Three estrogens drive the majority of estrogenic activity in global rivers

Estrogenic endocrine-disrupting chemicals pose a threat to aquatic ecosystems even at trace concentrations, yet their global occurrence, dominant contributors, and ecological risks in rivers remain poorly characterized. Here, we integrated machine-learning predictions, multiple bioassays, and both chemical- and effect-based risk assessments across 1096 rivers to map the pollution patterns of three representative estrogens (estrone (E1), 17β-estradiol (E2), and 17α-ethinylestradiol (EE2)) and their associated estrogenic activities. We revealed pervasive contamination of these three estrogens (global means: 7.3-10.3 ng/L) with pronounced latitudinal and socioeconomic heterogeneity. Together, the three target estrogens accounted for the majority (56-81%, assay‑dependent) of the estrogenic activity in global rivers. Despite similar concentration ranges, their ecological risks differed by five orders of magnitude, mainly due to differences in estrogenic potency. EE2 and E2 therefore emerged as priority targets for global monitoring and risk management in the context of the UNEP Global Framework on Chemicals. Under a comprehensive assessment incorporating unidentified or unquantified estrogenic contributors, the effect-based risks were 1.5‑to 7.4‑fold higher than the cumulative risks estimated solely from the three target estrogens, indicating that unknown or unquantified mixture components substantially drive the overall hazard. This systematic discrepancy highlights the inherent limitations of conventional targeted chemical analysis in capturing mixture-driven toxicity and underscores the necessity of integrating bioassay-based effect screening with high-resolution non-targeted chemical characterization for environmentally realistic risk assessment.