330 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS statistical controls of preservative efficacy tests, to determine synergistic effects of for- mula components, to set rational acceptance criteria, and to determine required D-val- ues (4-9). As originally outlined, the linear regression method specified that aerobic plate counts (APCs) be performed immediately after inoculation and at various times afterwards- typically at 2, 4, and 24 hr for bacteria and at 4, 8, and 24 hr for yeasts and molds. Additional samples were taken at 3, 5, or 7 d after inoculation unless the previous APC was 10/ml (4). These APCs were then used to determine the D-values for each test organism in product samples. It would be desirable to be able to determine the rate of microbial death--to determine D-values--using fewer APCs than required by the original linear regression method (4) because this would be less labor-intensive and would save on the cost of materials. This is now possible. This report describes a rapid screening procedure that allows calculation of D-values from two APC data points for each test organism. Our laboratory has performed several hundred rapid screening tests by use of this procedure to determine whether cosmetic and OTC-drug formulas meet acceptance criteria (5). EXPERIMENTAL TEST ORGANISMS The American Type Culture Collection (ATCC) organisms used in this study were received directly from the ATCC and consisted of Staphylococcus aureus (ATCC 6538), Pseudomonas aeruginosa (ATCC 9027), P. cepacia (ATCC 13945), Escherichia coli (ATCC 8739), Aspergillus niger (ATCC 16404), and Candida albicans (ATCC 10231). The Bacillus sp. was isolated from a product. The bacteria were maintained by weekly transfer on tryptic soy agar (TSA) with 0.07% lecithin and 0.5% Tween 80 (TSALT). A. niger and C. albicans were maintained by weekly transfer on potato dextrose agar (PDA). The bacteria and C. albicans were incubated at 32øC, and A. niger was incubated at 25øC. The bacteria were grown for 24 hr on TSALT agar prior to use in preservative efficacy testing. C. albicans was grown for 24 hr on PDA, and A. niger was grown for 7 d on PDA prior to use in preservative efficacy testing. TEST SAMPLES The test samples used in this study included proprietary formulations of o/w emulsions (creams, lotions, facial moisturizers), anionic surfactants (shower gels and cleansers), and OTC-drug products (sunscreens, anti-dandruff shampoos). TEST PROCEDURES Preservative efficacy tests were performed using saline suspensions from surface growth of each test organism as described elsewhere (4), with the following exceptions: 1) bacterial and C. albicans cultures were incubated at 32øC, and 2) APC determinations were made using single surface-spread plates of each o-ganism or each group of organ-

PRESERVATIVE EFFICACY TESTING 331 isms in "pooled" inocula. Pooled inocula consisted of saline suspensions prepared using organisms that had similar maximum acceptable D-values (i.e., P. aeruginosa and S. aureus E. coli and P. cepacia) and/or recovery media (i.e., C. albicans and A. niger). DETERMINATION OF D-VALUES D-values were calculated as described previously (4). Estimated D-values (ED-values) were calculated from the same data using APCs taken immediately after inoculation (i.e., at time = 0 hr) and at 24 hr for site-significant organisms (i.e., pathogens/ opportunistic pathogens) or 7 d for organisms that are not site-significant (i.e., non- pathogenic bacteria, yeasts, and molds). Thus, ED-values were equal to the negative reciprocal of the slope of a survivor curve constructed from APCs immediately after inoculation of test organisms into test samples and at 24 hr (for S. aureus, P. aeruginosa, or C. albicans) or 7 d (for E. coli, P. cepacia, Bacillus, and A. niger). D-values and ED-values, which were calculated using APCs of C tO/g, were expressed as "less than" a specific time (i.e., C3.5 hr). The "less than" signs were not used in determining mean APC values. STATISTICS Significant differences between D-values and ED-values were assessed by an independent t-test using Sigmaplot 4.0 (Menlo Park, CA). Linear regressions were determined using a hand-held calculator (4). RESULTS AND DISCUSSION Screening studies have been used in our laboratories for estimating D-values for several years however, ED-values obtained with the screening method have not been compared with D-values obtained by the linear regression method for cosmetic and OTC-drug products. D-values were determined using APCs at several times for each test organism. Direct comparison of D-values with ED-values was possible because both were calculated from the same experimental data: D-values were determined using 0, 2, 4, and 24 hr or 7 d APCs ED-values were determined using 0 and 24 hr or 7 d APCs. Table I compares 60 D-values and ED-values calculated using data obtained during preservative efficacy tests of cosmetics and OTC drugs including facial moisturizers, night creams, sunscreens, facial toners, shower gels, and antidandruff shampoos that were challenged with several different test organisms. The D-values ranged from CO. t hr (i.e., where the population of P. cepacia was not detected at the 2-hr reading, so that the D-value had to be estimated) to 39 hr. The ED-values ranged from C3.0 hr to 42 hr. The Student's t-test showed that the mean D-value (6.9 hr) and the mean ED-value (7.4 hr) were not significantly different (p 0. t0). When the D-values were C t0 hr, the differences in D-values and ED-values for the same experimental data ranged from 0 hr (where the values were the same) to 4.9 hr in a night cream (where the D-value was CO. 5 hr because P. aeruginosa died so quickly that no viable cells were recovered at 2 hr and the ED-value was C 5.4 hr because no viable cells were recovered at 24 hr). The differences in D-values and ED-values for the same experimental data ranged from 0 to