Steroid Estrogens Profiles along River Stretches Arising from Sewage Treatment Works Discharges RICHARD J. WILLIAMS,* ,† ANDREW C. JOHNSON, † JENNIFER J. L. SMITH, † AND RAKESH KANDA ‡ Centre for Ecology and Hydrology, Wallingford, Oxfordshire, OX10 8BB, U.K., and WRc-NSF Ltd., Medm enham , Marlow, Buckingham shire, SL2 2HD, U.K. Concentrations of estradiol, estrone, and ethinylestradiol were measured in the water column (daily for 28 or 14 days) and in the bed sediment (weekly over the same period) of the River Nene and the River Lea, U.K., upstream and downstream of sewage treatment works (STW). The concentrations of the three steroids in the STW effluents were also measured. Estrone was detected at the highest concentration and in almost all samples from the three STW effluents, concentrations ranging from <0.4 to 12.2 ng/L. Estradiol was also detected frequently (<0.4- 4.3ng/L),but ethinylestradiol was detected infrequently (<0.4- 3.4 ng/L). Positive detections were only found for estrone in the sediment, and these seemed to be unrelated to the water column concentrations. Levels of estrone were clearly raised above background levels in the rivers as a result of the STW discharges. Levels of estradiol and ethinylestradiol were too close to their detection limits to assess the STW impact. River water estrone concentration declined downstream at a rate that was in excess of that due to dilution. The most likely cause of this decline is a combination of sorption and biodegradation equivalent to a first-order decay half-life of 2.5 days for the River Nene and 0.5 days for the River Lea. Introduction Both natural and synthetic estrogenic substances are com- monlyenteringfreshwatersystemsthrough sewage treatment works (STW) effluents. Evidence for the estrogenic activity of such effluents comes from the United Kingdom (1-7), Germany (8, 9), Italy (10), The Netherlands (10, 11), Sweden (12), United States (13, 14), Canada (9, 15), and Israel (16). The compounds found to be responsible for the majority of the in vitro estrogenic activity of domestic sewage treatment works effluents have been the natural estrogens, 17-estradiol (E2) and estrone (E1) and the synthetic E2 derivate 17R-ethinylestradiol (EE2) (7). The concentrations of these steroid estrogens that find their way to the STW effluent pipe depend on the efficiencyoftheir removalin the various STW treatment processes. Such removal rates are high, usually above 75% and often up to 95% for E2 (10, 17). However,even at low concentrations,these compounds can be extremely potent; less than 1 ng/L EE2 can induce vitellogenin (an eggyolkprotein usuallyassociated with adult females) production in male rainbow trout (1, 18), and 4 ng/Lcaused male fathead minnowsto failto develop normal secondary sexual characteristics (19). Computer modeling has shown that, in summer, when sewage effluent can make up a significant fraction of river flows (particularly in the U.K.), steroid estrogens can reach these effect levels in receiving waters (20). In addition, these levels can persist downstream but are lowered by a range of removal processes. This paper presents the first data from a systematic survey of steroid estrogen concentrations in two rivers arising from the discharges from three sewage treatment works. Experimental Section Two intensive field surveys were undertaken in summer/ early autumn 2000 at a time of relatively low flows. These were conducted on the River Nene around the town of Northampton and on the River Lea around the town of Harpenden. Both towns are located in the southeast of England. Low-flow conditions were studied because steroid estrogen concentrations in the receiving rivers would be at their maximum when dilution of the effluent was at a minimum. The steroid estrogens of concern were 17- estradiol, estrone, and 17R-ethinyestradiol. The effluents dischargingto these rivers were known to cause an estrogenic effect in fish (6), and two had been shown previously to contain steroid estrogens (7). The field surveys themselves consisted of taking daily water samples from the effluents and from one upstream site and four downstream sites. The sampling periods were 28 days for the River Nene and 14 days for the River Lea. In addition, bed sediment samples were taken at various times and locations duringthe surveys. Further details ofthe sites and sampling programs are given below. Site Descriptions. Sampling took place at six sites along a 12-km stretch of the River Nene around Great Billing STW, between Northampton and Wellingborough, Northampton- shire, U.K. Great Billing is a relatively large works and serves a catchment of324 000population equivalents(average flow 690 L/s) of which approximately one-third is derived from industrial sources. Following primary settlement, the treat- ment stream is split equally between an activated sludge plant and a biological phosphorus removal plant, each side having a hydraulic residence time of ∼15-18 h. The sludge retention time is ∼17 days. The River Nene runs mainly through farmland and has considerable recreational use for canoeing and narrow boats. The names, distances from the Great Billing STW, and the U.K. grid references of the sampling points are given in Table 1. The section of the River Lea studied covered a distance of 11 km and is predominantly rural in character apart from where it flows through the northeast corner of Harpenden. The upstream site was 0.5 km above the East Hyde final effluent discharge point but still within the confines of the works. East Hyde is an activated sludge works that serves Luton with a population equivalent of ∼130 000 (average flow 530 L/s). The hydraulic retention time averages 13.5 h and the sludge retention time 13 days. Harpenden STW (population equivalent 30 000, average flow 95 L/s) using a biological percolating filter and sand filter system also discharges to this stretch of the River Lea just downstream ofHarpenden.Itshydraulicretention time from inlet to outfall is ∼10.5 h.The STWeffluents were sampled at the point they *Correspondingauthor.Phone: +44 1491 838800. Fax: +44 1491 692424. E-mail: rjw@ceh.ac.uk. † Centre for Ecology and Hydrology. ‡ WRc-NSF Ltd. Environ. Sci. Technol. 2003, 37, 1744-1750