Defined Approaches to Identify Potential Eye Irritants

Potential eye irritants can be identified without animal testing using defined approaches to testing and assessment. A defined approach consists of:

• Input data generated from identified methods.
• A data interpretation procedure (e.g., machine-learning model, flowchart, or decision tree through which the data are evaluated).

In June 2022, OECD issued Test Guideline No. 467, Defined Approaches for Serious Eye Damage and Eye Irritation. This guideline describes two defined approaches that can be used to test and assign eye irritation hazard classifications to nonsurfactant chemicals.

More about defined approaches to testing and assessment

NICEATM collaborated with the PETA Science Consortium International e.V., EPA, and member companies of the trade association CropLife America to develop in vitro defined approaches for classification of eye irritation potential of agrochemical formulations based on the GHS and EPA hazard classification systems. This public-private partnership was established as a means to develop confidence in nonanimal methods for identifying potential eye irritants and thus encourage adoption and use. A three-phase, multilaboratory evaluation assessed the applicability of different in vitro eye irritation test methods to this product class. A report published in 2021 describes generation and evaluation of data to support Phases 1 and 2 of the study; the results of Phase 3 testing were published in a January 2025 addendum to the report. A publication by van der Zalm et al. (2023) describes potential usefulness of defined approaches for classification of agrochemical formulations based under the EPA system. A forthcoming paper (Daniel et al. in preparation) will consider the Phase 3 data in a discussion of the potential usefulness of defined approaches to predict GHS and EPA classifications for ocular irritation potential of agrochemical formulations.

Details of the Study

NICEATM initially analyzed data from four in vitro assays paired with in vivo rabbit eye test data to determine if a defined approach using a combination of these assays could assess eye irritation potential. While the results of this analysis indicated that developing such a defined approach was feasible, it was determined that additional prospective testing was needed. Members of the collaboration team designed a multiphase prospective evaluation to (1) assess the applicability of selected in vitro eye irritation/corrosion methods to agrochemical formulations and (2) develop a defined approach for agrochemical formulations testing for prediction of U.S. and international irritancy classifications.

In Phase 1 of the study, six formulations classified by on the in vivo rabbit test as nonirritating (i.e., EPA Category IV; GHS Not Classified) or severely irritating (i.e., EPA Category I; GHS Category 1) were evaluated in seven in vitro or ex vivo methods. Ten additional formulations representing the full range of eye irritation classifications were tested in Phase 2. Based on an assessment of the results and considering the relevance of each method to humans, four in vitro or ex vivo assays were selected to proceed with Phase 3 testing of an additional 12 formulations classified based on the in vivo rabbit test as moderately or mildly irritating (i.e., EPA Category II or III; GHS Category 2A or 2B). Additionally, a subset of 13 formulations was tested in a fifth method. A total of 29 formulations have now been tested in as many as five methods:

• Bovine corneal opacity and permeability with histopathology (as described in OECD Test Guideline 437)
• EpiOcular EIT (as described in OECD Test Guideline 492)
• SkinEthic time-to-toxicity (as described in OECD Test Guideline 492B)
• In vitro depth of injury
• EyeIRR-IS

Supplemental Information file: Phase 3 testing results for all formulations (January 2025)

Data generated were used to analyze alignment of predictions across nonanimal assays and the rabbit test. Majority EPA predictions were achieved for 25 of 29 formulations (based on alignment among at least 2 of 3 methods). Majority GHS predictions were achieved for 27 of 29 formulations (based on alignment among at least 3 of 5 methods). These data will support ongoing work to develop a defined approach for assessing eye irritation potential of agrochemical formulations (van der Zalm et al. 2023; Daniel et al. in preparation).