Estrogens in Swiss rivers and effluents: Sampling matters

Estrogenic activity of treated sewage effluents and receiving waters surfaced as an issue of general concern some 15 years ago. Since then, a large number of studies investigated the distribution and nature of the estrogenic substances in various aquatic ecosystems. Within NRP50, a five-year Swiss research programme on endocrine disruptors, four field studies were performed to characterize the presence of environmental estrogens in rivers and effluents. The estrogenic activity was quantified with a biological assay - the yeast estrogen screen. An overview of the sampling approaches and results of the various studies is presented here. In a first study, using grab sampling, it appeared that estrogenic activity in river water was highly variable. Average estrogenicity values did not correlate with sewage treatment works and/or river characteristics, e.g. effluent dilution factor. However, variability was not 'random' but clearly associated with river size, and possibly its discharge. A second study specifically addressed this issue of variability of estrogenicity. The study was conducted at a single effluent source and its receiving river. Variability of estrogenicity in the grab samples was again large but again not 'random'; some of the variability was explained by the time over which the effluent resided in the treatment process. In a third study it was explored if passive sampling would be a better way to assess average estrogenicity. Indeed, passive samplers identified sources of estrogens, and passive sampling data correlated well with both repeated grab sampling and bioaccumulation data. Subsequently, passive samplers were deployed across Switzerland in a fourth study. It involved 22 effluent discharges and the associated rivers. Data analysis of the last study is still ongoing, preliminary observations are discussed here. Although a lot could be learned from repeated grab sampling campaigns, it emerged that passive sampling is a very effective and appropriate technique to assess: i) effluent treatment efficiency, and ii) the chemical load of river water. For these reasons it is a valuable monitoring tool for water quality criteria assessments as well as for studies that aim to link exposure and effect. The passive samplers that are currently available not only target estrogenic substances, but many other polar organic compounds of concern, such as antibiotics (see NRP49), other pharmaceuticals and biocides