Comparing and evaluating alternative (in vitro) tests on their ability to predict the Draize maximum average score.

Area of Science:

• Toxicology
• In vitro testing
• Regulatory science

Background:

• The Draize eye irritation test is a standard in vivo method.
• Alternative in vitro methods are being developed to reduce animal testing.
• The Cosmetic, Toiletry, and Fragrance Association (CTFA) initiated a program to evaluate these alternatives.

Purpose of the Study:

• To assess the effectiveness and limitations of various in vitro eye irritation tests.
• To determine the correlation between in vitro test results and in vivo Draize test data.
• To identify in vitro tests with sufficient predictive power for cosmetic and personal-care products.

Main Methods:

• Multi-phased study comparing in vivo (modified Draize test) and in vitro protocols.
• Testing of experimental formulations across three product types.
• Statistical analysis including concordance analysis and regression modeling.
• Evaluation of prediction intervals to measure in vitro test precision relative to Draize scores.

Main Results:

• In vitro tests were ranked based on their ability to predict Draize Maximum Average Score (MAS).
• Prediction interval widths, indicating prediction uncertainty, varied with the level of irritancy.
• In Phases I and II, intervals widened with increased irritation; Phase III showed wider intervals in the middle range.
• The primary source of variability (70-95%) was the deviation between observed and predicted Draize scores.

Conclusions:

• Selected in vitro endpoints showed similar average prediction interval widths.
• Differences in predictive ability often occurred at the low and high ends of the irritation range.
• Increasing replicate tests minimally improved overall prediction precision.
• The study provides a framework for evaluating the predictive capacity of in vitro eye irritation tests.