
View the document6.2 Estimation of supervised trial median values

6.2 Estimation of supervised trial median values

As a consequence of the recommendations made by the Joint FAO/WHO Consultation for Predicting Dietary Intake of Pesticide Residues an informal Workshop was convened in the Hague, Netherlands, in April 1996 to consider the effect of the recommendations on the work of JMPR and especially of those related to the reviews of residue data undertaken by the FAO Panel members. At this Workshop, guidelines were developed in order to give guidance on the implementation of the York Consultation recommendations to the FAO Panel reviewers, and to convert the recommendations into practical methods for evaluating data.

The Workshop focused on the reviews of data undertaken by FAO Panel members and the estimation of supervised trials median residue (STMR) levels. Several general recommendations and 27 specific recommendations for the evaluation of data were made.


Residues data from countries are evaluated against the GAP in the country of the trials or a neighbouring country with similar climate and cultural practices.

In identifying the STMR, the trials values selected should be comparable with the maximum registered use (i.e. maximum application rate, maximum number of treatments, minimum PHI) on which the MRL is based.

In establishing comparability of uses in the residue trials to the maximum registered use, the application rates in the trials should generally deviate no more than ±25-30% of the maximum application rate. Deviations from this should be explained in the appraisal. Similarly, maximum deviations of no more than ±25-30% should also be used as a guide for establishing comparability of PHI; however, in this case the latitude of acceptable PHIs will also depend on the rate of decline of residues of the compound under evaluation. Consideration of whether the number of treatments reported in trials is comparable to the registered maximum number of treatments will depend on the persistence of the compound and the interval between applications. Nevertheless, when a large number of treatments are made in the trials (more than 5 or 6), the residue level should be considered very little influenced by further treatments unless the compound is persistent or the treatments are made with unusually short intervals.

In establishing comparability of residue trials data in which more than one parameter (i.e. application rate, number of treatments or PHI) deviate from the maximum registered use, consideration should be given to the combination effect on the residue value which may lead to an underestimation or overestimation of the STMR. For example, a trial result should not normally be selected for the estimation of the STMR if both the application rate is lower (perhaps 0.75 kg/ha in the trial; 1 kg ai/ha GAP) than the maximum rate registered and the PHI is longer (perhaps 18 days in the trial, 14 days GAP) than the minimum registered PHI, since these parameters would combine to underestimate the residue. When results are selected for the estimation of STMRs, despite combination effects, the reasons should be explained in the appraisal.

If the residue value arising from a use considered comparable with the maximum registered use is lower than another residue value from the same trial which is within GAP, then the higher residue value should be selected in identifying the STMR. For example, if the GAP specified a minimum PHI of 21 days and the residue levels in a trial reflecting GAP were 0.7, 0.6 and 0.9 mg/kg at 21, 28 and 35 days respectively, then the residue value of 0.9 mg/kg would be selected.

Trials with more than one residue value

In identifying the STMR only one data point should be taken from each trial (i.e. in this aspect, treatments carried out within a study at the same location with different formulations, or different dosage rates or with the same dosage but at a different time, are considered to be different trials).

Where several residue values have been reported from replicate plots from a single trial (i.e. same dosage at one site location), the highest residue should be selected for the purpose of identifying the STMR.

Where several residue values have been reported from replicate analyses of the same field sample taken from a single trial (i.e. site location), the mean residue should be selected for the purpose of identifying the STMR.

Rounding of results

In identifying the STMR from a residue trial, the actual residue value should be used in the estimation of dietary intake without rounding up or down. This would even be the case where the actual results were below the practical limit of determination considered appropriate for enforcement purposes. Rounding of residue values is inappropriate since the STMRs are used at an intermediate stage in the dietary intake calculation.

Residue definition

The WHO Panel considers and indicates in its evaluations which metabolites should be included in the dietary risk assessment.

If it is recommended that the residue definition for the risk assessment be different from that for enforcement, this is clearly stated in the appraisal.

Close communication should be established prior to the JMPR meeting between the FAO Panel reviewers and the respective reviewers on the Toxicological and Environmental Groups, on questions such as which metabolites are of toxicological significance.

In tabulating the residue trials data the FAO Panel reviewer should indicate the levels of relevant metabolites separately from those of the parent compound, but in a way which would allow subsequent combination, in order to ensure that changes in the residue definition can be accommodated at the JMPR meeting.

In those cases where it is not possible to finalise the risk assessment at the JMPR (September, year 1), usually because of a change in residue definition, the MRLs would still be recommended to the CCPR (by way of Codex circular letter for comment at step 3) and the compound would be re-discussed at the following year's JMPR meeting (September, year 2). The recommended MRLs together with the conclusion of the risk assessment would be available for the next CCPR (April, year 3).

If two compounds for which STMRs can be calculated produce the same analyte in compliance monitoring (e.g. CS2 for dithiocarbamates), it is possible to separate the intake assessments, if required, because the intake assessment is no longer based on the MRL but is based on residue data specific to the individual compounds.

Combining of populations of data for the calculation of STMRs

In identifying the STMR, residue data reflecting different countries' GAPs would normally be combined. However, if the trials data reflecting different countries' GAPs appear to give rise to different populations of data then these data sets should not be combined. In these cases the STMR should be calculated from the population(s) of data which is (are) driving the MRL. In deciding whether the results of trials reflecting different countries GAPs give rise to different populations of residues data, the size of the database reflecting the different countries' GAPs should be taken into account.

Residues below the limit of determination

As a general rule, where all residue trials data are <LOD, the STMR would be assumed to be at the LOD, unless there is scientific evidence that residues are "essentially zero". Such supporting evidence would include residues from related trials at shorter PHIs, exaggerated, but related, application rates or a greater number of applications, expectations from metabolism studies or data from related commodities.

Where there are two or more sets of trials with different LODs, and no determinable residues have been reported in the trials, the lowest LOD should normally be used for the purpose of STMR selection (unless the residues can be assumed to be essentially zero as given above). The size of the trials database supporting the lowest LOD value should be taken into account in the decision.

Processing, cooking factors and edible portion residue data

In using data on the effects on residue levels of processing or cooking practices, the mean processing factor should be applied to the STMR estimated for the raw agricultural commodity as already described. The STMR value estimated in this way for the processed commodity should be referred to as the STMR-P.

If data are available for the residues in the edible portion of the commodity (e.g. banana pulp), a STMR should be estimated directly using the edible portion residue values from maximum registered use trials (as opposed to using pesticide values for the whole commodity).

Acute dietary intake2

2 The estimation of acute dietary intake and MRLs for products of animal origin are currently under reconsideration by FAO, WHO and the JMPR.

For the purpose of acute risk assessment the MRL, or the highest residue in the edible portion, should be used in estimating dietary intake.

Estimation of MRLs for products of animal origin2

2 The estimation of acute dietary intake and MRLs for products of animal origin are currently under reconsideration by FAO, WHO and the JMPR.

In estimating MRLs for products of animal origin, theoretical feed intakes for domestic animals should be calculated using the STMR for each feed item (derived from supervised trials comparable with the maximum registered use), rather than the MRL, together with the maximum feed incorporation rates. This is in conformity with past JMPR decisions.

Estimation of STMRs for commodity groups

Where there are adequate trials data the STMRs should, in principle, be identified for the individual commodities and these values used for the intake assessment. However, where the MRL has been established for a group of commodities (e.g. pome fruit) a single STMR should be calculated for the group of commodities.

Presentation of STMRs in the JMPR monographs and report

The GAP(s) on which trials data have been selected for the purpose of identifying the STMR should be clearly identified in the monographs.

In tabulating trials data in the monographs the reviewer should ensure that, in addition to the normal underlining of trials data that are within GAP (and therefore have been used for the MRL evaluation), the single residue values selected for the estimation of the STMR are double- underlined.

Information on the residue values on which the STMR is based should not only be identified in the tabulated trials data (see above) but a list of the residue values selected should be included in the appraisal, in numerical order, with the median residue underlined. Where the residue situation is complex (e.g. a number of metabolites to be considered) these data may best be tabulated in the appraisal. In addition, the STMR values should be included in the recommendation table in the appraisal and in Annex 1 of the report.

The range for the rates and PHIs used in the selection of residue values for STMR should be clearly identified in the appraisal (e.g. trials data with application rates from 1.8 - 3.0 kg ai/ha have been selected).

Methods for presenting estimated STMR levels are still being developed. The aim is to communicate the results as clearly and unambiguously as possible; experience may indicate that further changes are necessary. As pesticides are used in a wide variety of situations, methods for evaluating data must be developed to take into account cases that are not already covered by the suggested procedures. For instance, the recommendations on the estimation of STMRs in animal commodities arising from residues in feed, acute dietary intake and ectoparasite treatments of farm animals may require further development. Further guidance will be developed at future meetings of JMPR.