Assessment of the chorioallantoic membrane vascular assay (CAMVA) in the COLIPA in vitro eye irritation validation study

Toxicol In Vitro. 1999 Apr;13(2):285-93. doi: 10.1016/s0887-2333(98)00089-7.


The chorioallantoic membrane vascular assay (CAMVA) is an alternative to the Draize rabbit eye irritation method. The CAMVA employs the vascularized membrane of a fertile hen's egg to assess eye irritation potential. This irritation potential is a function of alterations in the vasculature following the administration of test material. Because of the history of use of the CAMVA it was selected as one of the methods for a validation study organized and sponsored by COLIPA. For this validation study mathematical prediction models (PMs) were developed to convert the CAMVA results into predicted Draize eye irritation scores known as a modified maximum average Draize score (MMAS). These predicted scores were statistically compared with the observed scores to assess the relevance of the CAMVA. The assay was conducted on the same set of test materials by two independent laboratories. These two sets of data were compared to assess the interlaboratory reproducibility of the assay. The results of this validation study of the CAMVA show that for test materials with MMASs in the 0 to 5 range or the 55 to 110 range, the CAMVA did not give a good prediction. The predictions were better for samples of mild to moderate irritation (MMAS 5-55). The difficulty in predicting at the low end of the irritation scale appears to be due to the biological variability of the test system and the subjective nature of the CAMVA evaluation. For those samples with an MMAS above 55, the CAMVA appeared to be limited in demonstrating the more severe response. This may be due to the fact that the PMs were developed using historical data sets of test materials with MMASs below this range. Two approaches for improving the CAMVA for eye irritation prediction are (1) to decrease the variability at the low end by reducing the subjectivity in the scoring and (2) to develop better prediction models using more data in the range of severe irritants.