184 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS small. Selection of atopic subjects for test panels is sometimes considered to improve predictive value but the evidence to indicate that atopics show enhanced susceptibility to topical allergens in general is questionable (16). Bearing in mind the objections to grossly exaggerated exposure testing and recognizing that sensitization testing at normal use-concentration often yields negative results that cannot be interpreted or which may be un- reliable, there are certain attractions in devising a test procedure to enhance responses to use-concentrations and to ensure that positive results are obtained even with moderate or mild sensitizers. Kligman (17) in proposing his 'maximization' test, offered a procedure giving 24/24 positive responses to p-phenylenediamine even though he was still unable to detect some known mild sensitizers such as lanolin. The value of this type of test for cosmetic evaluation does not therefore yet seem to have been established. At the present time, despite the energetic attempts by a number of highly-skilled investigators, there is clearly no satisfactory way of predicting the sensitizing potential of cosmetics and toiletries by means of human volunteer studies nevertheless, such studies carry a definite risk of sensitiz- ing volunteers, possibly for some years (18), and their justification is there- fore doubtful. As an alternative, a reliable test for sensitizing potential using laboratory animals would obviously be helpful. Whereas the response of some other mammals closely resembles the human response to irritants, there are marked interspecies differences in allergic responses. Suitability of the guinea-pig for sensitization testing has been validated to some extent (19) but it would be unwise to expect guinea- pig studies to eliminate any but the most potent sensitizers. Thus, although it is reasonably practical to test for the sensitizing potential of raw materials by conducting challenge tests at elevated concentrations using animals or man, no comparable procedure is yet available for studying complete formulations likely to display no more than mild sensitizing ability. Rather than applying maximizing procedures of dubious predictive value, it is probably better to allow a product to be used normally by gradually increas- ing numbers of individuals. This view takes for granted a prior scrutiny of the raw materials to eliminate any potent known sensitizers and an adequate scheme for monitoring adverse reactions if they are reported by users of the product.

EXAGGERATED EXPOSURE IN PREDICTIVE TESTING 185 CONCLUSION Unreasonable criteria for assessing the potential hazards of topical administration do not necessarily help to protect the consumer. Thus, although animal feeding tests on a proposed food colour may well show that the maximum no-untoward-effect level is several hundred times greater than the expected human intake, even the most harmless materials applied to the skin with such exaggeration are likely to prove injurious. Frazer (1) claimed that the acceptable usage level of a substance--he was referring specifically to food additives, though others have applied his concept more widelymshould be regarded as one-hundredth of the level required to pro- duce significant modification of structure or function in not more than 50•o of a group of test animals. Further, at a dose-level equal to one-tenth of the ED50, no significant changes of any kind should occur. Such margins, however, could not be applied generally to substances coming in contact with the skin or mucous membranes. For example, none of the synthetic anionic surfactants would be acceptable if a shampoo had to be formulated with no more than one-hundredth of the detergent concentration giving threshold irritation in a closed patch test. Nevertheless, present-day sham- poos are used almost universally with minimal known adverse effect a different basis for judging acceptability is therefore needed and one of the possible approaches might be to seek a tenfold margin in relation to experimental findings to allow for individual differences in susceptibility. As already shown, no allowance for interspecies differences need usually be made in irritancy testing. A tenfold safety criterion on these lines may prove quite helpful for the safety evaluation of raw materials but it will seldom be a technically feasible criterion for use in testing formulated products. Direct comparison of a newly-formulated product in a threshold irritancy test with other formulations of similar type, whose effects during normal use are known, will be more appropriate. Such a comparison will certainly give practical guidance on probable safety-in-use. A study on human volunteers will clearly be the most reliable and ideally the study should take the form of a comparison with a known 'safe' and a known 'unsafe' material of similar type (i.e. with 'negative' and 'positive' controls). Comparison with a 'positive' control (e.g. a known irritant) might facilitate quantitative expression of the findings, if a human tolerance test is carried out. In circumstances where laboratory animal studies are judged to be required, it will be equally desirable to conduct these as threshold irritancy tests to forecast the onset of hazard to man in normal use.