Endocrine disruption due to estrogens derived from humans predicted to be low in the majority of U.S. surface waters

Abstract

In an effort to assess the combined risk estrone (E1), 17β-estradiol (E2), 17α-ethinyl estradiol (EE2), and estriol (E3) pose to aquatic wildlife across United States watersheds, two sets of predicted-no-effect concentrations (PNECs) for significant reproductive effects in fish were compared to predicted environmental concentrations (PECs). One set of PNECs was developed for evaluation of effects following long-term exposures. A second set was derived for short-term exposures. Both sets of PNECs are expressed as a 17β-estradiol equivalent (E2-eq), with 2 and 5 ng/L being considered the most likely levels above which fish reproduction may be harmed following long-term and short-term exposures, respectively. A geographic information system-based water quality model, Pharmaceutical Assessment and Transport Evaluation (PhATE™), was used to compare these PNECs to mean and low flow concentrations of the steroid estrogens across 12 U.S. watersheds. These watersheds represent approximately 19% of the surface area of the 48 North American states, contain 40 million people, and include over 44,000 kilometers of rivers. This analysis determined that only 0.8% of the segments (less than 1.1% of kilometers) of these watersheds would have a mean flow E2-eq concentration exceeding the long-term PNEC of 2.0 ng/L; only 0.5% of the segments (less than 0.8% of kilometers) would have a critical low flow E2-eq exceeding the short-term PNEC of 5 ng/L. Those few river segments where the PNECs were exceeded were effluent dominated, being either headwater streams with a publicly owned treatment works (POTW), or flowing through a highly urbanized environment with one or several POTWs. These results suggest that aquatic species in most U.S. surface waters are not at risk from steroid estrogens that may be present as a result of human releases.