Endpoint sensitivity in fish endocrine disruption assays: Regulatory implications

Identifying potential endocrine disrupting chemicals (EDCs) needs screening and testing for mode of action (MOA) and intrinsic toxicological properties. MOA is often indicated by biomarker endpoints, whereas toxicity by apical endpoints. Risk assessment is mainly based on apical but not on biomarker endpoints. The 21-day fish assay (OECD TG229) is considered a screening test. But it includes both biomarker and apical endpoints. This study explores the utility of results of the 21-day fish assay for risk assessment purposes. Endpoint sensitivity was analysed by compiling 142 data sets for 21-day fish assays and 38 data sets for the fish sexual development test (FSDT), encompassing 62 chemicals with different MOAs. Conclusions from this analysis include: (1) vitellogenin (VTG), fecundity and gonad histology are the most sensitive endpoints for fathead minnow, medaka and zebrafish in 21-day fish assays; secondary sex characteristics (SSC) are a less sensitive endpoint and is likely inadequate to detect all known MOAs. (2) Biomarker endpoints like VTG and apical endpoints like fecundity from the 21-day fish assay can be used for risk assessment. (3) Lowest observed effect concentrations (LOECs) of the most chemicals are comparable for the 21-day fish assay and for the FSDT, further supporting that results of 21-day fish assays can be used for risk assessment. However, a significant difference in LOECs was observed for some chemicals, suggesting that chemical specific effects should be taken into account. This paper emphasizes that a weight of evidence approach is important for interpretation of results of the 21-day fish assay.