Step III: Evaluation and summary of the evidence for an adverse effect

The purpose of Step III of the ECETOC 7SI-ED is to examine the available evidence for adverse effects which may be indicative of endocrine disrupting properties and to determine the sufficiency of those data to conclude on adversity. There is nothing different about the way adverse effects are assessed for endocrine disrupting endpoints compared to established procedures for non-endocrine disrupting endpoints. When identifying endocrine disrupting properties, a specific next step after the determination that an adverse effect is present is its characterisation to possibly help direct which specific endocrine activity may merit follow-up in Step IV. The recognition of adverse effects and the differentiation between adverse and non-adverse effects is made as described by Lewis et al. (2002). For environmental assessments, adverse effects with population relevance are identified. For this purpose, it is important to determine which toxicity test endpoints are population-relevant and also what levels of effect on the endpoints may impact a population (Weltje et al., 2013).

OECD CF Level 4 and 5 studies (or other relevant studies that use different study designs or investigate endpoints, for which validated TGs are not yet available or listed in the OECD CF) provide information on whether a substance causes adverse effects on apical endpoints that may be indicative of endocrine disrupting properties. As compared to the Level 3 in vivo assays (that are not designed to assess adversity, but to assess a specific endocrine activity (Wheeler et al., 2014)), Level 4 and 5 studies are sensitive to a variety of toxicological MoAs, including one or more endocrine MoA, and they cover numerous apical endpoints across different life stages and larger portions or all of the life cycle of the test species. For this reason, the Level 4 and 5 studies can provide a more thorough assessment of the possible or actual endocrine disrupting effects of a substance in developing or adult organisms than the Level 3 assays. While apical effects identified in Level 3 in vivo assays should also be taken into consideration for a comprehensive WoE evaluation in Step III, such findings are not decisive in the absence of Level 4 or 5 data to identify adverse effects that may be indicative of endocrine disrupting properties.

Generally, in Step III a treatment-related, significant and biologically relevant effect that may be indicative of endocrine disrupting properties flags the need for further consideration (in Steps IV and V) to elucidate the MoA involved in the causation of the adverse effect. Effects which may be indicative of an endocrine MoA can also be caused by other (non-endocrine) MoAs. Many of the Level 4 and 5 studies (and similar study types) are designed to identify a range of adverse effects, and not only those which may be indicative of endocrine disrupting properties. These considerations notwithstanding, scenarios are conceivable in which diagnostic endpoint data from a Level 4 and 5 study could be used in Steps IV and V as part of the WoE approach, which is used to establish the biologically plausible link between an adverse effect elicited in this same study and the endocrine MoA. A substance that does not cause adverse effects on apical endpoints that are indicative of endocrine disrupting properties in a suite of Level 4 or 5 studies (or other relevant study designs) does not fulfil the criterion for adverse effects in the WHO/IPCS (2002) definition of an endocrine disruptor.

For human health evaluations of endocrine disrupting properties, adverse effects may constitute changes in the morphology, physiology, growth, development, reproduction, or life span of the organism that result in an impairment of functional capacity, an impairment of the capacity to compensate for additional stress, or an increase in susceptibility to other influences (WHO/IPCS, 2009). The Level 4 and 5 chronic / carcinogenicity and developmental / reproductive toxicity studies provide data on adverse effects and add to the WoE concerning the potential for impacts in humans, while at the same time providing data on dose/concentration-response.

For human hazard identification, it is essential to assess whether the adverse effect(s) observed in in vivo animal studies are relevant to humans. The current default assumption is to assume human relevance. However, this assumption can be re-evaluated if there are scientifically valid data to demonstrate non-relevance to humans. The most well-known example for non-human relevance of adverse endocrine-mediated effects in animals is the susceptibility of the rat to disruption of thyroid function, and it can be explained by species-specific differences in synthesis, binding, metabolism/clearance and transport of the thyroid hormones (Lewis, 2013). These considerations should also consider toxicodynamic and biological differences in species responses such as already accepted for the human versus rodent relevance of receptor-mediated toxicity via peroxisome proliferator-activated receptor alpha for certain liver tumours (Corton et al., 2014).

For wildlife evaluations of endocrine disrupting properties, the assessment of adverse effects needs consideration of the population relevance of effects in order to reflect the protection goal of environmental assessments. The WHO/IPCS (2009) definition of adverse effects includes consideration of population effects. Therefore, it is crucial to determine which apical endpoints are population-relevant, as well as what level of effect on these endpoints may impact a population, bearing in mind that the magnitude of change will depend on the dose or concentration tested. Population-relevant endpoints are generally assessed by survival, growth, development (including sexual) and reproduction in Level 4 and 5 studies. Whilst it is important to determine what level of effect on these endpoints may impact populations, guidance on this aspect is still lacking. However, these endpoints can also be affected by generalised and specific non-endocrine toxicities, which highlights the need for mechanistic evaluation (covered in Step IV). A number of mechanistic endpoints that are diagnostic for endocrine disruption are included in more recently adopted OECD TGs; e.g. OECD TG 234 and 241 (Level 4) and OECD TG 240 (Level 5), whereas others do not provide any mechanistic insight; e.g. OECD TG 206 (Level 5).

For human health and wildlife evaluations of endocrine disrupting properties alike, effects seen above the recommended limit dose or concentration for the studies should not be considered, and those observed at the highest tested dose/concentration require careful consideration (Wheeler et al., 2013; Mihaich et al., 2017). Systemic toxicity at these doses/concentrations can lead to non-specific adversity which can result in apical changes that could be associated with endocrine disruption. However, such changes may be related to the overload of the organism, resulting in stress, or occur concomitant with generalised, systemic toxicity (Wheeler and Coady, 2016). If the MTD or MTC is reached or exceeded in a study, and generalised, systemic toxicity is observed, then any endocrine relevant-effects observed at these doses/concentrations should generally be excluded from considerations on endocrine activity/disruption (Wheeler et al., 2013). The relevance of effects occurring at doses that exceed the inflection point of nonlinear toxicokinetic behaviour, i.e. a ‘kinetically derived maximum dose’, should also be carefully considered, particularly, if there is an adequate margin with human exposures and the relevance of the toxicokinetic processes in the test species to humans is reasonably established (Creton et al., 2012).

In determining the consistency and coherence of effects amongst different studies, it should be taken into consideration that the route of exposure may not be representative of normal conditions of use making direct extrapolations difficult (e.g. intraperitoneal exposure in a study when human exposure is dermal or oral). Similarly, species-specific differences in substance adsorption, distribution, metabolism, and elimination (ADME) should be taken into account in assessing the consistency and coherence of effects amongst different studies. This is particularly important when considering the consistency and coherence of mammalian and non-mammalian toxicity data. Expert judgement is required on a case-by-case basis to evaluate whether data are conflicting and to determine the implications of conflicting data.

OECD GD 150 provides guidance on the specific apical endpoints that are covered by the individual Level 4 and Level 5 studies and on the relevant endocrine MoA(s) that are indicative of these apical endpoints.