222 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS molecular weight aliphatic, steroid, or triterpenoid alcohols and fatty acids. Various fractions and chemical modifications of these mixtures are used in cosmetics, each with its unique physical and chemical characteristics. The comedogenic potential of these raw materials in our sampling ranged from negative to severe. Lanolin acid was more active than the whole lanolin mixture and lanolin oil liquid fractions. Acetylation of lanolin alcohol resulted in a significant increase in comedogenic potency. Kligman and Mills (7) previously found nine of eleven lanolin samples tested to be mildly to moderately comedogenic in the rabbit ear. While they found lanolin steroIs and de-waxed liquid lanolin inactive, they also report that acetylated lanolin alcohol and lanolin acid were strongly comedogenic. Fulton et al. (8), reporting varying degrees of activity with lanolins and their derivatives, also determined that acetylation usually resulted in increased comedogenic potency but reported lower levels of activity for lanolin acid. A report by Fulton et al. (8) of significant comedones resulting from applications of Lanogene, differs from our assessment of minimal activity with this ingredient obtained from the same supplier. An evaluation of vegetable oils by Kligman and Mills (7) indicated that most had some degree of comedogenicity, usually rather low, and all were inactivated when diluted to a 25% concentration in mineral oil. In our study, test results are similar however, the absence of activity from olive oil differs significantly from the moderate response reported by these authors. Cottonseed oil, not included in our assay, was negative in two animals treated by Kligman and Mills (7), while Kligman et al. (14) report this material to be mildly comedogenic but borderline when diluted with an equal part of mineral oil. A relationship between comedogenic potential and carbon-chain length of fatty acids and corresponding triglycerides has been reported (11,14). Although vegetable oils consist of varying mixtures of triglycerides, an inspection of relative activity based on the predominating portion of these esters revealed no consistent trend with respect to chain length. Despite the documented potent activity of oleic acid (7-8,14), two oils containing 77% to 84% of its triglyceride ester demonstrated no activity. There appeared to be a trend toward increased potency in vegetable oils containing comparatively high combinations of unsaturated oleic and linoleic acid esters. The fatty acid esters tested in this study are extensively used by the cosmetic industry as emollients or emulsifiers, often at high concentrations. With respect to the activity of these chemicals, Fulton et al. (8) report that although oleic and stearic acids were comedogenic, esterified products, such as butyl stearate and isopropyl isostearate, were always more potent. Our data on two derivatives of oleic acid do not support this generalization since isodecyl oleate reflected minimal comedogenic grades, and the decyl oleate ester was moderately active. With isopropyl myristate, perhaps the most widely used cosmetic emollient ester, Fulton and co-workers (8) observed severe lesions similar to those seen following applications of coal tar. Our similar findings of the high activity with this chemical conflict with the mild grades reported by Kligman and Mills (7) and Kligman and Kwong (18). The data we report on surfactants concur with those of Kligman and Mills (7) who found these cosmetic ingredients uniformly negative. Both in their study and our laboratory observations, Triton X-100 (octoxynol-9) was assessed to be noncomedo- genic at 50% concentration. Mills and Kligman (22) also show that this material

COSMETIC INGREDIENT COMEDOGENiCITY 223 produced no effects in a human test panel when applied for six weeks under occlusive patches at 0.25% concentration. Strauss et al. (23), however, using the same technique, concluded that Triton X-100 appeared to be comedogenic. Our determination, and that of Kligman and Mills (7), of the noncomedogenic activity of sodium lauryl sulfate conflicts with the high scores published by Fulton et aL (8) on this surfactant. This diversity in comedogenic findings among investigators using the same or similar test agents may be attributed to differences in techniques, purity of test samples, biological variations, interpretation criteria, or other unapparent factors. With a simple chemical mixture, although source and purity may differ, several dermatologists found that hydrophilic ointment (USP), for example, was noncomedogenic in both animal and human studies (7,17,22). Fulton et al. (8), on the other hand, report that significant comedones were produced in rabbit ears after a two-week application of a USP ointment sample and ascribe this result to the presence of sodium lauryl sulfate (1%) in the formula. Investigators attempting the utilization of rabbit test dat a to predict the potential for comedogenicity in humans have emphasized several factors bearing on such compari- sons. The ultrasensitivity of the rabbit follicle to respond readily to test materials has been extensively documented in published work (6-7,15,17,22-23). Differences between this animal model and the human pilosebaceous apparatus in bacterial residence, morphology, and comedo development have been outlined (9-11,15,24). With respect to the ability of the rabbit follicle to respond quickly to many test materials, the effects of dermal irritants have not been thoroughly studied. Kligman and Katz (15) assert that hyperkeratosis of the follicular epithelium cannot be ascribed to simple irritancy and were unable to induce follicular hyperkeratosis with several strong irritants. Kligman (10) conversely states that in the first week of application it is difficult to distinguish among epidermal acanthosis, hyperkeratosis, and inflammatory changes produced by an irritant and a comedogen. He also explained that with irritants, while hyperplasia of the epidermis may be great, the follicular epithelium is usually slightly affected, and the loose, horny squame in the follicle which may accompany an irritation reaction are usually sloughed too quickly to allow comedo formation. Kligman et aL (14) also report that irritants us/aally stimulate the formation of non-adherent horny cells, while comedogenic agents cause these to stick tightly together to form a comedone. In one's observation of slight to moderate increases in follicular hyperkeratosis, confusion may exist in the interpretation of an irritant versus a comedogenic response. CONCLUSION The purpose of this study was to expand on data available on the comedogenic potential of cosmetic ingredients and attempt to corroborate the findings of other investigators. In some cases our test results conflict with those previously published. The discrepancies in comedogenic findings among investigators on the same or similar chemicals may be due to the use of different techniques, purity of test agents, biological variation, and other factors. Compared to established animal methods used to predict or approximate effects on human skin, the Rabbit Ear Test is a fairly new laboratory procedure. Since the comedogenic bioassay lacks standardization, we