Validation d'une méthode d'analyse de composés ½strogèniques (incluant les formes libres et conjuguées) dans diverses matrices aqueuses

International audience; During the last fifty years, consumption of estrogenic compounds (natural hormone or synthetic analogue) for human medicine contraception, management of menopausal and post-menopausal syndrome, physiological replacement therapy in deficiency states and treatment of prostate and cancers) and animal farming (developers of single-sex populations of fishes in aquaculture) has considerably increased. Estrogenic compounds are usually not entirely metabolized and reach aquatic environment mainly via effluents of wastewater treatment plants. The lack of knowledge on toxicity of these compounds and their impact on ecosystems and human health has raised public concern about their occurrence in the environment. For this reason, highly sensitive and selective analytical method is needed to identify and quantify these emerging contaminants in various environmental compartments. The aim of this communication is to show the different steps for validation of an analytical method for five estrogenic compound in aquatic environments (estrone [E1], 17aestradiol [17a-E2], 17bestradiol [17b-E2], 17aethynylestradiol [EE2] and estriol [E3]) chosen because of their strong endocrine-disrupting potency in river and wastewater. The sample pretreatment consists of an extraction on Oasis HLB cartridges followed by purification on Florisil cartridges. Four stable isotopically labelled estrogenic compounds were used as surrogate standards to overcome matrix effects during the quantitation. Liquid chromatography coupled to electrospray tandem mass spectrometry is performed in the multiple-reaction monitoring mode and molecule identity is confirmed by two quantitation and confirmation transitions according to the EEC decision 2002/657/CE. Stability of these compounds in extracts from Oasis HLB (stored at ambient temperature during 3 days) or in mobile phase (stored at 18°C during 1 month) have been verified. Yields and repeatability have been evaluated in three kinds of matrices at two contamination levels by quintuplate analysis. Yields higher than 75% (except for 17a-E2 at low level: in the range of 60%) and relative standard deviation lower than 20% (except for E3 in influent: at 36%) have been obtained. Reproducibility has also been studied from real matrices. Method reaches limits of quantification around 1 ng/L in waters from rivers and wastewater treatment plants. Specificity has been statistically verified for the five analytes, thus we can consider that matrix effects are corrected by the use of deuterated surrogates.