CEA presented 17 posters and platforms at SETAC Copenhagen in May 2022. Over the coming weeks we will be showcasing each of these presentations in a series of 'SETAC Spotlight' articles. This week it is:
Authors: Amy Brooks and Sarah Bull
Under the EU regulation for authorising plant protection products (Regulation (EC) 1107/2009), the potential risks to wild mammals must be assessed. The current default approach for wild mammal risk assessments is to use the worst-case no observed adverse effect level (NOAEL) that was concluded from the toxicology package for the human health risk assessment. However, there are many parameters that are measured in toxicology studies that are relevant for individual humans but not for wild mammal populations.
The current bird and mammal guidance (EFSA 2009) theoretically allows the derivation of an ‘ecologically relevant’ wild mammal toxicity endpoint that only takes into account those effects relevant for wild mammal populations. It was later recommended by EFSA in 2015 that the assessment of ecological relevance should be made at the active substance level. In September 2021, EFSA released the draft new bird and mammal guidance, which includes recommendations for the derivation of ecorelevant endpoints for wild mammals.
The aim of this poster is to review the proposals under the draft new guidance (EFSA 2021) for deriving ecorelevant endpoints for wild mammals, including issues such as which studies to use, species sensitivity, types of effect, and magnitude of effect. The approach for deriving ecorelevant endpoints proposed in EFSA (2021) was reviewed. The rationale behind the recommendations was considered, together with implications for the risk assessment.
Which studies to use
The toxicity dataset available for mammals is vast in comparison to other vertebrate wildlife groups for which risks from exposure are assessed. This is due to the mammal data being primarily collected for the human health assessment, for which effects on individuals are critical, whereas for birds and fish, only effects at the population level are considered. Hence for birds, a single one-generation study is required to assess long-term effects. For fish, either an early life stage or full life cycle study is required to assess the risks from chronic exposure. Thus, for both birds and fish, usually endpoints from a single study on a single species are used in the risk assessment. This is in contrast with the wild mammal risk assessment, for which it is proposed that the rodent multi-generation study and rodent/rabbit development studies are taken into account, with other toxicology studies also being considered e.g. neurotoxicity studies. In addition, some of the mammalian studies (such as OECD 416) perform exposure over longer durations than performed in bird studies, and for longer than would be predicted in the wild. Therefore, a higher level of protection is being afforded to wild mammals compared to birds, which are under the same risk assessment framework and have the same protection goals.
Instead of using all available toxicology studies, which were not designed to detect population relevant effects in wild mammals, the focus should be on the most relevant study i.e., the multigeneration rodent study. The other studies should only be used to derive endpoints where there are very good reasons to do so.
Alternatively, if the additional toxicology studies are included in the derivation of the wild mammal endpoint, then this should be reflected in the assessment factor used in the risk assessment. This would acknowledge that by increasing the dataset used for deriving the endpoint, the level of uncertainty is reduced. As such, the assessment (or uncertainty) factor can also be reduced to <5.
The multi-generation mammal study is a much longer study than used for birds and fish, covering at least two generations of offspring (F1, F2) instead of just one. The rodent multi-generation study has been designed to address the risk to human health, whereas the avian and fish studies were designed for wildlife risk assessments. However, even for human risk assessment purposes there are already frameworks under which the study can be shortened, using the so called Extended One-Generation Reproductive Toxicity Study (EOGRTS; OECD 2018c). If the same approach used for wild birds and fish were followed for wild mammals, then only the effects observed in the F0 and F1 animals should be taken into account in the risk assessment, and thus an EOGRTS would be a more appropriate study design.
This raises the question, just because the additional mammalian data (i.e., 2nd generation) are available, should we be using them for the wild mammal risk assessment? If the answer to that question is yes, and we genuinely need to use all these additional data to adequately protect wild mammal populations, why aren’t we also demanding this same level of data for birds (which are assessed under the same guidance document as wild mammals and thus have the same protection goals), and indeed other wildlife groups e.g. fish? If the answer to the question is no, then we should focus on the data that are key for addressing the protection goal, with the most relevant of those studies being the multi-generation study or EOGRTS, which are most comparable to those studies required for birds and fish.
When deciding whether or not the 2nd generation data should be considered for endpoint setting, the decision criteria used for extending the EOGRTS could be used. If the F0/F1 results wouldn’t have required the study to proceed to F2, the results from the F2 phase (and F3 etc) could be excluded from the endpoint selection.
The 28-d and 90-d studies (including a recovery phase) can provide useful information for the argumentation around the time weighted average factor (ftwa), providing insight into whether effects are driven by short term or long-term exposure.
It is known that some active substances are more toxic to lagomorphs than rodents, and vice versa, due to their mode of action and the physiological differences between lagomorphs and rodents. As such, if this is a genuine difference based on robust scientific evidence, it is proposed that this can be taken into account within the risk assessment. For example, an endpoint based on the rabbit development study could be used for any lagomorph-type scenarios in the risk assessment, and the remaining rodent scenarios could be based on endpoints from rat/mice studies.
Furthermore, the scenario type could also be considered when selecting the most relevant endpoint. For example, for seed treatments, it would be inappropriate to use a toxicity endpoint based on effects in lagomorphs for a seed-eating rodent scenario.
Types of effect
In terms of which effect types to take into account for ecologically relevant endpoints, the focus should be on those that can potentially affect population growth i.e., reproductive success and offspring viability. As such, the main study type that should be used is the multi-generation study or EOGTS. Related effect types measured in other chronic/dietary toxicology studies should be reviewed to determine if they support what was observed in the multi-generation study. They may also provide useful information to explain why particular effects were observed in the multi-generation study, and whether effects observed are ‘real’ effects or due to natural variability and/or an artefact of dose-spacing.
In terms of effects on bodyweight, the multi-generation study may be overly conservative for this parameter given its prolonged exposure period (56 days in mice, 70 days in rats); instead, it may be more appropriate to use the shorter dietary studies where available in the toxicology package to determine body weight effects, or focus only on effects observed within a short time of exposure initiation in the multi-generation study. Where possible, body weight effects should be further distinguished between those arising from reduced food intake due to unpalatability, and those arising due to toxic effects.
The additional toxicology studies can also be used to determine whether there is any indication that there is a genuine difference in sensitivity between rabbits and rodents, and if so, derive separate NOAELs for the two groups to use in the wild mammal risk assessment for lagomorphs and small mammals, respectively.
Magnitude of effect
The draft new guidance states that an effect of 10% or greater will be considered relevant; however, it is not clear what the 10% refers to. Does this mean a 10% increase/decrease in a particular parameter e.g., startle reflex? Or a 10% increase/decrease in the number of organisms expressing the behaviour?
It is unclear why a 10% threshold has been selected for behavioural effects. It seems unlikely that such a difference could be reliably detected in toxicology studies, and indeed, some behavioural effects are not currently quantitatively measured. It also seems unlikely that such a small behavioural change would significantly impact wild mammal populations. It may also be the case that the artificial conditions of the laboratory (and perhaps the specially bred lines used in testing) result in expressions of behaviour which would not occur in the wild (e.g., due to other survival pressures). Research should be undertaken to determine which types of behavioural changes and at what magnitude are relevant at the population level. This should include a consideration of not only the behavioural change being detectable by human observers, but whether the change is perceptible to the wild mammal population or its predators i.e., does it actually impact on their ability to breed or survive?
In the meantime, it is suggested that a similar approach to that taken for birds is applied to wild mammals i.e., use the acute LD50/10 as a proxy for sublethal changes that may impact pair formation, etc.
It is stated in the draft new guidance that the power to detect effects is likely closer to 20%. Therefore, the decision to propose a 10% effect level for benchmark dose (BMD10) modelling is not consistent with this. Furthermore, the use of the lower limit of the BMD10 (BMDL10) will add uncertainty to an endpoint of already questionable statistical power.
The use of a 10% effect level threshold regardless of the effect type is also problematic. It is likely that an effect of e.g., 15% on pup body weight will have a different impact at the population level compared to e.g., 15% effect on pup survival. Wild mammals and birds will have different tolerance thresholds to different types of effects, and this should be taken into account in setting appropriate thresholds for those effect types.
No evidence has been presented in the draft new guidance to illustrate that 10% effects would impact wild mammal populations. We suggest if such evidence exists, that it is presented in the guidance. If no such evidence currently exists, we suggest research into this area to determine what types of effects are important for regulating bird and wild mammal populations and at what thresholds.
The draft new guidance has expanded the recommendations for the derivation of ecorelevant endpoints for wild mammals. The increased guidance on this topic is welcomed, as this will encourage consistency between regulatory authorities and between active substances; however, the nuances of the recommendations require further consideration to ensure that we are not over or under protecting wild mammals due to the assumptions being made or the data being used.
Further communication is encouraged between toxicology and ecotoxicology experts in deriving ecologically relevant wild mammal endpoints. This is important not only during the initial dossier preparation, to ensure robust and relevant data interpretation, but also during EU expert reviews. The setting of wild mammal NOAELs requires a multi-disciplinary approach, in terms of what the effects mean, how the various studies and effect types are interlinked, and which studies are the most relevant for each effect type.
European Food Safety Authority (EFSA). 2009. Guidance document on risk assessment for birds and mammals on request from EFSA. EFSA Journal 7:1438.
European Food Safety Authority (EFSA). 2015. Technical report on the outcome of the pesticides peer review meeting on general recurring issues in ecotoxicology. EFSA Supporting Publications 2015:EN-924.
European Food Safety Authority (EFSA). 2021. Risk assessment for birds and mammals. https://connect.efsa.europa.eu/RM/s/publicconsultation2/a0l1v00000E7Zev/pc0090
You can find all of the other posters and platforms that CEA presented at SETAC Copenhagen here
. You can also find all of our publications from previous conferences and links to journal articles we have authored on our library page