|
(79-08) Comparison of a mechanistic and a statistic method to evaluate standard toxicity tests with eight alkylated phenols in perspective of mixture toxicology. Smit, Mathijs*,1, Cova, Alberto2, Buffagni, Melania2, 1 TNO-MEP, Den Helder, Netherlands2 ENI Agip, Milano, Italy ABSTRACT- EC50 and NOEC values are estimated from the results of standard toxicity experiments using different statistical evaluation methods and assuming a constant exposure level. These statistical methods assume that the obtained data can be described by a certain pre-defined set of distributions (log-normal, log-logistic, Weibull, etc.), from which the end points are estimated. The DEBtox model developed by Kooijman and Bedaux (1996) follows a different approach. Derived from the DEB theory (Dynamic Energy Budgets) this model defines simple mechanistic descriptions for the toxicological processes, which are likely to occur when organisms are exposed to toxicants. From the experimental results, parameters are estimated to calibrate these mechanistic descriptions. This also includes the decrease of the exposure concentration as a result of degradation. The obtained descriptions can be used to calculate the effect at any point at given time and (variable) exposure. Both methods are applied and evaluated on the basis of results of standard acute toxicity experiments with the marine zooplankton species Acartia tonsa, exposed to single alkylated phenols and exposed to a mixture of eight alkylated phenols. The methods result in different values for the NOEC and EC50. The results from the experiments with single components were used to predict the toxicity of the mixture in terms of concentration addition, effect addition and stress addition. The predictions were later tested by the results from a mixture experiment by applying both the mechanistic and the statistic approach. Conclusions concerning the applicability and the pros and cons of both methods will be drawn Key words: effect modeling, DEBtox, Alkylated phenols, mixtures |
