PREDICTIVE PATCH TESTING 325 ferent times, even on the same panel. A test subject who experiences reactions should not be used again for sixty to ninety days, since the reac- tion site is more sensitive to irritants and may give a false positive. It must also be remembered that the usual irritancy tests will generally not uncover such undesirable effects as drying, stinging, scaling, aeneform reactions, etc., which may be of the greatest significance in consumer use. A test procedure which incorporates features adequate for irritancy evaluation has been reported by Finkelstein, Laden, and Miechowski (2). It involves the repetitive application of the test substance and of a suit- able control formulation on cotton flannel pads, occluded by polyeth- ylene film. For materials which are actually used on a daily basis, ap- plication is made for a seventeen-hour period on four consecutive days. Substances used once or twice weekly are applied for five hours a day on five consecutive days. Irritancy reactions are scored daily, and weight is given both to the severity of reaction and the number of days required to produce the reaction. Finkelstein and his co-workers found this a fairly sensitive and reliable indicator of the irritancy of agents which fall in the low range on the Draize test. Results of the test indicate whether (as regards irritancy) it is worthwhile going ahead with usage studies and trial sales with the new formulation. With this and all similar proce- dures, selection of proper control formulations is important. Predictive tests for sensitization are even more complicated than irritancy tests. It is relatively simple to screen out potent sensitizing agents, since even casual contact with these may induce reactions in lab- oratory or factory personnel. However, the problem is to predict accu- rately the sensitization rate when the formulation is only a weak sensi- tizer and is used by the consumer in a way which differs from the usual test procedure exposure. The predictive tests for sensitization in gen- eral use a formalized method of application, repeated one or more times in the various methods. The aim is, of course, to determine how often sensitization can be purposely induced by repeated application of the test substance. The frequency of reactions is regarded as an indicator of how the substance will act under use conditions a judgment can then be made as to whether the level is tolerable. However, questions have been raised concerning the sensitivity of the test methods and the relation- ships of test results to usage results. The first of the predictive tests was the Schwartz (3) prophetic patch test. Schwartz suggested that new cosmerles should be tested by the closed patch test method on at least 200 subjects, using as a control an old formula with a known record of safety. The patches are left on for

326 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS forty-eight hours. Repeat patch tests are made two to three weeks later. Those subjects showing reactions on the second contact which were not present on the first are deemed sensitized. According to Schwartz, if the new formula shows more reactions than the control it is unsafe. Schwartz admits that the test may "not give an accurate idea of what may happen under conditions of actual use." Therefore, a four- week paired comparison use test of the cosmetic on the same 200 sub- jects is recommended before trial sale. In this part of the test the occur- rence of more than one case of dermatitis indicates that the formulation is unsafe. Trial sale, the final step by Schwartz' definition is the sale and use of 5-10,000 units in one community. The Schwartz prophetic patch test has certain inherent defects. False positives may occur, since borderline primary irritants can sometimes produce reactions which can be confused with sensitizations when only a patch test reading is made. More important are the false negatives, which are due to the fact that the single application of a small amount of the cosmetic is often inadequate to produce sensitization, except in the case of strong allergens. The use test which Schwartz recommended to follow the initial patch application is probably a recognition of this inadequacy. Most of the predictive burden is shifted to the use test, but this part of the procedure is prob- ably numerically inadequate to reveal low reaction rates. Also, some products are used only once weekly, or every four to six weeks, and can- not be evaluated for sensitization in a four-week use test. The Brunner-Smiljanic test (4), reported in 1952, attempted to sur- mount some of the problems of the prophetic patch test by increasing the frequency of application and using a larger area of contact with the test substance. Following an initial standard patch test, daily thirty-minute repetitive applications of the test substance are made to the forearm using saturated 7.5 X 7.5 cm gauze squares on three different sites in ro- tation. After an initial five-day application period, there is a rest period of one week followed by another ten-day application period. A stand- ard patch test is repeated at the end of the series. With this method primary irritant reactions are easily distinguished from sensitization since irritant reactions fade by the day following application, when the test sites are inspected. Sensit/zation reactions consist of persistent red- ness, sometimes with blistering and swelling, at the application sites. In many cases spontaneous flares occur in the sites of previous applica- tions at the time sensitization is induced. Standard patch tests also be- come positive in the subjects devdoping sensitization. Groups of 12 subjects are tested consecutively if more than one reaction occurs in the