Towards safer fruit and vegetable products: Global meta-regression modelling of vegetative pathogen inactivation by non-thermal high-pressure processing

Abstract

This study aimed to collect and meta-analyse data on inactivation of Escherichia coli, Salmonella and Listeria in fruits and vegetables processed by high pressure (HPP) under non-thermal conditions (<45 ◦C). An extensive literature search yielded 55 selected articles, 1633 Log reduction values were extracted and 97 DP (time for 1-log reduction) were estimated. Time for 5-Log reduction (5DP) was also calculated, considering pressure come-up inactivation, shoulder and biphasic behaviour. Pressure and product pH category explained 91–93 % data variability, both being significant fixed effects factors along with microorganism. Random effects (study and strain nested within microorganism) did not improve model fit. A global regression Bigelow-type model for LogDP and Log5DP resulted in a LogDPref and Log5DPref parameter for each pH category and a common zP for all microorganisms. The model’s predictive performance was assessed with 1126 independent Log reduction data. Fail-dangerous predictions occurred in only 20 % of cases for E. coli, 6 % for Salmonella and 10 % for L. monocytogenes, thus confirming the model as a conservative tool for risk assessment, benchmarking and setting HPP parameters to meet a given performance criterion. At 600 MPa, up to 19 min are required to achieve 5-Log reduction of vegetative pathogens in low-acid products, while 3 min and 1.5 min are sufficient for acid and high-acid matrixes, respectively. A user-friendly decision support system prototype integrating the validated model will assist stakeholders in setting and validating HPP conditions that comply with regulatory requirements and ensure food safety in the production of fruit and vegetable-based foods.