Inappropriate use of D-values for determining biocidal activity of various antimicrobials

The objective of this study was to investigate the application of established D-value calculations to survival curves for various bacteria using the following antimicrobials: acidified sodium chlorite, triclosan, octanoic acid, and sodium hydroxide. D-values can be calculated in 3 ways, a linear regression, an endpoint calculation, or an average of multiple endpoint calculations. The assumption made in calculating a D-value is that the rate of kill follows 1st-order kinetics under specified treatment conditions. Each antimicrobial solution was challenged with approximately 108 CFU/mL of Staphylococcus aureus, Listeria monocytogenes, Salmonella enterica subsp. enterica, and Escherichia coli independently and in triplicate. Test systems were sampled at each of the 10 time points over a period of 7 min, neutralized, pour plated then incubated at 35 °C for 48 h (AOAC official method 960.09). Survival curves using the log-transformed data were calculated using regression analysis. Correlations coefficients for all linear regression analyses ranged between 0.291 and 0.982, with 6 of the 16 different treatment systems having an R2 value below 0.7. Methods used for calculating D-values should lead to the same result if the survival curve in a given condition is linear. The calculated D-values were different using endpoint analysis (Stumbo method), linear regression, and average of multiple endpoints. This study demonstrates the nonlinearity of inactivation curves of antimicrobials. D-value estimations cannot be reliably used to illustrate biocidal activity in antimicrobial test systems.