Benzoyl peroxide: an integrated human safety assessment for carcinogenicity.

Topical benzoyl peroxide has been used in the treatment of acne for over 30 years, with no reports of adverse effects that could be related to skin carcinogenesis. Two case-control epidemiological studies have found a lack of association between the specific use of benzoyl peroxide and skin cancer. In addition to these findings in humans, 23 carcinogenicity studies in rodents with benzoyl peroxide, including 16 employing topical application, have yielded negative results. An increase in skin carcinomas was reported in 1 study in which benzoyl peroxide in acetone was applied to the skin of SENCAR mice for a 1-year period; however, this study did not employ adequate control groups to fully understand the unusual findings, and the results were inconsistent with those of 6 other similar studies. While benzoyl peroxide is not a complete carcinogen in animals and has weak or no mutagenic potential, it has been found to be a tumor promoter in mouse skin using experimental two-stage models of carcinogenesis. Consistently positive results have been obtained in tumor promotion studies in which SENCAR mice were exposed to initiating doses of potent experimental carcinogens followed by promotion with benzoyl peroxide in acetone. Negative results have been obtained in similar studies with commercial formulations. However, the results of promotion studies with benzoyl peroxide do not carry significant weight for human safety assessment as evidenced by (i) the absence of demonstrated carcinogenicity in humans of a number of rodent tumor promoters despite long-term human exposure; (ii) the observation that tumor promotion in mouse skin occurs only under specific experimental conditions and predominantly in highly sensitive strains; (iii) clinical use scenarios markedly different from the conditions resulting in tumor promotion in mouse skin; and (iv) the significant physiological differences between mouse and human skin. Thus, to date, available scientific evidence does not allow the results of these rodent promotion studies to be meaningfully applied to human safety assessment. As such, significant scientific progress must be made before human safety estimations can be derived from rodent promotion data.(ABSTRACT TRUNCATED AT 400 WORDS)