Reliability of using fewer rabbits to evaluate dermal irritation potential of industrial chemicals.

To investigate the feasibility of using fewer than six animals for dermal irritation testing, a retrospective evaluation was conducted on data from 224 studies performed within the last 10 years using standard methodology. For each study, an irritation classification was made from primary dermal irritation indices calculated using irritancy data available for five, four, or three rabbits randomly selected from the original six-rabbit pool. The original six-rabbit analyses resulted in classification of 49 non-irritants, 75 negligible irritants, 66 mild irritants, 24 moderate irritants, and 10 severe irritants. Agreement with the six-rabbit classification was obtained for five-, four-, and three-rabbit groups, respectively, as follows: for negligible irritants, 95, 85, and 69%; mild irritants, 85, 80, and 76%; moderate irritants, 88, 88, and 71%; severe irritants, 100, 90, and 90%. Differences in the dermal irritation indices between groups of six and five, four, and three rabbits were small, with an average difference no larger than 0.3. When disagreements in classification occurred, the test material was most often classified less severe. Using 90% or greater agreement as acceptable criterion for reducing the number of test animals, our results suggest that the use of less than six rabbits would not be suitable for evaluating dermal irritation based solely on a qualitative classification. However, the use of as few as three rabbits would be adequate to evaluate dermal irritancy based on quantitative primary dermal irritancy values.