Performance of porcine corneal opacity and permeability assay to predict eye irritation for water-soluble cosmetic ingredients.

The purpose of this paper is to report on the ability of an in-house porcine corneal opacity and permeability assay (PCOP) to predict eye irritation for cosmetic ingredients. Preliminary studies showed that the PCOP assay could accurately predict eye irritation class for liquid and water soluble materials. To broaden our experience a larger study on 50 cosmetic ingredients of this group was conducted. A prediction model (PM) was obtained based on only one endpoint-permeability measured after 30-min exposure O.D.30. This PM allows to distinguish nonirritating compounds (if O.D.30 < 0.35) from irritating (if O.D.30 > or = 0.35). Forty-nine of the 50 ingredients tested in the PCOP assay were accurately classified. The agreement was high (concordance 98%-kappa = 0.96). For 43 of the test substances an equation PM was obtained to predict the MAS. Despite satisfactory statistical coefficients this algorithm is not recommended due to wide 95% confidence intervals. These results confirm the usefulness of the PCOP for water-soluble cosmetic ingredients to discriminate nonirritants (MAS < or = 15) and irritants (MAS >15). For this type of ingredients the PCOP seems to be better than the BCOP to predict irritation class. Future work will be done to compare the BCOP and PCOP performances and to develop an appropriate protocol for water insoluble compounds.