Choksi NY, Haseman JH, Allen DG, Tice RR, Stokes WS. Estimation of the underprediction rates for the in vivo rabbit dermal irritation assay. Poster presented at Society of Toxicology Annual Meeting, Baltimore, MD, March 2004.
Alternative in vitro test methods proposed to substitute or replace an in vivo test method should provide equivalent or improved protection of human or animal health in order to gain regulatory and general acceptance. The ICCVAM and NICEATM are collaborating with the ECVAM to conduct a validation study of three in vitro dermal irritation assays. To assess the acceptability of these in vitro assays, an effort was undertaken to estimate the false negative rate of the in vivo test as defined by its ability to consistently identify irritants, mild irritants, and non-irritants according to the Globally Harmonized Classification Scheme. Data for 187 substances was obtained from the ECETOC database for skin irritation and corrosion. The distribution of rabbits with mean erythema or oedema scores of <1.5, between 1.5 and 2.3, or >2.3 was determined for each of the substances classified as negative, mild irritant or irritant. Since the true classification of each substance is unknown, a simplifying assumption was made that the results are correct for substances tested once only. For multiple-tested substances, the classification obtained from a majority of the studies was used. The analysis indicated: (1) the likelihood of a mild irritant being under-classified as a non-irritant was <5 percent when based on all substances and <10percent when based on multiple-tested substances, (2) the under-classification rate of irritants as non-irritants was <1 percent, and (3) the under-classification rate of irritants as mild irritants ranged from 9-30percent, depending on whether all substances or only multiple-tested substances were considered. Additional in vivo irritation data for studies using currently accepted procedures was requested from US federal agencies and industry. Appropriate data received will be added to the database and the false negative analysis refined. This evaluation emphasizes the need for high quality in vivo dermal irritation data that can be used to assess the performance of proposed new alternative test methods. ILS staff supported by NIEHS contract N01-ES-35504.