Registration Dossier

Ecotoxicological information

Short-term toxicity to fish

Currently viewing:

Administrative data

Link to relevant study record(s)

Description of key information

In the absence of experimental data, the acute toxicity to fish has been assessed using a QSAR approach. The QSAR approach works within the spirit of REACH and the stated aim of the legislation to reduce animal testing where possible. Three predicted values are used in a weight-of-evidence approach; one using a category formed from the aquatic toxicity databases available in the QSAR Toolbox 3.4.0.17 (96h LC50 = 1.5 mg/L) and the other two using ECOSAR v1.11 (Neutral Organic SAR, 96h LC50 = 0.628 mg/L; Ester SAR, 96h LC50 = 0.759 mg/L). The prediction results are considered reliable according to OECD principles. The prediction models are valid and the target substance falls within their applicability domain. The models are independent from each other (e.g. based on different training sets). Thus, the agreement among the predictions increases the confidence in the reliability of the predictions. For the purposes of the chemical safety assessment, the lowest predicted value of 0.628 mg/L has been chosen.

Key value for chemical safety assessment

Fresh water fish

Fresh water fish
Effect concentration:
0.628 mg/L

Additional information

The QSAR predictions were performed on the main constituent of Oxyoctaline formate, which is 2,4a,5,8a-tetramethyl-1,2,3,4,4a,7,8,8a-octahydronaphthalen-1-yl formate (concentration range 75-85%). Oxyoctaline formate contains an isomeric impurity, 2,5,5,8a-tetramethyl-1,2,3,5,6,7,8,8a-octahydronaphthalen-1-yl formate (CAS 132346-28-2, concentration range 10-23%), which is predicted to have the same 96-h LC50 as the main constituent since it is assigned to the same ECOSAR classes and has the same estimated log Kow. The registered substance also contains traces of the corresponding alcohol (0.1 to 2%). This impurity is expected to be less toxic that the ester components because of its lower uptake potential (as modelled by log Kow, estimated Kowwin v1.68 value = 4.315) and lack of any excess toxicity (i.e. it is assigned only to the neutral organic class by ECOSAR). The estimated 96-hr LC50 is 1.427 mg/L. Therefore, the predicted value for the mono-constituent component is considered a suitable representation for the acute toxicity of Oxyoctaline formate to freshwater fish. The predicted fish 96h LC50 values using ECOSAR are 0.628 mg/L (Neutral Organic SAR) and 0.759 mg/L (Ester SAR). The predicted values can be considered reliable taking into account that predicted and experimental data for chemicals from the "Ester" training set, which are close neighbours to the target in the Log Kow descriptor space, are in close agreement. The 4 identified analogues were predicted to have LC50 values of 0.716-1.036 mg/L (Neutral Organic SAR) and 0.911-1.113 mg/L (Ester SAR), which is comparable to the experimental values of 0.48 to 5.53mg/L. The applicability of both models in the assessment of the aquatic toxicity of Oxyoctaline formate is further supported by a comparison of experimental and predicted values for the daphnia and green algal endpoints.

In exploring the using of the OECD Toolbox for the prediction of the acute toxicity to fish, a trend analysis was performed on several fish species. The target is classified as an Ester by the ECOSAR profiler. Thus a category was formed from chemicals classified as Esters present in the acute aquatic toxicity databases publicly available in the OECD QSAR Toolbox (Aquatic ECETOC, Aquatic Japan MoE, Aquatic Oasis, ECHA, ECOTOX). The category consisted of 1478 esters. The fish species with the most 96h LC50 values were chosen for subsequent trend analysis: Brachydanio rerio (75 chemicals), Danio rerio (49), Oncorhynchus mykiss (257), Oryzias latipes (45) and Pimephales Promelas (154). The same sub-categorisation approach was applied in all cases. First, chemicals were excluded if their EC50 value exceeded the water solubility or if the EC50 value was reported as > x mg/L (i.e. exceeds highest test concentration). Then analogues were eliminated that were assigned additional ECOSAR classes, assigned additional MOAs by OASIS or have a protein binding alert by OECD (i.e. analogues that might act via additional mechanisms). The number of analogues in the final categories were 5 (Brachydanio rerio), 11 (Danio rerio), 21 (Oncorhynchus mykiss), 4 (Oryzias latipes) and 27 (Pimephales Promelas). The regression equations derived for the first four species were not applicable to the target chemical because the latter was outside the log Kow domain; the extrapolated LC50 values ranged from 0.511 to 3.12 mg/L but are considered unreliable because the target is OUT OF DOMAIN. Therefore, only the reliable prediction for pimephales promelas was retained which gave an estimated LC50 of 1.5 mg/L.