Summary and highlights from recent EFSA publication ‘Recurring issues in ecotoxicology’

Published on


Summary and highlights from recent EFSA publication ‘Recurring issues in ecotoxicology’

On 8th July 2019, EFSA published a technical report containing the outcome of discussions within the pesticide peer review meeting of general recurring issues in ecotoxicology that have been observed by EFSA and MS regulatory authorities during active substance and zonal product evaluations for plant protection products (PPPs).

This 117-page report contains discussion of general and specific issues relating to the risk assessment for birds and mammals, aquatic organisms, non-target arthropods, soil organisms and non-target terrestrial plants. In this article we have highlighted some of the key decisions and recommendations made within the technical report, but for full details reference should be made to the EFSA publication.

General issues

Extrapolation of studies between different agro-climatic conditions

For aquatic mesocosms, experts agreed that an ETO-RAC can be used to extrapolate between different regions. Extrapolation of an ERO-RAC would be considered on a case-by-case basis. However, regulatory acceptance of an ERO-RAC, even within a region, is currently uncertain and further discussion is needed to address these uncertainties.

For extrapolation of ecological refinements used in bird and mammal risk assessments (e.g. focal species, PT, PD), it needs to be demonstrated that the refinements are relevant for each proposed region.  The recommendations in EFSA (2009) for spray applications are considered to be sufficient on this point. Further attention is required for seed treatment uses to take into account specific agronomic practices (e.g. sowing rate) and conditions. It was suggested that this can considered in the context of the revision of the EFSA (2009) bird and mammal guidance.

How to consider studies when the analytical methods are not validated

Where analytical methods are not considered to be valid, a case-by-case assessment of the study should be made e.g. toxicological profile, margin of safety in the risk assessment, whether the study was on a vertebrate species, etc. The implication here appears to be that a study won’t be rejected if it can be demonstrated that it is not critical for the outcome of the risk assessment.

Risk assessment for PPPs within the evaluation of the active substance

A formulated product may be considered more toxic than the technical active substance if it is >3x more toxic in a comparable study. If the product is more toxic (i.e. >3x more toxic than the a.s.) then the lower endpoint should be used in the risk assessment, with an option to also present the risk assessment using the a.s. endpoint.

For formulated products containing multiple actives, analytical measurements should be taken for all of the a.s. within the product unless it is clear which drives the toxicity. A proposal for how to express endpoints from formulation studies is included in Appendix J (p.96) of the technical report.

Use of residue data to support ecotoxicological assessments

A summary of the types of data available in the residue dossier section that might be useful for ecotox risk assessments is available in Appendix B (p. 41) of the technical report e.g. identify plant metabolites. A template to collate such residue data potentially relevant for the ecotox risk assessment has been developed (Appendix C (p. 48) of the technical report).

Equivalence of batches

If the test item used in a study had a <90% purity, and nominal concentrations were used to express endpoints, the endpoints should be corrected for the purity of the test item. An overview of the batches used in the ecotox studies should be included in the dossier.

Use of lower limit (EC10, HC5) as endpoint in the risk assessment

The calculation and reliability assessment of EC10 values is described in Appendix E (p.54) of the technical report. This section highlights under which circumstances the EC10 can be used instead of the NOEC and explains the statistical background. The EC10 median should be always reported including its normalised width of the confidence interval (NW) and only used for the risk assessment if it does not fall into the confidence interval (CI) of the EC20 or EC50. If the median EC10 is within the CI of the EC20 or EC50 the 95% lower limit of the EC10 has to be stated. We are able to help you with your ECx calculations.

Risk assessment for rice paddies

The Southern Zone (IT) presented a draft guidance for the assessment of risks for proposed uses in rice paddies. No details regarding the contents of the draft guidance are provided in the technical report, but communication with the Southern Zone task force indicates that there is a plan to submit the draft guidance to the EU commission in late September, in order to obtain feedback from MSs, EFSA and, if agreed with by the Commission, industry, with a view to vote at a future Standing Committee.

Risk assessment for bananas

It was agreed that the ‘orchard’ scenario can be used as a surrogate for banana uses in a Tier I bird and mammal risk assessment. Higher tier risk assessments reliant on ecological refinements would require the identification of focal species etc that are relevant to banana-growing areas.

Birds and mammals

Trials for residue decline

The criteria used to assess residue trials in terms of reliability, extrapolation, dissipation kinetics, etc were discussed and agreed, being presented in detail in Appendix F (p. 64) of the technical report.

Experts broadly agreed with extrapolation of residue data within groupings proposed by EFSA e.g. dicot plants, monocot plants, fruits, etc. Extrapolations between groupings was generally not considered appropriate.

It was agreed that dissipation kinetics should follow FOCUS (2006) and use the single first-order (SFO) model, though the use of the first-order multi-compartment (FOMC) model would also be possible in some instances. Guidance is provided in the technical report regarding the number of time points and trials required for model fitting, how to interpret values below the LOQ/LOD, dealing with outliers, dealing with toxic metabolites, etc.

The difficulties in defining residue decline for invertebrates was discussed. In cases where the DT50 is not considered reliable, it may be possible to calculate a Ftwa by using the area under the curve (AUC) and the averaging period (21 days).

The required number of trials and sites required for invertebrates was highlighted as an issue to be dealt with in the revision of the EFSA bird and mammal guidance (2009).

Several worked examples are presented in the technical report, illustrating how to interpret different residue data sets.

21-d PT

The use of a 21-d time-weighted average value for PT (published by Ludwigs et al 2017) was discussed. This approach has been evaluated in detail by the UK (available by request from HSE CRD) and though the potential for such an approach is recognised, it was agreed that further consideration was needed before it could be used in the risk assessment.  It was recommended that this point be considered in the ongoing revision of the EFSA (2009) bird and mammal guidance.

Aquatic organisms

Use of geomean and weight of evidence for acute data

It was agreed that when the RACgeomean > lowest endpoint, the lowest endpoint should be used to calculate the RAClowest. The modified acute AF for deriving the RAClowest should be ≥20 for invertebrates and ≥30 for fish. No agreement was reached for cases where RACgeomean < lowest endpoint in terms of the RAC and AF to use in the risk assessment.

Use of geomean and weight of evidence for chronic data

There was no agreement for using a geometric mean or reduced assessment factor for chronic data. Experts noted that for some Member States the use geometric mean for chronic data is limited to algae and aquatic plants. The experts expressed the wish to reconsider the issue in the next revision of the aquatic guidance. 

General recommendations on mesocosm experiments

It was agreed that the absence or low abundance of vulnerable groups, i.e. EPT, should not necessarily result in the invalidation of the mesocosm. However, their absence should trigger the need for further considerations, e.g. the selection of a higher AF and/or request for further testing to confirm that EPT are not among the most sensitive species. In such assessments, particular consideration should be paid to the mode of action of the active substance.

Although many EPT taxa are mainly prevalent in running waters, there are representatives that occur in sufficient numbers in mesocosms, which allow assessment of EPT taxa in mesocosms. For example, within the CEA mesocosm facility we regularly find larvae of mayfly (Cloeon dipterum, Ephemeroptera) and caddisfly (Phryganidae, Limnephilidae (Trichoptera)). We also find Coenagrionidae (Odonata) and Chaoborus sp. (Diptera), which are not EPT taxa and are considered to be vulnerable due to their life history and sensitivity to certain modes of action. When setting up our mesocosm studies at CEA, we always ensure that these taxa are evenly distributed between replicates before test item application, and that their life histories are taken into account (e.g. seasonality) to ensure reliable data.

Several recommendations for the experimental design of mesocosms were discussed and agreed, in terms of establishment time, recolonisation, emergence, insect instar, replicates, number of samples and sampling times. At CEA, we have already implemented the majority of these recommendations following the publication of the EFSA (2013) guidance and have presented our experiences to date.

To analyse complex mesocosm data of a high number of species, effect classes are included in the current aquatic guidance (EFSA, 2013). It was noted that effect class 2 (slight effects) is currently not well defined in the guidance, but is discussed in Brock et al. (2015) and should be updated in the revision of the current guidance. At CEA, the Brock et al. (2015) paper is already routinely used in our mesocosm data analysis.

Representativeness of mesocosm studies when the risk assessment at lower tiers is triggered by a non-freshwater species

A step-wise approach for interpreting the relevance of mesocosms when triggered by marine species was discussed and agreed. The first step is to check whether there is a species in the mesocosm study that is taxonomically closely related to the sensitive marine species. If present, check whether the taxonomically related species shows a clear response to treatment in the mesocosm study. If no clear response is observed, further analysis/experiments may be requested.

Use of refined exposure studies as Tier 2C

Concerns were raised regarding modified exposure studies e.g. accuracy of exposure profiles, extrapolation of results to the field, sensitivity of life stages used in studies, testing of individuals versus populations, etc. It was agreed that the scheme for assessing Tier 2C should be reconsidered and possibly further developed in the revision of the EFSA (2013) aquatic guidance document.

Alternative test design in Myriophyllum studies

It was agreed that Myriophyllum studies performed to OECD TG 239 (OECD, 2014b) but with an alternative test design (i.e. one shoot per pot per test vessel) should be considered acceptable.

How to express the endpoint for sediment-dwelling organisms when tested in the presence of sediment

It was agreed that endpoints for sediment-dwelling organisms, when tested in the presence of sediment, should be determined using a mass balance calculation. Submission of mass balance calculations as part of the dataset for sediment-dwellers was recommended, particularly in the case of substances which are difficult to test (concentrations poorly maintained in the test system). Worked examples for calculating mass-balance are available in Appendix G (p. 81) and J (p. 96) of the technical report.

Minimum detectable difference (MDD)

The MDD (EFSA 2013; Brock et al. 2015), is considered to be a valid tool to assess the statistical power of a study to detect treatment-related direct effects, and should preferably be reported on non-aggregated data for the relevant taxon and time points. It was concluded that the use of the MDD is supported and that further considerations and clarifications will be addressed in the revision of the EFSA (2013) aquatic guidance. We are able to help you with MDD analysis.

Primary producers

Concerns were raised by some MSs that the exclusive use of ErC50 endpoints for primary producers instead of EbC50 or EyC50 endpoints (that could be lower) could reduce the level of protection. Experts agreed to further consider this in the context of the revision of the EFSA (2013) aquatic guidance. However, some MSs (e.g. Germany) are already applying an additional assessment factor to ErC50 endpoints for national authorisations of plant protection products to account for this potential reduction in protection.

Non-target arthropods (NTAs)

Use of de Jong et al (2010) guidance

Experts agreed that the de Jong et al. (2010) guidance is useful, and some aspects should be used for EU-level assessments until further guidance for the evaluation of NTA field studies is available. Appendix H (p. 83) of the technical report provides detail of how studies can be assessed in terms of their reliability and the information that should be included in the study summary.

Aged residue trials and recovery

No agreement was reached regarding whether the substrate used in NTA aged residue studies needs to be relevant for the crop under assessment, with further guidance being required.

The use of aged residue studies to demonstrate the potential for recovery was discussed, in terms of the maximum aging period that would be acceptable. Some experts considered that the aging period should be lower than the generation or life span of the species tested (up to 1 week for the standard indicator species). However, this was not agreed by all MSs and was noted to be much shorter than the 1 year recovery period stated in the current guidance. No agreement was reached regarding the maximum aging period permitted; however, given the concerns raised regarding these types of studies, it may be that they will no longer be sufficient to demonstrate recovery under the new NTA guidance.

Minimum time considered acceptable for an in-field recolonisation

Concerns were raised with the current NTA in-field recovery period of up to one year. Experts agreed that this should be addressed in future guidance documents, and therefore it is possible that a shorter recovery period will need to be demonstrated for future NTA risk assessments.

Vegetation distribution factor (VDF)

The experts recommended that a vegetation distribution factor (VDF) of 5 should be applied at all tiers of the risk assessment, whereas currently, according to ESCORT II, a VDF of 1 is used for studies where a 3D structure has been treated e.g. whole barley seedling.

Risk assessment for non-target arthropods when oral exposure is relevant

It was agreed that, until guidance is developed and adopted, data for herbivorous species should not be requested. In cases where a concern is raised for oral exposure, then this should be highlighted in the risk assessment and acknowledged in the EFSA conclusion.

Soil organisms

Use of de Jong et al (2006) guidance

It was agreed that the de Jong et al. (2006) guidance can be used for reporting earthworm field studies, in terms of assessing reliability and summarising studies. Some modifications were proposed, which are summarised in Appendix I (p. 89) of the technical report.

Minimum detectable differences (MDDs)

Although MDD is considered a valid concept for the post hoc evaluation of the statistical power of earthworm field studies, additional guidance is needed on, for example, classes of MDD (%) and the minimum number of vulnerable taxa with an acceptable MDD.

Non-target terrestrial plants (NTTPs)

Endpoint based on phytotoxicity

It was agreed that an ERx endpoint based on phytotoxic effects should be reported in the study summary and in the list of endpoints. Moreover, such an endpoint should also be used in the risk assessment where relevant.


It is unclear whether the implementation of the above conclusions will be immediate or whether there will be a lead in time; in some instances the recommendations may be incorporated into updated guidance documents. However, it is advisable to consider how the above conclusions might affect your risk assessments when submitting dossiers for national authorisations or EU active substance assessments, as it is possible that regulators will implement these changes now and this may, in some instances, affect the overall outcome of the risk assessment.

The full technical report with associated appendices can be found on the EFSA website here (EFSA supporting publication 2019: EN-1673. 117 pp. doi: 2903/sp.efsa.2019.EN-1673).

To discuss how the recommendations within the EFSA technical report might affect your ongoing or future dossier submissions, or to discuss any other aspects of your ecotoxicology submission that you need help with, please contact Amy Brooks (

Number of views (262)

Share this article:
Copyright 2019 RSK ADAS Ltd.