SAFETY TESTING OF COSMETICS 139 trations which can be e•nployed without inducing sensitization by finished formulations. Experiences of some investigators, however, have shown some false positive results that is, sensitization seemingly induced in the guinea pig that was not confirmed by subsequent tests in humans. The possibility of such differences in opinions is readily apparent when one considers the variations in procedure which have been employed in the tests on guinea pigs and on humans. There are as many, probably more, variations in the predictive patch test on humans as there are in the tests employing guinea pigs. Recent reviews of the importance of predictive testing and of the use of guinea pigs and humans in predictive testing include one by Griffith and Buehler (3). In this review, some of the findings of Buehler were analyzed. The studies were carried out to determine the relative sensitivity of variations in procedures utilizing guinea pigs. The techniques compared were the intradermal injection method of Land- steiner and Jacobs (4) and a new procedure which employed a closed patch. In this series of studies, it was determined that the closed-patch procedure de- tected sensitizing potential of eight materials known to be contact allergens, while only two of the eight were detected by the intradermal procedure. The application schedule and test material concentration were the same in both series. This difference in sensitivity was confirmed by additional studies in which sensitization of 100% of the animals was obtained by the patch proce- dure with concentrations of dinitroehlorobenzene (DNCB) which failed to induce sensitization of any animals tested by the intradermal technique. The value of the intradermal technique cannot be questioned, but it has limitations. With many materials, the concentrations which can be employed without producing severe primary irritation, including necrosis, are limited to levels somewhat below the practical use level of the material. An equally important factor is the dittleulty in interpreting the reactions to the intrader- real injections. Unfortunately, sensitization reactions in guinea pigs are not characterized by the typical grossly visible papular or vesieular reactions com- monly observed in humans. Judgment must be made on the basis of the extent of erythematous or mfi•imal edematous reactions and comparing the challenge reactions to similar but less extensive reactions to initial injections. When properly applied, however, it is a valuable tool for the detection of strong sensitizers. HUMAN CLINICAL STUDIES After the animal toxicological studies are completed and the product has been shown to be safe for human clinical studies, the latter are then initiated. Clinical studies performed with cosmetics on humans are referred to by our laboratories as proprietary clinical studies to distinguish them from clinical studies which are carried out on ethical drugs.

140 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Patch Tests Two basic procedures have been most frequently employed and numerous variations of these have been used. One of the basic procedures, the Schwartz and Peck (5), has limited value since it will detect only very strong sensitiz- ers. Present consideration of the human patch test will primarily be limited to the repeated insult procedure and its various modifications. Two variations were originally presented-one by Shelanski and Shelanski (6) and the other by Draize (2). Both of fl•ese procedures employ repeated applications lnade over a period of 4 or 5 weeks. They differ in that 10 sensitizing applications are made in the Draize procedure and 15 are made in the procedure of She- lanski. They also differ in that challenge applications are made to the original test sites only in the Draize procedure and are made to the original test site and to previously unpatched sites in the procedure of Shelanski. A short review of these and other procedures employing human panelists has been presented by Kligman (7), who suggested the maximization proce- dure as an alternative. Various methods for achieving the objective of Klig- man's maximization procedure have been employed. We used a procedure in which the skin is abraded prior to sample application. The test materials are applied to abraded sites under patches which are applied to the upper arms. Reactions to the test applications are scored on each Wednesday, Friday, and Monday following applications on the preceding Monday, Wednesday, and Friday. The patches are removed by the panelists 9,4 hours after application. Nine applications are made in 3 successive weeks, and the challenge applica- tion is made on Monday of the sixth week. The challenge applications are made to previously unpatched sites. Abrasions of the test sites are made just prior to sample applications on test days 1, 4, 7, and 11, or preferably, prior to each application. This more fre- quent schedule of abrading can be done when up to four materials are evalu- ated on abraded skin only. The abrasions are in the form of two concentric circles with diameters of % in. and "% in. and are inflicted by a specially con- structed apparatus. The skin areas are cleansed with isopropyl alcohol just prior to abrading. A separate sterile abrader is used for each panelist. All ad- hesive patches and absorbent pads applied over abraded areas are sterilized by exposure to ethylene oxide prior to use. The usual grading scale is utilized. Another procedure employed is to strip the skin by repeated application and removal of adhesive tape from the test sites prior to sample applications. All of these procedures have value in certain applications. The one chosen should provide the required degree of sensitivity as well as data from which valid decisions can be derived. A variation of the repeated insult patch test has proved to be the most val- uable procedure available. Most of the criticisms of the basic procedure should have been criticism of the way in which the procedurc •vas carried out. Products are frequently formulated on the premise that this can be safely