74O JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS To be suitable, the tests should take into account the actual part of the human body to which the cosmetic will be applied, the maximum concentra- tion of the substance likely to be present in the cosmetic, and the greatest frequency at which the preparation might be used by an individual. As in most toxicological studies, initial safety evaluation tests are carried outusing an animal species and only the confirmatory tests are conducted in man. Substances which irritate the skin are of three types, namely: (a) Primary irritants. (b) Skin fatigue inducing agents. (c) Contact sensitizing agents. A primary irritant, as defined by Fisher (2) is a substance which produces inflammatory changes in the skin as a result of a direct damaging effect. An alternative definition is that it is a substance capable of producing damage without the mediation of an antibody system (3). The response is an extreme reaction of the tissue to injury, which is not physiological and is characterized by the presence of inflammation, vesiculation and necrotic changes at the contact site. Some substances may not produce visible damage in the skin after a single application but, when administered repeatedly, irritation is produced. Such an agent brings about a state of 'skin fatigue', a condition which is characterized by hyperirritability, local hardening and thickening of the skin (2). Contact sensitizing agents are distinguished from the two classes of irritants mentioned above, in that they produce epidermal damage which is mediated by allergic changes in the tissues (3). Often the cutaneous response occurs some time after the original contact and at a site remote from the point of application of the sensitizor. This response is known as 'delayed hypersensitivity'. Since this type of reaction is not relevant to this review, it will not be discussed further. In cosmetics, strong irritants are rarely a problem since they are readily detected by animal tests and consequently would be rejected without further consideration. However, weak irritants present a greater problem, since they may not be detected either in the initial animal tests or in sub- sequent clinical trials. Once a product has been marketed, information on minor reactions is difficult to obtain since users tend to revert to using an alternative preparation, without making a complaint to the manufacturer. Cosmetics and toilet preparations were classified by Rostenberg (3) under the headings--hair and eye cosmetics face powders, creams and lipsticks nail cosmetics antiperspirants and deodorants depilatories

APPRAISAL OF METHODS FOR DETFCTING PRIMARY SKIN IRRITANTS 741 perfumes, sunburn preventitives and special creams and lotions. Some of these products remain on the surface of the skin for prolonged periods, whereas others, such as shampoos, are rinsed off quickly. Some of the problems encountered in testing for primary irritants were mentioned by Fisher (2), Rostenberg (3, 4) and Rieger and Battista (5). In a brief discussion of the many methods available to a dermatologist evaluat- ing drugs for use on the skin, Calnan (6) referred to four possible approaches: (1) An examination of the gross macroscopic appearance of the skin. (2) A determination of the effect of the substance on separate elements of the skin, e.g. keratinization, sweating, hair growth and pigmentation. (3) An observation of the structure of individual cells. (4) An investigation of biochemical pathways in the skin. These parameters of investigation are applicable also to a safety testing programme for cosmetics. In this review, the tests frequently employed in the safety evaluation of compounds used in skin preparations are discussed with reference to the various modifications which have been suggested for overcoming the difficulties experienced in using them. STANDARD TESTS The patch test, as originally described in 1895 by Jadassohn (7) and later adapted by Bloch (8, 9) and by Draize, Woodard and Calvary (10), appears to be the most widely adopted test for detecting primary skin irritants (1). The technique is simple and has been found to give reproducible results with a wide range of materials both in animal tests and in clinical trials. In practice, initial tests for the irritancy of substances are performed in animals and then, if warranted, tests on human subjects are carried out. Animal patch tests The patch test as described by Draize et al (10) involves single applica- tions of the test substance to the shaved skin on the back of albino rabbits. Accurately measured quantities (0.5 ml or 0.5 mg) of the substance are applied to the skin under 25.4 mm square gauze patches which are secured with adhesive tape. To prevent any volatile constituents from evaporating, the trunks of the rabbits are wrapped in rubberized cloth. The animals are placed in restrainers and the patches left intact for 24 h. This procedure ensures that the skin is in contact with the test substance for the entire period