Non log-linear bacterial survival curves: a new primary model

Introduction Growth and thermal inactivation of microoganisms were extensively studied in order to get information on the microbiological impact of manufacturing process or formulation of foods. Nevertheless, there are still many unexplored points in the field of predictive microbiology: not controlled in order to be able to predict in the evolution of bacterial populations during the production or the storage of a food. In particular, slow decline of a population during storage, or during an acid or osmotic stress cannot be accurately predicted. This late development of the field of the non-thermal inactivation modelling is partly due to the heterogeneity of shapes of the bacterial survival curves which can be concave, convex, sigmoidal, or simply log-linear destruction. The same bacterial species can present the different shapes of curves according to experimental conditions(2, 6). As a consequence, a lot of primary models were published in literature. After having selected the primary model which yields the best fit of observed survival kinetics, authors deduce from it the t 4D (the time of storage corresponding to a 4 log decline). This has the advantage of reflecting the evolution of the rate of inactivation as a function of the various investigated physicochemical factors regardless of the pattern of curves. On the other hand this modelling does not give information about the time for a given decimal log decline. This overall parameter simplifies the use of secondary model to predict the inactivation of microorganisms in unfavourable media. However, many models describing non log-linear curves of thermal inactivation were published (1, 3). The description of shapes relatively complex like sigmoids requires the implementation of new parameters which are generally without chemical, physics or biological meaning or which do not evolve in a logical way with the level of stress. Lastly, most of these models assume that the probability of survival tends towards an asymptote when time tends towards infinity. If their use does not set a problem for short processing times, they seem to overestimate survival for prolonged times. Materials and methods Brain heart infusion broth was supplemented with HCl to give pH values ranging from 3.0 to 4.0. Overnight cultures of strains of Salmonella typhimurium (isolated from brine, ADRIA-France), or Listeria monocytogenes (isolated from pork-butchery, SOREDAB-France), were grown. Flasks of brain heart infusion were stored at 12°C under agitation (100 rpm) and sampled at appropriated intervals. The survivors were counted by plating on brain heart …