This section includes the presentation of the case studies and their further evaluation:

Overview of the case studies for pyrimethanil, boscalid, metribuzin and ethiprole; the database considered for each case study and details on relevant findings are provided in the Supplementary Information SI-2,

MoA and human relevance assessment for the case study substances,

Introduction of the CLE proposal for a structured approach to support the ED HH classification for the thyroid modality and its application to the case study substances.

Pyrimethanil, boscalid and metribuzin were subject to renewal approval in accordance with Article 15 of the Plant Protection Products Regulation (European Parliament and Council 2009) after 10 November 2018. This is the deadline laid down in Commission Implementing Regulation (EU) 2018/1659 in view of the scientific criteria for the determination of endocrine disrupting properties (European Commission 2018b). Similarly, ethiprole was assessed to support an import tolerance application in accordance with Article 3.2(g) of Regulation (EC) No 396/2005 on maximum residue levels in pesticides (European Parliament and Council 2005) after 10 November 2018. Accordingly, a comprehensive toxicological database that includes ED-related data is available for all four case study substances. Relevant studies include developmental and reproductive toxicity studies using rats and/or rabbits as well as repeated-dose toxicity and chronic toxicity/carcinogenicity studies using rats, mice and/or dogs. Such studies were generally conducted in accordance with formally agreed test guidelines (TGs) such as those adopted by the Organisation for Economic Co-operation and Development (OECD 2025) or the United States Environmental Protection Agency (US EPA 2024). Relevant studies may also include non-standardised in vitro or in vivo assays to assess thyroid-related MoAs and in vitro assays to investigate species differences of effects (OECD 2018). Please see the introduction to the Supplementary Information SI-2 for details on test methods to assess ED via the thyroid modality.

The case studies consider all available statistically significant in vivo and in vitro data from the regulatory submissions that are relevant to assess ED via the thyroid modality (and few non-statistically significant data that contributed to the pattern of effects).

The relevant in vivo toxicological data inform on:

1.

Thyroid-related endocrine activity, i.e. altered serum T4, T3 and TSH levels, altered thyroid weight and/or non-adverse thyroid histopathological changes, such as diffuse follicular cell hypertrophy/hyperplasia, which reflect physiological, adaptive reactions of the thyroid and are generally reversible when the underlying insult is stopped (Lewis et al. 2002; Huisinga et al. 2020), and, possibly, histopathological changes of the pituitary, as these may indicate higher-level effects on the HPT axis,

2.

Thyroid adversity, i.e. thyroid tumour formation and/or focal follicular cell hyperplasia as pre-neoplastic alteration of the thyroid gland that may or may not be reversible (Huisinga et al. 2020; further discussed in Section "Thyroid adversity"),

3.

Neurodevelopmental adversity, i.e. changes in offspring brain weight, morphometry and histopathology and/or offspring neurodevelopment and neurobehaviour,

4.

The substance’s MoA, with liver function examined via liver weight and histopathology, the induction of Phase I and II liver enzymes and specific clinical chemistry parameters,

5.

Signs of systemic toxicity, examined via e.g., clinical observations, clinical chemistry and haematology; such data may indicate that the maximum tolerated dose (MTD) was exceeded so that any endocrine activity observed at these same dose levels should be considered secondary to systemic toxicity; see introduction to the Supplementary Information SI-2 for establishment of MTD.

The relevant in vitro mechanistic and species differences data inform on:

1.

Molecular initiating events, i.e. inhibition of NIS, TPO and/or DIO, interaction with thyroid- and TSH-receptors,

2.

Species differences in liver effects, i.e. Phase I and II liver enzyme activities in hepatocytes from laboratory animal species and human donors.

Of note, none of the case study substances showed evidence for mutagenicity or genotoxicity in any of the available peer-reviewed regulatory studies (in vitro and in vivo; data not shown). Therefore, mutagenic and genotoxic MoAs are not further considered in this article.

Case study 1 —pyrimethanil

Please see Supplementary Information SI-2_CS1 including Table SI-2_CS1_1 to Table SI-2_CS1_3 for pyrimethanil database and for details on the findings from the pyrimethanil studies.

Key findings regarding thyroid-related endocrine activity and thyroid adversity: In rats, pyrimethanil elicited diffuse follicular cell hypertrophy/hyperplasia with a dose–response- and duration-of-exposure relationship; non-neoplastic effects regressed within 14 days post-exposure. Increased incidences of focal follicular cell hyperplasia and follicular cell adenomas after 2-year exposure to pyrimethanil were observed in the high dose group (HDG) animals only. The applied dose level (males/females: 221/291 mg/kg body weight/day [mkd]) was established as close to the MTD in the males but as exceeding the MTD in the females. Therefore, the thyroid adversity observed in the females was assessed as secondary to systemic toxicity. Moderate T4 reductions (< 50%), T3 reductions and TSH increases were observed after 14-day exposure to 379 mkd pyrimethanil in male rats (females not included in study); all hormone changes were reversible within 14 days post-exposure. Serum levels of T4 and TSH remained unaffected in the rat offspring from pregnant/lactating dams exposed to up to 466 mkd pyrimethanil. Also, pyrimethanil did not elicit TPO or NIS inhibition in adult rats in the perchlorate discharge test (for test protocol, see Atterwill et al. 1987; Coelho-Palermo Cunha and van Ravenzwaay 2007). In mice, thyroid gland effects were only observed at dose levels exceeding the MTD, and there were no thyroid effects in dogs.

Summary—thyroid-related endocrine activity: Thyroid-related endocrine activity upon exposure to pyrimethanil was observed in rats, but not in mice or dogs.

Summary—thyroid adversity: Focal follicular cell hyperplasia and follicular cell adenomas were observed in male rats after long-term exposure to high dose levels of pyrimethanil.

Key findings regarding DNT: In the DNT cohorts from the extended one-generation reproductive toxicity study (OECD TG 443), pyrimethanil did not elicit any adverse effects on rat offspring brain weight or histopathology. Isolated brain morphometric findings were assessed as incidental; all neurodevelopmental and neurobehavioural tests were negative. In the two-generation reproductive toxicity study (OECD TG 416), one isolated neurobehavioural finding was assessed as secondary to reduced body weight.

Summary—DNT: Pyrimethanil did not elicit DNT in rats.

Key liver-related findings: In rats, exposure to pyrimethanil generally increased the weight and incidences of histopathological changes in the liver as well as the induction of hepatic enzymes, including UGT. In mice, hepatic glycogen depletion was observed at a dose level that exceeded the MTD (males/females: 1,864/2,545 mkd pyrimethanil). In dogs, exposure to pyrimethanil did not affect liver function.

Summary—liver-related findings: It is postulated that the induction of hepatic enzymes mediating thyroid hormone clearance observed in rats upon exposure to pyrimethanil may be linked with the thyroid-related endocrine activity.

Summary—in vivo and vitro mechanisms: In vivo mechanistic studies provided evidence that pyrimethanil induces Phase II liver enzymes leading to thyroid hormone changes in rats. In vitro assessments showed that pyrimethanil does not inhibit NIS, TPO or DIOs and that it does not interact with human thyroid or TSH receptors.

Case study 2 —boscalid

Please see Supplementary Information SI-2_CS2 including Table SI-2_CS2_1 to Table SI-2_CS2_3 for boscalid database and for details on the findings from the boscalid studies.

Key findings regarding thyroid-related endocrine activity and thyroid adversity: In rats, boscalid elicited diffuse follicular cell hypertrophy/hyperplasia that exhibited a dose–response- and duration-of-exposure relationship and were reversible within 4 weeks post-exposure. Increased incidences of focal follicular cell hyperplasia and thyroid follicular cell adenomas were observed in the male and female rats after lifetime exposure to boscalid at ≥ 22 and ≥ 30 mkd, respectively. Slight and transient T4 reductions (≤ 27% in males, ≤ 13% non-statistically significant in females) and moderate TSH increases (≤ 183% in males; ≤ 177% in females) were observed across studies after up to 4-week exposure to boscalid (dose levels up to 957 and 1,197 mkd in males and females, respectively) and returned to control levels within 4 weeks following 4-week limit dose exposure (1,137 mkd) as shown in a study with male rats. Thyroid-related effects were often more pronounced in male rats than in female rats. Exposure to boscalid did not elicit TPO or NIS inhibition in the rat perchlorate discharge assay, and it did not cause effects on the thyroid gland in mice. In dogs, increased thyroid weight was observed upon exposure to boscalid without concomitant histopathological findings.

Summary—thyroid-related endocrine activity: In rats, boscalid showed thyroid-related endocrine activity. In mice, boscalid did not show any signs of thyroid-related endocrine activity. In dogs, increased thyroid weight was observed, however without associated histopathology findings (thyroid hormone not measured in dogs).

Summary—thyroid-related adversity: Increased incidences of focal follicular cell hyperplasia and follicular cell adenomas were observed in rats after lifetime exposure to boscalid.

Key findings regarding DNT: The rat DNT study (OECD TG 426) yielded no adverse findings at dose levels up to 1,442 mkd boscalid.

Summary—DNT: Boscalid did not elicit DNT in rats.

Key liver-related findings: Changes in liver histopathology were observed in rats and mice upon subacute, subchronic and long-term exposure to boscalid; also, a variety of Phase I and II liver enzymes were induced in the rat studies. In dogs, increased liver weight and few changes in liver-related clinical chemistry parameters were recorded.

Summary—liver-related findings: It is postulated that the liver enzyme induction observed in rats upon exposure to boscalid may be linked with the thyroid-related endocrine activity. Similarly, it is postulated that the increased liver weight and increased thyroid weight observed in boscalid-exposed dogs may be linked.

Summary—in vivo and in vitro mechanisms and in vitro species differences: In vitro assessments showed that boscalid does not inhibit NIS, TPO or DIOs and that it does not interact with human thyroid or TSH receptors. Further, boscalid induced liver enzymes mediating thyroid hormone clearance in cultured rat hepatocytes, but not in human hepatocytes (Fig. 2).

Fig. 2

Effects of boscalid on T4-UGT activities in primary Wistar rat and human hepatocytes compared with reference compounds. Delta activity to the mean (basal) control activity per specific hepatocyte lot; adapted from Wiemann et al. (2023). BNF: ß-naphthoflavone; D day, DMSO dimethyl sulfoxide, PB phenobarbital, PCN 5-pregnen-3ß-ol-20-one-16a-carbonitirile, RIF rifampicin, T4 thyroxine, UGT uridine diphosphate glucuronyltransferase. Whisker plots of n = 9 activity measurements, statistical analyses performed using GraphPad Prism V08, *p < 0.05, **p < 0.01, ***p < 0.001. Solvent control: 0.2% DMSO; reference compounds: 5 µM BNF, 6 µM PCN, 1000 µM phenobarbital; test compound: 5, 10, 20 µM boscalid

Case study 3 —metribuzin

Please see Supplementary Information SI-2_CS3 including Table SI-2_CS3_1 to Table SI-2_CS3_4 for metribuzin database and for details on the findings from the metribuzin studies.

Key findings regarding thyroid-related endocrine activity and thyroid adversity: In rats, exposure to metribuzin elicited follicular cell hypertrophy in some, but not all, repeated-dose toxicity studies (e.g. OECD TG 407, 408), diffuse follicular cell hyperplasia in one chronic toxicity study (at 13.8 and 17.7 mkd in the males and females, respectively), but not in two others (e.g. OECD TG 453), and no thyroid tumours (testing up to 42.2 and 53.6 mkd metribuzin in the males and females, respectively). Thyroid weight was increased in many rat studies. However, most effects on the thyroid gland only occurred at dose levels at which signs of excessive toxicity were also observed. At lower dose levels of metribuzin, and at shorter exposure durations, rats showed increased serum T4, whereas T4 reductions and compensatory TSH increases were recorded at higher dose levels (above approximately 60 mkd metribuzin) and longer exposure durations. In the comparative thyroid assay (US EPA 2005), thyroid hormone changes (in the absence of thyroid histopathological changes) were recorded at the same dose level in the dams and offspring (approximately 138 mkd). In rabbits, 3-week dermal exposure to metribuzin (OECD TG 410) resulted in increased serum T4 and decreased serum T3 (but no effects on thyroid histopathology), whereas all HPT-related parameters remained unaffected by exposure to metribuzin in mice and dogs.

Summary—thyroid-related endocrine activity: Metribuzin shows thyroid-related endocrine activity with a “bi-phasic” T4 effect pattern in rats (i.e. serum T4 increases/decreases at lower/higher dose levels and shorter/longer exposure durations; Bomann et al. 2021). Metribuzin also showed thyroid activity in one rabbit study, but not in mice or dogs.

Summary—thyroid adversity: Thyroid adversity (i.e. focal follicular cell hyperplasia or thyroid tumours) was not observed upon long-term exposure to metribuzin in rats, mice or dogs.

Key findings regarding DNT: There are currently no data on offspring brain weight, morphometry or histopathology or any neurodevelopmental for metribuzin.

Summary—DNT: DNT has not been investigated for metribuzin.

Key liver-related findings: In rats, liver histopathological findings upon exposure to metribuzin were always recorded at the same dose levels as thyroid histopathological findings, whereas increased liver weight was sometimes already recorded at lower dose levels. Metribuzin induced Phase I/II liver enzyme activity in rats (at higher dose levels), dogs and rabbits (liver enzymes not measured in mice). In rats, UGT activity also showed a bi-phasic effect pattern i.e. it was reduced at low dose levels and increased at higher dose levels.

Summary—liver-related findings: It is postulated that the modulation of hepatic enzymes mediating thyroid hormone clearance observed in rats upon exposure to metribuzin may be linked with the thyroid-related endocrine activity.

Summary—in vivo and in vitro mechanisms and in vitro species differences: Metribuzin induced liver enzymes mediating thyroid hormone clearance (in particular T4-UGT) both in rat studies and in cultured rat hepatocytes, but not in human hepatocytes (Fig. 3). In vitro data showed that metribuzin does not inhibit NIS, TPO or DIOs and that it does not interact with thyroid receptors.

Fig. 3

Effects of metribuzin on T4-uridine diphosphate glucuronosyltransferase (T4-UGT) activities in primary Wistar rat and human (Hu) hepatocytes. Delta activity in pmol/min/mg protein to the mean (basal) control activity per specific hepatocyte lot

Case study 4 —ethiprole

Please see Supplementary Information SI-2_CS4 including Table SI-2_CS4_1 to Table SI-2_CS4_4 for ethiprole database and for details on the findings from the ethiprole studies.

Key findings regarding thyroid-related endocrine activity and thyroid adversity: In the rat subacute and subchronic studies, follicular cell hypertrophy and/or diffuse follicular cell hyperplasia were observed in the males and females at ≥ 6.3 mkd and ≥ 7.6 mkd ethiprole, respectively. After 2-year exposure to ethiprole in rats, in addition, slightly higher incidences of focal follicular cell hyperplasia and follicular cell adenomas (not statistically significant) were observed in the HDG (males/females: 10.8/14.7 mkd). Serum T4 levels were consistently reduced and serum TSH levels were consistently increased in several rat repeated-dose toxicity studies assessing ethiprole (up to 52% at 220 mkd), whereas serum T3 levels were only occasionally increased. Also, increased T4 clearance was observed at 20 mkd in a rat 14-day oral gavage study.

No thyroid-related effects (weight, histopathology, thyroid hormone, TSH) were observed in any of the mouse or dog studies in which the respective parameters were measured.

Summary—thyroid-related endocrine activity: Ethiprole elicited thyroid-related endocrine activity in rats, but not in mice or dogs.

Summary—thyroid adversity: Increased incidences of focal follicular cell hyperplasia and follicular cell adenomas (not statistically significant) were observed in the rat after 2-year exposure to ethiprole, but not in mice or dogs, after 18-months and 1-year exposure, respectively.

Key findings regarding DNT: There are currently no data for ethiprole on offspring brain weight, morphometry or histopathology or any neurodevelopmental endpoints.

Summary—DNT: DNT has not been investigated for ethiprole.

Key liver-related findings: Liver changes upon exposure to ethiprole were concomitant with the thyroid observations in all rat repeated-dose toxicity studies and mainly consisted of hepatocellular hypertrophy (generally diffuse, centrilobular) and increased liver weight. In the rat 2-year study, a marginal increase in liver adenomas was observed in the HDG males. The liver was also a target organ in the mouse ethiprole studies with changes consisting of increased liver weight, hepatocellular hypertrophy and, at the end of the 18-month carcinogenicity study, an increased incidence of hepatocellular adenomas in the HDG females (73.5 mkd). In the dog, the liver was only marginally affected upon exposure to ethiprole.

Different Phase I liver enzymes, reflecting pregnane X receptor (PXR)/constitutive androstane receptor (CAR) activation, were increased upon exposure to ethiprole both in the rat and the mouse. Indirect evidence for in vivo T4-UGT induction in the rat was provided in the 14-day T4 biliary clearance study, in which increased excretion of T4 was observed (see above).

Summary—liver-related findings: It is postulated that the liver effects observed in rats upon exposure to ethiprole may be linked with the thyroid-related endocrine activity.

Summary—in vivo and in vitro mechanisms and species differences: Ethiprole induced Phase I liver enzymes in both rat and mouse studies; in addition, there is evidence for in vivo T4-UGT induction in the rat 14-day biliary clearance study (see above). In the in vitro species comparisons, Phase I and Phase II enzyme activities (especially T4-UGT) were markedly increased following exposure to ethiprole in rat hepatocytes but to a much lesser extent in human hepatocytes, (Fig. 4). In vitro assessments showed that ethiprole does not inhibit NIS or TPO.

Fig. 4

Effects of ethiprole on T4-uridine diphosphate glucuronosyltransferase (T4-UGT) activities in primary Wistar rat and human (Hu) hepatocytes. Delta activity in pmol/min/mg protein to the mean (basal) control activity per specific hepatocyte lot

MoA and human relevance assessment

All four case study substances, i.e. pyrimethanil, boscalid, metribuzin and ethiprole, elicited thyroid-related endocrine activity in rats. Generally, these thyroid effects coincided with liver effects, including the induction of liver enzymes that mediate thyroid hormone clearance.

Considering thyroid and liver effects in further species, 90-day and 1-year exposures to boscalid led to increased thyroid and liver weight without concomitant histopathological findings in dogs, and 3-week dermal exposure to metribuzin caused altered thyroid hormone levels, reduced liver weight and Phase I liver enzyme induction in rabbits. By comparison, ethiprole caused thyroid effects only in the rat, but not in the mouse or dog, whereas liver effects were also recorded in these species.

On account of their effect patterns, it is postulated that pyrimethanil, boscalid, metribuzin and ethiprole elicit effects on the HPT axis via a liver enzyme induction-related MoA. Based on the broader scientific knowledge (Section "State-of-the-science thyroid-related MoAs"), it is also postulated that this MoA, when leading to thyroid adversity in rats, is generally not relevant to humans. By comparison, there are no human studies to establish a link between substance-mediated liver enzyme induction and increased thyroid hormone clearance, let alone further to maternal hypothyroxinaemia and child neurodevelopmental impairment (Sauer et al. 2020). Therefore, the human relevance of thyroid-mediated NDT in rats is currently unclear. Nevertheless, the authors postulate that the non-relevance to humans of a liver enzyme induction-related MoA (regardless of the adverse outcome in rats) can be established via in vitro liver enzyme induction species comparisons (see, however, Section "Uncertainties related to the MoA assessments" for scientific limitations of this assay). This is the background to the MoA and human relevance assessment for the four case study substances.

The MoA and human relevance assessment is organised applying information from the AOP concept, as recommended by Melching-Kollmuss et al. (2023). The AOP Wiki (OECD 2024) includes one AOP that describes events leading from liver enzyme induction to thyroid hormone imbalance and ultimately loss of cochlear function, i.e. AOP 8 Upregulation of thyroid hormone catabolism via activation of hepatic nuclear receptors, and subsequent adverse neurodevelopmental outcomes in mammals (Fig. 5). The MoA and human relevance assessments for pyrimethanil, boscalid, metribuzin and ethiprole are structured by the key events for AOP 8 while focussing more broadly on NDT as neurodevelopmental adverse outcome and not only on ototoxicity (for rationale, see Supplementary Information SI-1.2). In addition, the MoA assessments consider focal follicular cell hyperplasia as pre-neoplastic change and follicular cell adenomas as thyroid adverse outcomes. The formation of these tumours in rats (or of malignant thyroid carcinomas, which, however, were not observed for any of the case study substances) is generally considered non-relevant to humans if caused by long-term stimulation of thyroid gland growth due to substance-mediated UGT induction and increased thyroid hormone clearance (Dellarco et al. 2006; Elcombe et al. 2014; Bartsch et al. 2018; Foster et al. 2021; Section "State-of-the-science thyroid-related MoAs").

Fig. 5

AOP 8: Sequence of events that may lead from liver enzyme induction to adverse neurodevelopmental outcomes in mammals and opportunities for their further investigation (adapted from Melching-Kollmuss et al. 2023). AhR aryl hydrocarbon receptor, AO(P) adverse outcome (pathway), BROD benzoxyresorufin, CAR constitutive androstane receptor, Cyp cytochrome p-450, KE key event, LDG lower-dose groups, MIE molecular initiating event, PBK physiologically based kinetic, PPAR peroxisome proliferator-activated receptor, PROD pentoxyresorufin, PXR pregnane X receptor, T3 triiodothyronine, T4 thyroxine, TDG top-dose group, UGT uridine diphosphate glucuronyltransferase. Colour legend: yellow arrow: available data indicating that AOP 8 may be relevant for the MoA and human relevance assessment. Boxes with blue shading: MIE, early KEs and AO. White boxes with blue text: Supportive in vivo or in vitro evidence that may inform on the MIE or specific KEs for AOP 8 (linked by blue arrows; black arrow for enhanced traceability). Boxes with green shading: KEs relating to serum/brain T4 decrements; these KEs are central to five of the six potentially relevant AOPs included in the AOP Wiki (exception: AOP 300 on thyroid receptor antagonism; Supplementary Information, Table SI-1). White box with green text: In vivo data or PBK modelling to inform on thyroid-related events. Boxes with ochre shadings: KEs relating to the hippocampus; these KEs are central to five of the six potentially relevant AOPs (exception: AOP 54 on NIS inhibition leading to impaired learning and memory). [a] See Tinwell and Bars (2022) for details on the indirect assessment of CAR/PXR activation in rat studies via induction of transcript level and corresponding enzyme activity associated with each receptor (Cyp2b/PROD and Cyp3a/BROD for CAR and PXR, respectively). [b] In the AOP Wiki, the MIE of AOP 8 is recorded as PXR activation. Noyes et al. (2019) indicated CAR, AhR and PPAR activation as further MIEs leading to liver enzyme induction. However, since all these MIEs are not indispensable to trigger UGT upregulation, their assessments may not be needed for the MoA assessment of the substance of interest (Melching-Kollmuss et al. 2023). [c] PBK modelling, if available: Estimate serum/brain T4 levels in rat vs human considering relevant parameters, such as binding constants, potencies of MIEs and/or liver enzyme inductions in rat vs human tissue. [d] The AOPs in the AOP Wiki only generally refer to “T4 in serum, decrease” without distinction between maternal and offspring serum T4 levels; also, none of the AOPs considers serum T3 (or TSH). Following the observations by Marty et al. (2022), maternal serum T4 levels do not appear predictive of neurodevelopmental effects. However, there seems to be some association between ≥ 60% / ≥ 50% offspring serum T4 decrements in the TDG/LDGs (and ≥ 20% and statistically significant offspring serum T3 decrements) and the occurrence of statistically significant neurodevelopmental effects (see also Supplementary Information, Table SI-2). Therefore, Melching-Kollmuss et al. (2023) recommended considering offspring serum T4 as predominant parameter related to serum thyroid hormone levels. In addition, information on maternal serum T4, maternal and/or offspring serum T3 and offspring brain T4/T3 should be considered, if available

To assess the causality of observed associations between substance exposure and adverse outcomes, the MoA and human relevance assessments are based upon the five “evolved Bradford Hill considerations” described in the World Health Organisation/International Programme on Chemical Safety MoA and species concordance analysis framework (Meek et al. 2014a, b):

Consistency: “Is the pattern of effects across species/ strains/ organs/test systems what would be expected?”

Essentiality of key events: “Is the sequence of events reversible if dosing is stopped or a key event prevented?”

Temporal (sequential) concordance: “Are the key events observed in hypothesized order?”

Dose–response concordance: “Are the key events observed at doses below or similar to those associated with the end (adverse) effect?”

Biological concordance: “Does the hypothesized MoA conflict with broader biological knowledge? How well established is the MoA in the wider biological database?”

Case studies 1 and 2 —pyrimethanil and boscalid

Pyrimethanil and boscalid are discussed together in this section because their databases include data on NDT. Generally, the MoA and human relevance assessments for pyrimethanil and boscalid focus on the findings from the rat studies. Boscalid also elicited increased thyroid and liver weight in dogs. However, the absence of histopathological findings in these organs in both dog studies (Section "Case study 2—boscalid") indicates that boscalid does not elicit thyroid adversity in dogs, and it shows that this species is less sensitive to boscalid-mediated thyroid effects than rats. Consistent with this estimation, serum levels of thyroxine binding globulin are considerably lower in rats than in dogs (Choksi et al. 2003; Daminet and Ferguson 2003; Jahnke et al. 2004), resulting in a lower capacity to compensate for serum thyroid hormone imbalance (Section "State-of-the-science thyroid-related MoAs").

The rat studies evaluating pyrimethanil and boscalid showed liver effects and thyroid-related endocrine activity. Sustained thyroid-related endocrine activity led to increased incidences of focal follicular cell hyperplasia and follicular cell adenomas in male rats exposed to pyrimethanil and in male and female rats exposed to boscalid (observed in the 2-year studies). By comparison, the DNT assessments did not provide any indication for neurodevelopmental adverse outcomes in the rat offspring after in utero/lactational exposure to pyrimethanil or boscalid.

The thyroid and liver effect patterns recorded in the rat studies assessing pyrimethanil and boscalid were evaluated for consistency and essentiality of the key events (Table SI-2_CS1/CS2_1) and for the dose–response and temporal concordances of the sequence of liver and thyroid effects that reflect the early events of AOP 8 (pyrimethanil: Table SI-2_CS1_4, boscalid: Table SI-2_CS2_4). For both pyrimethanil and boscalid, liver- and thyroid-related findings reflecting all early key events of AOP 8 occurred consistently across rat studies, and they were linked in a biologically plausible manner as was shown in the mechanistic in vivo thyroid hormone level and liver enzyme induction studies. The no/lowest observed (adverse) effect levels (NO(A)ELs/LO(A)ELs) recorded for the liver-related events were generally similar to (or lower than) those for the thyroid-related events. Also, the liver- and thyroid-related effects observed for pyrimethanil and boscalid in rats exhibited both dose–response and temporal relationships. The effects on liver enzyme induction and liver weight and histopathology were seen at 4 days and 2 weeks after initiation of administration of pyrimethanil and boscalid, respectively, indicating that the liver changes were early effects. Altered serum levels of T4 and TSH as well as thyroid histopathological findings were also recorded within the first days of treatment, and effects were generally reversible when exposure was discontinued. Importantly, pyrimethanil and boscalid did not elicit thyroid-related effects in the absence of changes in liver parameters (apparent thresholds for liver and thyroid effects across studies: above approximately 50–60 mkd and 20 mkd for pyrimethanil and boscalid, respectively). These findings are consistent with current knowledge for the postulated MoA. For all early events of AOP 8, consistency, essentiality of key events and dose–response and temporal concordance could be demonstrated across the rat studies evaluating pyrimethanil and boscalid. This provides strong and consistent support for the biological plausibility of the postulated MoA for pyrimethanil (Table 2) and boscalid (Table 3). By comparison, all potentially relevant alternative MoAs could be reasonably ruled out (Section "Case study 1—pyrimethanil" and "Case study 2—boscalid").

Table 2 Case study 1—pyrimethanil: summary of the MoA analysis and conclusions on biological plausibility for link between eve