Development of a decision support tool for corrective actions during storage of fermented sausages in case of pathogens survival during their production - Specific application to Listeria monocytogenes

OBJECTIVES Depending on fermentation and ripening conditions of sausages, Listeria monocytogenes may survive well during their production. Given the absence of an additional antimicrobial step after manufacturing, most often storage at high temperatures before distribution could be the only option and this may accelerate the inactivation of several pathogens including L. monocytogenes. Therefore, the objective of this study was to compile the available data on L. monocytogenes survival during storage of vacuum-packaged fermented sausages at various temperatures in order to develop a generic model for decision-making purposes. METHODS Literature search was performed to identify all the studies reporting quantitative data relative to in situ survival of different L. monocytogenes strains being in different physiological states at the time of inoculation. A part of the studies were kept aside for validation purposes. Tool performance was assessed by the bias (Bf) and accuracy (Af) factors. The log-linear and two populations models were used to calculate the kinetic parameter of interest, i.e. inactivation rate (kmax). The 4D-value parameter was originated from the corresponding kmax. Linear regression was used to model log4D-value vs. temperature for determining z4D-values. A secondary model was developed for modeling the effect of sausage water activity (aw) on z4D- and √4D-values at 25oC. Multiple regression was employed to identify significant predictors of the secondary model. Model fitting evaluation was done by inspecting the R2 and Root Mean Sum of Squared Error (RMSE) indices. Model fitting and parameters calculation were carried out using GInaFiT v1.6, linear regression using Excel and multiple regression using SPSS v15.1. RESULTS Generally, the inactivation curve of L. monocytogenes appeared in two forms: the linear and biphasic. Because it was not possible to describe all inactivation curves by a common model, the time needed for 4 logs reduction (4D-value) was used instead the classical D-value (time needed for 1 log reduction of the pathogen). Modeling log4D-value=f(T) showed good correlation (R2=0.76-0.99). The observed 4D-values were compared to predicted ones obtained by the Pathogen Modeling Program (PMP). The comparison revealed that 4D-values at 25oC agreed relatively well (Bf=1.01 and Af=1.40) and therefore this temperature was chosen as reference (4D25-value). Multiple regression indicated that 4D25- and z4D-values were dependent only on aw (P=0.007 and P=0.041 for 4D25- and z4D-values, respectively) and not on pH (P=0.266 and P=0.769 for 4D25- and z4D-values, respectively), explaining a significant part of the variability observed in 4D25- (79.3%) and z4D-values (92%). Therefore, only aw was included as predictor variable in the secondary model. Finally, the decision support tool was successfully validated against the studies not initially used (Bf=1.08 and Af=1.28). CONCLUSIONS AND IMPACT OF THE STUDY The developed √4D25- and z4D-value models can be used to predict the desired time-temperature combinations that lead to additional pathogen reduction. The decision support tool can predict the fate of L. monocytogenes at a specific storage temperature and based on this prediction a decision could be made (corrective action) about the time needed to store the product before its distribution in order to achieve the additional desired pathogen inactivation. It should be noted, however, that model applicability lies within the studies domain used to develop the decision support tool, i.e. vacuum-packaged fermented sausages with post-ripening pH from 4.5 to 5.0 and aw from 0.82 to 0.92, and storage at temperatures from 4 to 30oC. Such tools can also be incorporated in HACCP studies of a food-producing company to assure food safety of its products.