Implications of asymmetric concentration–response relationships on effect concentration estimation: a case study regarding the chronic toxicity of NaCl to <i>Ceriodaphnia dubia</i>

An important limitation of concentration-response (C-R) modeling of toxicity test data is the imposition of a shape for this relationship that can deviate from the underlying true relationship, thereby biasing estimates of effect concentrations (ECp's). In particular, the imposed mathematical model is often symmetric, whereas considerable asymmetry might be present in the underlying relationship that is either not evident or not addressed in standard toxicity tests with limited numbers of treatments. To evaluate asymmetry and its implications for ECp estimation, six simultaneous tests of NaCl chronic toxicity to Ceriodaphnia dubia were conducted, providing extensive information on inter-replicate variability and on the shape of the C-R relationship. An asymmetric C-R relationship derived from this large data set was used to simulate data sets with a more typical, smaller configuration, which were subject to C-R analysis using both symmetric and asymmetric models. Both models resulted in substantial uncertainties for estimating ECp's at low p values for this test configuration. There is a need for more work and care regarding the use of C-R analysis in developing effects assessments for low levels of effect.