Changing concepts of ocular irritation evaluation: pitfalls and progress.

A series of studies has been carried out to evaluate methods of modifying existing ocular test protocols with the aim of reducing the numbers of animals used and lessening the severity of the response with the minimum effect on the predictive value of the tests. Initial studies showed that when the dose administered is reduced from 100 to 10 microliter, the severity of the response is reduced but the rank order of severity of a series of test compounds is unchanged. This series of studies also demonstrates that the greatest source of variability in obtaining irritation scores is in the rabbit population and that the numbers of rabbits used per test can be chosen on the basis of the required degree of accuracy. In a related study it is shown that direct corneal application of the test material (10 microliter) increases the corneal response but has no effect on the conjunctival response in comparison with reactions after application in the conjunctival cul-de-sac. Finally, a retrospective study was performed to determine whether severe skin irritation is predictive of a similar level of response in the eye. Of 60 compounds known to be severe skin irritants, only 39 are severe eye irritants whereas 15 are mild to non-irritants. The remainder were found to cause moderate responses in the eye. These data demonstrate that existing ocular irritation test protocols can be modified to use fewer animals under less extreme conditions without significant compromise of predictability. However, dermal irritation, determined using current test protocols, does not appear to be directly predictive of ocular irritation potential.