Robust models for the growth of Listeria monocytogenes in foods as a function of temperature based on systematic review and meta-regression.

Listeria monocytogenes is a major food safety concern, especially for refrigerated ready-to-eat products. As this pathogen requires high doses to cause illness, reliable models for its growth kinetics as a function of temperature are essential for risk assessment. This study conducted a (PRISMA) systematic review to compile growth data for L. monocytogenes across seven food categories (dairy, milk, egg, fish, plant-based, meat, and mixed products) as well as laboratory media. A total of 918 growth curves (16,234 time points) were manually extracted and modeled using the Baranyi (primary) and Baranyi-Ratkowsky (primary+secondary) models. Model performance was assessed through residual analysis, and goodness-of-fit indicators were compared concluding that the Standard Error of Residuals (SER) is more robust than Root Mean Squared Error or Accuracy Factor for evaluating goodness of fit, obtaining the following benchmark SER values: 0.33 log10 CFU/g (10th percentile 0.15; 90th percentile 0.64) for the one-step Baranyi-Ratkowsky model; 0.23 log10 CFU/g (10th percentile 0.09; 90th percentile 0.50) for the primary Baranyi model. Parameter analysis showed that studies with fewer temperature levels tended to underestimate cardinal temperature Tmin. Meta-regression was applied to the square root of growth rates (√μ) as a function of temperature. Food matrix was a significant random effect, with the overall model yielding an intercept of 0.043 (log₁₀ CFU/h)1ᐟ2 and a slope of 0.015 (log₁₀ CFU/h)1ᐟ2C-1, with residual variability of 0.045 (log₁₀ CFU/h)1ᐟ2. Product-specific effects were quantified for each category, providing a comprehensive and robust framework for modelling the growth of L. monocytogenes under sub-optimal temperature conditions and supporting improved risk assessment applications.