742 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS of treatment, and that the patches are not dislodged by the movement of the animals. In addition to applying the patches to areas of intact skin, patches are applied to areas of skin which have been abraded by scoring through the stratum corneum, but not sufficiently to cause bleeding. Test areas are examined at 24 and 72 h after treatment and scores allocated according to the degree of erythema, oedema and eschar formation. The average of the combined scores given for the areas of intact and abraded skin is referred to as the Primary Irritation Index. The test as described has been widely adopted and found to give useful information for placing sub- stances in order of their irritancy. The use of positive and negative controls is essential in animal tests (11, 12). Sodium lauryl sulphate is regarded by some workers as a satisfactory positive control for primary irritants, since it does not produce an allergic reaction (13, 14). Although some workers believe that sufficient evidence is available to indicate that tests of this type are suitable for showing the irritancy of substances to the human skin (15), their views are not generally accepted for the reasons discussed below. A disadvantage of the test as recommended, is that the results rely entirely on subjective assessments of the appearance of the test sites. No objective measurements are made nor is any microscopic examination of the skin carried out to detect changes which are not visible to the naked eye. Schmid (16) suggested that alterations in the thickness of the skin might be taken as an indication of the degree of irritations. He indicated that histo- logical studies might be carried out to substantiate this conclusion. Where it is possible to examine areas of treated skin histologically, a more detailed assessment of the damage caused by an irritant can be made (17). The standard patch test as outlined above, has been modified on numer- ous occasions, but in general, these modifications concern the size of the patches, the quantity of test substance applied to the skin, the duration of the exposure period, and the animal species employed (18, 19). The degree of occlusion and the pressure with which the patch is applied are other variable factors which have been investigated as possible sources of error (2, 20-25). A semi-occlusive patch test described by Holland (20) resembled the Draize test in having the test materials applied on absorbent pads, but the rubberized cloth or Cellophane used to prevent any volatile materials escaping, is omitted. Using detergents, these authors obtained different scores from those given by the standard patch test. A pressure patch test was devised by Fernstrom (26, 27). This patch consisted of an 8 x 8 mm square of blotting paper impregnated with the test substance which was applied to the skin under a sheet of rubberized cloth. On this

APPRAISAL OF METHODS FOR DETECTING PRIMARY SKIN IRRITANTS 743 was placed a plastic sponge and the whole secured with bandages. Rosten- berg (22) suggested that in such a test the skin will be kept in contact with the test substance more consistently than if the standard test is used. Despite any advantages that this method may have in animal tests, Bett and Calnan (28) considered that it offered no improvement over the standard procedure in clinical trials. Another modification, aimed at improving the contact between the skin and the test compound, is the cup technique (29, 30). Cups made of impervious material are filled with the test substance (liquid) and applied to the skin. Rostenberg (22) reported that this method was particularly satisfactory for screening substances for irritancy in human tests but it had the disadvantage of being more cumbersome and slower than the standard procedure (31). The reproducibility of the results obtained using this technique is not known. False positive reactions in the occluded patch test have been related to the effects of the patch on the microclimate of the skin at the test site (4) and the irritancy caused by the patches themselves. Although in safety testing, it is the aim of the experimentalist to exaggerate the conditions under which the test substance ultimately will be used, Idson (25) holds the view that the conditions used in patch testing are too stringent and the results obtained likely to be misleading. With occlusive patches, the skin is under duress and is unable to function normally, in terms of sweat evapora- tion, cooling and cutaneous respiration (4, 32). Furthermore, volatile solvents are prevented from evaporating as they would do in normal use of the product (32, 33) and may modify the irritant response (34). A reduc- tion in the evaporation of sweat may result in an increased hydration of the skin (35), an effect which is likely to increase skin permeability and thus increase sensitivity. An example of misleading reactions given by patch tests was provided by Holland (20), who exposed volunteers to solutions of detergents and found very different results using occlusive and non-occlusive conditions of testing. These observations emphasize the importance of testing substances under conditions which resemble the intended conditions of use (25). The components of the patches themselves have been shown to bring about changes in the skin, irrespective of whether irritants are present or not. Rubberized fabric, adhesive plaster and other materials commonly used, have been shown to produce changes in the skin which include inflammation, increased hydration and damage to the stratum corneum (36-38). Hence, adequate precautions should be taken to ensure that the materials used in patch tests are of minimal irritancy to the skin.