INOCULUM PREPARATION IN PRESERVATION TESTING 201 It is believed that decreases in the rates of inactivation of test organisms during preser- vative efficacy testing of samples containing TSB or TSB(-G), as reflected by increases in the D-values for these organisms, are due to the inactivation of a portion of the preservative system or to a protective effect of nutrients in the culture medium (3). Moss and Speck (17) reported that peptides present in trypticase, which is a component of TSB and TSB(-G), were responsible for improved recovery of freeze-injured E. coli. The effect of components of these broths on repair of sublethally injured bacteria or on D-values was not determined in the current work. These data demonstrate that the addition of broth inocula or sterile broth media to test samples undergoing preserva- tive efficacy testing produces errors in the test results, namely, an increase in D-values, in most cases. Use of broth inocula constitutes a form of abuse testing (i.e., determining tolerance to the addition of extraneous materials, dilution, or adverse physical conditions). Signifi- cant decreases in the rates of inactivation of bacteria were produced with 0.2% broth (0. ! mL/50 mL sample) in these studies. Thus, broth inocula should not be used in routine preservative efficacy testing. A similar recommendation was made in an earlier report from our laboratory (3). The data presented here illustrate that the culture medium affects the metabolism of bacteria and that the type of inocula (i.e., saline vs. broth) affects the kinetics of micro- bial death during preservative efficacy testing. The kinetics of microbial death in pre- servative efficacy testing were discussed by Bean (18) and in reports from this laboratory (3,7,8, 19). We are unaware of reports in recent literature that attempt to address the difference in rates of death obtained in preservative efficacy testing of cosmetic or phar- maceutical products using broth-vs.-saline inocula. It is believed that the use of non- standardized procedures, such as different culture media, growth conditions, and in- lOO 8o 6o 4o I r• 2o / TSALT/SiNE i• TSB(-G) CREAM LOTION C LOTION D LOTION E LOTION F TEST SAMPLE Figure 4. Comparison of D-values obtained in preservative efficacy testing of a cream and four lotions using saline and TSB(-G) inocula of S. aureus.

202 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table V Effect of TSB on D-values of Test Organisms in Lotion A Addition of 0.1 ml of TSB or sterile -- Test organism deionized water D-value X S. aureus Water S. algrelg$ P. aeruginosa Water TSB TSB Water 6.8 5.6 12 15 0.7 1.1 1.1 14' 6.2* 0.9 Water P. aeruginosa TSB TSB 1.2 B. ereus Water 1.5 1.3 Water 1.0 B. ereus TSB 1.6 1.5 TSB 1.4 E. coli Water 14 15' Water 16 E. coli TSB 24 24* TSB 24 1.2 Table values are D-values in hours. * Mean D-values obtained with TSB addition were significantly larger than with water addition (p = 0.05). ocula may help explain the different preservative efficacy test results obtained by dif- ferent laboratories. The need for standardized test conditions in antimicrobial evaluation is recognized. Thus, the influence of growth phase of a number of organisms, including S. aureus and P. aeruginosa, on microdilution susceptibility tests was examined by Barry et al. (20). They observed that 18-24-hr surface growth on Blood Agar produced similar results as broth cultures (5-6-hr BHI and 2-4-hr TSB) when the turbidity of the inocula were standardized by adjusting turbidity to match that of a McFarland 0.5 standard. Wicks et al. noted the need for controlling the bacterial population of the inoculum within defined limits when conducting antimicrobial susceptibility tests (21). Mayhall and Apollo studied the effect of S. aureus growth phase on antibiotic resistance and found that log phase organisms were killed much more rapidly than stationary phase inocula (22). Although the current studies demonstrated that the growth medium affects the results