Advancing next-generation risk assessment for cosmetic ingredients: transferring the 10-step read-across framework from human to environmental safety using triclosan as a case study

screening; Tier 1 for bioavailability estimation and mode of action hypothesis formulation; and Tier 2 for refinement, hazard characterisation, risk quantification and uncertainty assessment. Four core principles are proposed for environmental NGRA: the assessments should (1) be exposure-anchored, (2) be mechanism-based, (3) have translatable thresholds, and (4) be both protective and transparent. Exposure was assessed using a two-tier use-based predicted environmental concentration (PEC) approach. The lower tier applied a conservative aggregate cosmetic exposure scenario, yielding a screening PEC of 0.74 µg/L. This value exceeded both the ecological threshold of toxicological concern (ecoTTC) and Tier 0 read across thresholds, triggering refinement using year specific exposure data. Integrated bioactivity, QIVIVE and conservation analysis identified lipid and steroid metabolism as the most sensitive pathways. Tier 2 yielded two NGRA predicted no effect concentrations (PNECs): a species sensitivity distribution (SSD)-based PNEC of 0.020 µg/L and a lowest environmental point of departure (ePoD)-based PNEC of 0.007 µg/L. Risk characterisation ratios (RCRs) were below 1 from 2019 onwards under both routes, supporting an environmental NGRA approach at least as protective as conventional ERA. While the proposed framework has been demonstrated for an insecticide, this study extends it to a cosmetic ingredient, broadening the transferability of the 10-step framework for NAM-based ERA.