Evaluation of the fish short term reproduction assay for detecting endocrine disrupters

► Quantifying the variability of the fish short term reproduction assay (FSTR) in detecting EDCs. ► The FSTR not sensitive enough for chemicals with anti-androgenic and steroidogenesis activities. ► Suggestions for a testing strategy are given. In a fish testing strategy, positive results of the fish short term reproduction assay (FSTR), often trigger a definitive test like the fish sexual development test (FSDT) or the fish full life cycle test (FFLC), entailing ethical and economic problems. This study analysed 137 studies encompassing 35 chemicals with different modes of actions (MOAs). Variability is quantified for MOA endpoints vitellogenin (VTG) and secondary sex characteristics (SSCs) as well as for apical endpoints. Two MOA endpoints could indicate estrogenic, anti-estrogenic, androgenic, anti-androgenic and steroidogenesis activities. Great variability, however, has been observed for chemicals with anti-androgenic and steroidogenesis activities, suggesting that TG229/230 may not be sensitive enough to detect these types of chemicals and may produce false negatives. Changes in apical endpoints like fecundity are not limited to endocrine disrupting chemicals (EDCs). Non-EDCs could induce the similar effects on these apical endpoints. If elucidating MOA is needed, targeted in vitro MOA tests are suggested. Positive in vitro MOA results trigger a definitive test, which could be used for confirmation of the MOA in vivo and for deriving a no observed effect concentration (NOEC). Based on positive MOA results of TG229, a definitive test such as the FSDT or the FFLC is still needed, because the current TG229 has limitation on the derivation of a NOEC. An extended TG229 with more power to detect reproduction effects, as recently proposed in the OECD test guideline program, would improve the possibility to derive a NOEC and increase its usefulness in risk assessment.