Predicting the effect of salt on heat tolerance of Listeria monocytogenes in meat and fish products.

Listeria monocytogenes is a potentially fatal foodborne pathogen that can be found in various ready-to-eat (RTE) products. It tolerates adverse conditions such as high salt concentrations and refrigerated storage, thus, the elimination of the pathogen in food processing often relies on heat processing. The objective of this study was to create a model to predict the effect of salt on heat tolerance of L. monocytogenes in meat and seafood products during heat treatments conducted at 57 to 65 °C to reduce numbers by ≥3 log10 cycles. Salt concentrations, up to 6% in the water phase (WPS%), were applied to cover a variety of lightly salted RTE meat and seafood products. The experimental work involved samples of ground pork tenderloin, ground chicken breast fillet and skinned, ground salmon fillet adjusted to different WPS% i.e., 3.6 and 5.2 WPS% for pork samples, 2.0, 3.0, 3.5 and 6.0 WPS% for chicken samples and 3.0 and 6.0 WPS% for salmon samples. All samples were inoculated with late-stationary phase L. monocytogenes cultures. For pork samples, a two-strain mixture of a pork isolate (MS22254) and an environmental isolate (MS22246) was applied. For chicken and salmon samples, a seafood isolate (MS22258) and isolate MS22246 was applied as single cultures. Samples were vacuum-packed in sterile bags, immerged in water bath, and held at constant temperatures of 57, 60 and 65 °C for pork samples and 58, 61 and 62.5 °C for chicken and salmon samples. For survivor curves, where at least 3 log10-reduction were obtained, heat tolerance was expressed as decimal reduction times, D-values. D-values were observed to increase with increasing WPS%. The effect of salt on heat tolerance of L. monocytogenes was defined as the relative increase (RI-value) in D-value obtained when salt had been added to the food. The effect of WPS% on RI-values was independent of heating temperatures, foods and strains. For secondary modelling, RI-values were transformed using the natural logarithm, ln(RI) and fitted to a linear model as a function of WPS%. Model validation, with 56 independent values collected from the scientific literature, resulted in bias and accuracy factors of 0.89 and 1.26, respectively, suggesting acceptable performance with tendency to slightly under-predict. The developed predictive model can be used to guide the design of heat processes for manufacturers of lightly preserved and mildly processed meat and seafood products requiring more than 3 log10 reduction of L. monocytogenes to ensure safety.