COMPARISON OF NOEC VALUES TO EC10/EC20 VALUES, INCLUDING CONFIDENCE INTERVALS, IN AQUATIC AND TERRESTRIAL ECOTOXICOLOGICAL RISK ASSESSMENT. CALL REFERENCE: GP/EFSA/PRAS/2013/01

Year: 2015
Category: Scientific articles
Language: English
Authors: Flavio Marchetto and Francesco Galimberti, Giovanna Azimonti, Luca Menaballi, Sonia Ullucci, Federica Pellicioli, Alessandra Caffi, Lidia Ceriani, Alessio Ippolito, Waldo de Boer, Hilko van der Voet, Angelo Moretto
Open the attachment

Ecotoxicological studies performed for the authorization of plant protection products (PPP) usually result in the reporting of endpoint values in terms of effect concentration (EC) affecting a percentage x of test organisms or where a x percentage of an effect is observed (ECx). The new Regulation (EC) No. 1107/2009 for the authorization of PPPs and the related data requirements provide that ecotoxicological endpoint data from chronic or long-term studies submitted by the Applicant are reported as EC10 or EC20 values together with the NOEC. NOEC values have been criticized since their values strongly depends on the experimental study design, whereas ECx values take into account the whole concentration-response curve and are therefore considered more appropriate. The aim of the project is to investigate the comparability of the ECx approach to the current NOEC approach on a larger data sets in view of the new Regulation requirements. Ecotoxicological data gathered from 70 active substances’ approval dossiers were collected and stored into a MS Access database. All the extracted ecotoxicological data were analyzed in order to derive NOEC and calculate EC10, EC20, EC50 with confidence intervals, using statistical models from the exponential and Hill families for continuous data, and logistic, log-logistic and complementary log-log models for quantal data. The optimal model was selected based on likelihood ratio tests and the Akaike Information Criterion. ECx/NOEC ratio distributions were calculated considering the whole set of data and model outputs; data were grouped in different categories to remark any differences in the ECx/NOEC ratio distributions.


Back