212 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS dehyde-3-phosphate dehydrogenase, and asparaginase). Initially, the isothiazolinone forms a disulfide link with the thiol group on the amino acid. Occasionally, the chloromethyl-isothiazolinone may facilitate linkage with another thiol group to estab- lish a new disulfide linkage and release the biocide as a mercaptoacrylamide. This mercaptoacrylamide can tautomerize to a thioacyl chloride that may react again by denaturing nucleic acids (136). Formaldehyde also denatures protein but by alkylating amino and sulfhydril groups it can also alkylate the nitrogens of purine rings to denature DNA (137). Most of the formaldehyde donors (e.g., DMDM hydantoin, imidazolidinyl urea, Quaterium 15, polymethoxy bicyclic oxazolidine, etc.) act in this basic manner since these compounds release formaldehyde into the product or the microbial cell. Differences seen between the formaldehyde donors may exist as a result of when or what triggers the compound to release or "donate" formaldehyde. For example, a compound with a long hydrophilic chain connected to the formaldehyde-donating region (e.g., polymethoxy bicyclic ox- azolidine) may release formaldehyde only when the long chain enters into the lipopoly- saccharide portion of the membrane. Brominated compounds such as bromo-nitropropanediol and bromo-nitrodioxane act by oxidation of thiol groups (138-141) or by causing thioIs to convert to disulfides (142,143) where the thiol group first becomes brominated and then reacts with another thiol group to yield a disulfide and free bromide. As a result, enzymes involved in respiratory activity (e.g., dehydrogenases) and nucleic acid synthesis are inhibited, cell membrane integrity is compromised, and the cell wall may even be affected (144). One compound that is not technically a biocide but rather a biocide adjuvant is ethyl- enediamine-tetra-acetic acid (EDTA). This and other chelating agents remove magne- sium and calcium divalent cations from the cell wall, which is needed for stability (145). Once destabilized, they permit easier access of biocides into the cell. SELECTION OF PRESERVATIVE The ideal preservative would be broad-spectrum, safe and completely free of any sen- sitization issues, completely water-soluble, completely stable to all extremes of pH and temperature, completely compatible with all ingredients and packages, and impart no color or odor to the product, be inexpensive, and comply with government regulations. This ideal does not exist. One must select a preservative based on empirical testing. The only approach bordering on a theoretical basis for choosing a preservative is a qualified microbiologist's intuition, finely honed by experience. Selection of preservative may also be from published lists of available preservatives (146,147). These provide good sources for getting ideas of what might work in a formulation. Every formulation must be considered unique. Factors such as the physical and chemical nature of the product, how it is to be used, the container type and closure, and the shelf life must be considered when choosing the preservative (30). Often the selection of a preservative must be a compromise between efficacy, stability, and safety. More detail on the selection process of preservatives can be found by referring to several books and articles on the subject (148-150). SAFETY CONSIDERATIONS OF PRESERVATIVES One must always balance the risk of microbial contamination with the risk that a biocide

COSMETIC PRESERVATION 213 may give to a product. For example, many eye area products were permitted to contain phenyl mercuric acetate because the risk of infection to the eye was greater than the risk of exposure to the compound. The key consideration is to judge whether the product will be safe for the consumer under normal use and foreseeable misuse conditions. One of the first considerations of a preservative is its acute toxicity. Ocular irritation and subchronic and chronic toxicity tests are performed via the expected consumer exposure route to determine at what level the preservative can exhibit any irritant, toxic, or carcinogenic properties. Perhaps more important than these tests are the skin responses to biocides. Basic irritant responses can be a result of corrosion, acute irritation, cu- mulative irritation, or photoirritation. Skin sensitization is another key concern when using biocides. Nearly all biocides used today will elicit sensitization. Sensitization testing is performed in much the same way as irritant patch testing, except that much lower concentrations and repetitive appli- cations are used. Another concern for biocides is mutagenicity testing to determine if the biocide has the potential for mutating somatic or germ cells. In addition to this testing, embryological (or developmental) toxicity testing is done to determine if the biocide may be a teratogen capable of causing birth defects. In all these tests, the results must be compared to the ordinary-use and foreseeable- misuse exposure levels to give us a reasonable risk assessment. The definition of rea- sonable risk must include considerations based on the benefits from using the biocide, the ability to use less risky biocides for the same use, the economic benefits from using the biocide (can the biocide help prevent costly recalls due to contamination?), even how the biocide may affect the quality of life, the environment, and public opinion of the company. More detail on the safety considerations of cosmetic ingredients may be found in books by Waggoner and Whittam (151,152). CONCLUSION This article does not detail or discuss the pros and cons of the various methods used in cosmetic microbiology. There are plenty of references available from which the serious student can get this information (56,153-156). Regardless of which methods are in use by any particular company, the fact that the cosmetic industry has been so successful in providing adequately preserved products for its consumers is commendable and rein- forces the wisdom that we are capable of self regulation. The cosmetic microbiologist must balance a variety of factors to provide for safe, unspoiled quality products (157). In addition to knowing preservatives, he or she must understand microbial physiology, pathogenic microbiology, and microbial ecology. In addition to microbiology, he or she must understand organic and physical chemistry, toxicology, engineering, manufacturing and processing, sanitation, and regulatory/ environmental law. The cosmetic microbiologist must use all this education and knowl- edge within the context of the business needs of the company and be able to balance risk/benefit to the consumer using the product. Finally and most importantly, this person must have the highest of ethical standards, considering himself or herself as part of the cadre of health care providers in the world dedicated to serving humankind via the mission of providing microbially safe and efficacious products.