482 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS METHODS PRE-TEST EVALUATION Initially, the adequacy of the preservative system is evaluated through in vitro challenge testing. An accelerated preservation test (APT) is conducted to rapidly assess preserva- tive efficacy (Table I). The product is inoculated with 106 cfu/gram of product for bacteria and 105 cfu of mold and yeast/gram of product. At the time of inoculation, extra nutrients in the form of glucose and a minimal salts buffer are added to the product sample. Assays are performed at 2 and 7 days after inoculation. If mold is recovered at Day 7, an additional Day 12 assay is made. A reduction of bacteria to 10 cfu/g in 7 days and greater than 99.9% reduction of mold and yeast in 12 days are the criteria for passing the APT. A long-term in vitro challenge test is initiated on those formulations which meet the criteria of the APT. Selected microorganisms are inocu- lated into the product at Week 0 and again at Week 2 microbiological assays are performed weekly. At the completion of 8 weeks of testing, the product is deemed adequately preserved if bacteria are reduced to 10 cfu/g one week after each challenge inoculation (Weeks 1 and 3) and maintained at that level throughout the test period. The reduction of fungal recoveries should initially be greater than 99.9% of the original inoculum concentration and continue to 10 cfu/g of mold and yeast by test comple- tion. When the formulation has demonstrated in vitro preservation efficacy, it is then tested for in vivo preservation efficacy. CONSUMER-USE TEST REGIMEN A consumer-use panel representative of the target audience for the product's intended use is selected. A dermatologist will supervise studies as required. Panel size is selected for statistical significance, averaging 15-30 panelists per product. Mascara panels are larger, with 40-60 people per product. The test panels designed for mascara and other eye products have more stringent requirements because concern exists in the industry over ocular infections associated with improperly preserved mascara (8,9). The associa- tion of Pseudomonas-induced infection leading to corneal ulceration is also documented (10). A microbiological prescreening of the microflora type and level in the area around the panelist's eyelids and lashes is performed. The microflora is then monitored throughout the test and after panel completion as a measure of the potential microbial load which may be introduced into the product. Table I In Vitro Challenge Testing Accelerated Standard Preservation Test Challenge Test Test Length 7 days 8 weeks Challenges 1 2 Organism Concentration 10 6 cfu/g bacteria Same 10 5 spores/g of mold and yeast Nutrification Glucose and minimal salt buffers None Assay Intervals 0 time, 2 days, 7 days 0 time, Weeks, 1, 2, 3, (12 days if mold is recovered 4, and 8 at Day 7)

MICROBIOLOGICAL INTEGRITY OF COSMETICS 483 A microbiological assay is performed prior to product distribution. In the case of eye product panels, microbiological assays of the product are conducted at several time intervals throughout the test rather than only at test initiation and completion. Products are coded and distributed randomly to the panelists, using a double crossover study design in which panelists first use one test product and are then "crossed over" to use the second test product. Panelists are given instructions for product use and a calendar to record the frequency of product use. Panels are conducted for 3-8 weeks. The actual length is dependent on the product and its intended use. More intensive evaluation of sensitive eye area products require longer test panels mascara in use testing averages eight weeks or greater in length. At the conclusion of the use test, the microbiological evaluation is performed first, followed by any additional physical/chem- ical testing. MICROBIOLOGICAL EVALUATION/CRITERIA Following the consumer-use test, an assay is performed on the used product 24 hours after the last application of the product. Preservative efficacy is more accurately mea- sured by allowing 24 hours before assaying since this more closely simulates ordinary use. Assays are performed on nutrient agar (BBL 12174) with 2% polysorbate 80, which is added for neutralization of the preservative, Vogel-Johnson agar (BBL 11812 for detection of Staphy/ococcus species), Pseudomonas Isolation agar (Difco 0927-01), and Mycophil agar (BBL 11450 for detection of yeasts and molds). Plates are incubated for 48 hours at 35øC and the test results are recorded. Enrichment tests are performed for mascara and eye products. A broth dilution of the product [1/10 in AOAC Letheen Broth (BBL 10914)] is incubated at 35øC for 24 hours. These broth dilutions are then streaked onto enrichment plates of Vogel-Johnson agar, MacConkey agar (BBL 11387 for the detection of gram negatives), and Pseudomonas Isolation agar, which in turn are incubated for 48 hours at 35øC. Criteria for formula acceptability after use is recovery of • 100 cfu/g total aerobic plate count including yeasts and molds and 10 cfu/g gram negatives on initial assay or recovery of 10 cfu/g total plate count on retest of the product. In addition, the product should not contain Pseudomonas species, Escherichia coli, or Staphylococcus aureus, All microorganisms recovered after retest are identified. RESULTS One-hundred forty three products were consumer-tested from 1980 through early 1985. A full complement of cosmetic and toiletry products containing a wide range of cosmetic preservatives were evaluated after consumer use. Preservative efficacy under various use conditions was measured. Table II itemizes the creams and lotions which have been tested. Twenty-eight creams were panel-tested these include night creams, cleansing creams, moisturizers, sunscreens, and eye creams. Seventy-three lotions were evaluated, including moisturizers, cleansers, skin lotions, milky lotions, fresheners and toners, foundations, sunscreens, and body lotions. Other product types evaluated during this five-year period are listed in Table III. The complete history of consumer-use testing is summarized in Table IV. Only four of the 143 products failed to meet the test criteria, i.e., three lotions and one shower gel. With 2182 lotion samples assayed, five samples (0.22%) were unacceptable. With 254