PREDICTIVE PATCH TESTING 327 first group, the test is generally not continued. Usually three to five groups are tested. With this technic it has been possible to show lack of allergenicity in substances such as ammonium thioglycolate and ammo- nium bisulfite, which produce few allergic reactions in consumer use, while 50-100% of subjects reacted with potent sensitizers such as mer- captohydrazides, mercaptoethane sulfonic acid, etc. The latter agents produced a few sensitization reactions in laboratory personnel who had only casual contact with them during testing. Thioglycerol, which has produced a moderate number of sensitization reactions in consumer use, sensitized about 10% of the test subjects. This test is based on the following concepts: First, for any given sub- stance there is a certain upper limit in the number of individuals who will become sensitized to it. Secondly, that one of the factors which deter- mine the completeness with which this potential index of sensitization is exploited is the number of contacts. It is not necessary to keep up ap- plications indefinitely, since after a certain number of contacts there is a diminishing return of newly sensitized individuals. Calnan, Epstein, and Kligman (5), for example, report that in animals, ten injections are optimal for sensitization fifteen are no better and five give fewer reac- tions. In experimental sensitization of humans with Krameria and monobenzone (monobenzylether of hydroquinone), they found four weekly exposures gave more reactions than three. They suggest that with very weak sensitizers more exposures may be needed. Subsequent to the report by Brunner and Smiljanic, other predictive tests for sensitization were described by Shelanski (6), Draize (7), and Traub and his co-workers (8). As in the Brunner-Smiljanic procedure, all these tests involve multiple contacts, but the factors of occlusion, site, duration and area of contact are varied. The pertinent factors in each test are shown in Table I. The Traub test obviously has only limited predictive value since it consists of three weeks of use application in a group of only 200 subjects without intensification of exposure conditions. Also, one cannot repetitively apply many substances, such as cold wave formulations or hair dyes exactly according to use in the three-week test period. One the other hand, in the Shelanski test occlusive repetitive applications to the same site give a higher yield of sensitized subjects but tend to magnify the irritant effects of the test formulation. The result- ing so-called "skin fatigue," due to summation of irritations under the test conditions, may make the differentiation between sensitization and irritancy more difficult. The Draize test also uses occlusion to increase the yield of sensitized individuals but varies the application site so that

328 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS + o + + o