Evaluation of eye irritation potential is essential to ensuring the safety of cosmetics and other personal care products.Among others, in vitro three-dimensional human corneal models have been adopted as a part of testing strategies.EpiOcular tissue model, for example, has proven to be a useful tool to partially replace Draize eye irritation test.The long history of reproducibility and proven utility for ultra-mildness testing make EpiOcular a particularly useful model for cosmetics.However, there is a fundamental lack of accessibility to the three-dimensional tissue models outside their country of origin due to individual customs regulation, transport constraints or other reasons.Even after the similar tissue models become available, additional challenges still arise regarding intrinsic differences inherent to their tissue of origin and manufacturing process, and transferability of the test protocol from one model to another.BioOcular, a three-dimensional human corneal epithelium construct, has recently been developed in China as an equivalent of the EpiOcular model.Comparative studies of the two tissues models for predicting eye irritation potential of cosmetic products have not yet been reported.To examine the suitability and performance of the BioOcular model for predicting eye irritation potential of cosmetics, ten blind coded facial and eye area products were tested by using BioOcular or EpiOcular assay at two different laboratories, i.e., Shaanxi Biocell Biotech Co.(Xian, China, for the BioOcular assay) and Institute for In Vitro Sciences (Gaithersburg, MD for the EpiOcular assay).The same, time to toxicity protocol was chosen for both assays.Briefly, formulations were applied neat or in diluted form (10% for Sample #10 only) to tissue construct along with controls for 0, 4, 8, 16, and 24 hours and the exposure time required to reduce the tissue viability to 50% (ET50) was determined using the MTT cytotoxicity assay.The ten cosmetic products showed ET50 values of 1422, 1014, >1440, 354, >1440, >1440, 1146, 198, 138, and 126 minutes in the BioOcular assay and 1176, 870, 780, 540, >1440, 1200, 565, 214, 120, and 28 minutes in the EpiOcular assay, respectively.The products were all considered "non-irritating" based on the ET50 obtained from either test.The ET50 values were highly consistent with the low levels of irritation potential observed in the consumer studies and market history of the respective products.Interestingly, the ET50 values obtained from the BioOcular assay had a high level of agreement with that obtained from the EpiOcular assay (intraclass correlation coefficient =0.8444), indicating a high interchangeability between the two methods.It should be noted that BioOcular assay tended to yield slightly longer ET50 than EpiOcular assay for given samples.Compared with EpiOcular assay, BioOcular assay showed a longer average ET50 (872 vs.693 minutes) and a higher ratio of samples with longer ET50 (7/10 vs.2/10) for the 10 samples tested and positive control (0.3% Triton X-100) as well (45 vs.27 ± 6 minutes).Larger studies are needed to verify this finding and factors involved in the differences.Nevertheless, use of the BioOcular tissue model shows promise as an in vitro assay to assess the ocular irritation potential of cosmetics.It is therefore concluded that the BioOcular model is suitable for predicting eye irritation potential of cosmetics.