SOAP BAR ANTIBACTERIAL ACTIVITY 233 For these experiments, a deodorant soap containing 1.5% triclocarban, sodium tal- lowate, sodium cocoate, water, glycerin, fragrance, sodium chloride, preservatives, and color was compared to non-germicide placebo bar soap. CALCULATION The number of CFUs (colony forming units) on each plate was converted to logs, and mean log CFU/plate determined for each panelist's forearms (treatment). Statistical differences between treatments were determined using the Student's paired t-test. The geometric mean CFU/plate is the antilog of the mean log CFU/plate. RESULTS AND DISCUSSION The results of three experiments using the agar patch method are shown in Table I. Regardless of whether nutrient agar or mannitol salt agar was used, the residual anti- bacterial activity of deodorant soap (1.5% triclocarban)-washed forearms after eight supervised washings showed a 79 to 88% reduction in the geometric mean CFU count compared to the placebo soap-washed forearms. All reductions were statistically sig- nificant at the •.025 probability level. Note that the coefficient of variation (standard deviation/mean) of the log CFU/plate of the control plates ranged from 3.3 to 6.4%. These values are comparable to those obtained in previous experiments using pipets (10), indicating that the use of an inoculation loop is as reproducible as the use of a pipet to inoculate the agar plates. Table I A Comparison of Residual Antibacterial Activity of a Deodorant Soap (1.5 % Triclocarban) vs. a Placebo Soap (no Triclocarban) by the In Vivo Agar Patch Method • Panehst Initials/ Test Soap • % Number Used Mean Log CFU/Plate + S.D. Reducnon c p-Value d Imt•als LS RD WF PF RB RC Placebo 1.9318 1.8949 2.3514 2.3784 1.9898 1.8647 .6874 79.5 .005 ñ .3165 Deodorant t.1t27 1.7703 1.3809 1.7807 1.0249 1.2173 Initials RL RD PF PW BB GG CS MH Placebo 1.0887 2.2535 2.2503 2.5692 2.087 1.8221 1.5519 1.8345 .732 81.5 .025 ñ .7120 Deodorant 1.1937 1.5576 0.8844 2.2329 0 1.6806 1.t032 0.9488 Initials GG CS SE JM PF RB MH RC Placebo 1.2915 1.2896 1.1286 1.7654 1.9211 1.7472 1.4053 1.5853 .9401 88.4 .001 q-.3916 Deodorant 0.477t 0.7101 0.6990 0.6667 0.2817 1.0195 0.4407 0.318t • All panehsts washed with the test bars for 8 applications. The placebo soap had a d•fferent base composition from the deodorant soap. Th_e test bacterium was S. epidermzdit ATCC t55. b d = mean difference of the placebo _mean log CFU/plate--the deodorant mean log CFU/plate. S.D. = standard deviation. • The % reduction = (1 - 1/antilog d) 100. d p-value = the smallest alpha level at which the null hypothesis of no differences can be relected. e The desired control inoculum was t00 CFU/plate on nutrient agar using a 1/10,000 dilution of the overnight culture in AATTC bacteriostasis broth. Actual control mean log CFU/plate q- S.D. = 2.0188 q- 0.1137 (n = 7). The contact time was 4 hours between the washed forearms and the inoculated plate. f The desired control inoculum was 200 CFU/plate on mannitol salt agar using a 3/10,000 dilution in 0.1% peptone. Actual control mean log CFU/plate q- S.D. = 2.2703 ñ . 1460 (n = 12). The contact t•me was 4 hours. Because some of the test plates contained no surviving CFU, "1" was added to all CFU/plate counts to allow mean log CFU/plate to be calculated. g The desired control inoculum was 400 CFU/plate on manmtol salt agar using a 6/10,000 dilution in 0.1% peptone. Actual control mean log CFU/plate q- S.D. = 2.4642 q- .08t0 (n = 8). Contact time = 2 hr.

234 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Results from the series of tests shown in Table I indicate that the percent reductions could be duplicated between tests when 6 to 8 panelists were used. Reducing the contact time between the skin of the forearm and the streaked plate from 4 hours to 2 hours did not affect the residual antibacterial activity of the skin and did reduce discomfort to the panelist. In later experiments a 30-minute contact time of the streaked petri dishes with the skin was shown to be equally effective. During these first experiments, some of the skin-attached streaked plates became con- taminated with resident skin bacteria. To avoid such problems in subsequent experi- ments, a spontaneous streptomycin-resistant colony from Staphy/ococcus epidermidis ATCC 14990 was isolated. This bacteria was designated as S. epidermidis JEH-51. One mil- ligram of streptomycin sulfate was also incorporated per milliliter of trypticase soy agar to inhibit the growth of resident skin bacteria on the stre'aked plates. In an attempt to improve test sensitivity and to further decrease the coefficient of variation, the inoculum on the streaked agar plates was also raised from 100-300 CFU to 1000-2000 CFU. The next series of experiments was designed to determine if a buildup of triclocarban germicidal activity on the deodorant soap-washed forearms compared to placebo soap (no triclocarban)-washed forearms could be shown as a function of the number of washings. Seven or eight different panelists were used for each experiment. The panelists washed for an average of 5 to 13 washes over a period of 2 to 4 days, with at least a 2-hour drying out time between washes. The data for these runs are presented in Table II. The reduction in bacterial levels observed with the deodorant soap after 5 washings Table II A Comparison of the Residual Antibacterial Activity of a Deodorant Soap (1.553 Triclocarban) as a Function of the Number of Washings • Test No. of No. of Applications Number Panelists b Mean (range) c • + S.D. % Reduction p-Value 4 8 5.25 (4-7) 0. 1722 q- .5172 32.7 N.S. d 5 8 5.13 (5-6) 0.0821 q- .1548 17.2 N.S. 4 + 5 e 16 5.19(4-7) 0.1271 q- .3717 25.4 N.S. 6 8 7.25 (7-8) 0.1400 q- . 1574 27.6 .05 7 7 9.57 (8-10) 0.2429 q- .2520 42.8 .05 8 8 13.0 (13) 0.1439 q- .1170 28.2 .02 • The placebo soap used had the same base composition as the deodorant soap. The test bacterium was S. epidermidis JEH 51. The desired control inoculum was 1,000-2,000 CFU/plate using a 3/2,000 dilution of the overnight culture. Actual control mean log CFU/plate ranged from 3.0365 to 3.2764 with coef- ficients of variation of 1.21 to 3.38%. The contact time of the inoculated plates on the washed forearms was 30 minutes. Other abbreviations are given in Table I. b Different panelists used in each test. c All panelists participated in a minimum number of supervised washings (the minimum number given under "Range") over a 2-4 day period. In addition, panelists were issued color-coded test bars, and instructed to perform any ad. lib. washings of their forearms with the test bars and to record the number of such washings. The total of the supervised plus ad./lb. washings is the maximum number of applications for any panelist in the test (the maximum number under "Range"). The "mean" is the average number of applications per panelist. d N.S. = nonsignificant (p .05). e Data from test 4 and 5 combined.