Not yet nixing rabbit eyes from the six: the TTT is not predicting the full range of in vivo eye irritation classifications for agrochemical formulations

Since its adoption as a test guideline (EIT, OECD test guideline (TG) 492) in 2015, the EpiOcular™ eye irritation test has been the in vitro method of choice to identify agrochemical formulations non-irritant to the eye (neither category 1 nor 2, according to the Globally Harmonized System of Classification and Labelling of Chemicals, GHS). For serious eye damage (GHS category 1), however, none of the in vitro test methods including the bovine cornea opacity and permeability test (BCOP, OECD TG 437, and protocol modifications), the isolated chicken eye test (ICE, OECD TG 438), the ocular irritation assay (OECD TG 496) or the hen's egg test on the chorioallantoic membrane (HET-CAM) assay proved sufficiently sensitive to correctly identify agrochemical formulations classified as seriously eye damaging in vivo. In 2022, the OECD adopted a new in vitro test guideline for identifying chemicals inducing serious eye damage, eye irritation, and chemicals not requiring classification for eye irritation (OECD TG 492B). However, within the formal validation study of this time-to-toxicity (TTT) test method, agrochemical formulations were not included. To assess the applicability of this test to agrochemical formulations 25 formulations with historical in vivo data were assessed in the TTT. The TTT correctly predicted 0 %, 90 % and 50 % of the agrochemical formulations which were classified by the in vivo data into the three categories 1, 2 and “not classified”, respectively. The overall accuracy of the TTT to predict agrochemical formulations of the three categories was 52 %. Therefore, to facilitate non-animal-based identification of the full range of ocular irritation of agrochemical formulations, testing protocols and/or prediction models must be revised or classification guidelines should be adjusted to align with human relevant in vitro testing rather than relying on in vivo Draize data as a reference point.