Meta-regression models describing the effects of essential oils and added lactic acid bacteria on L. monocytogenes inactivation in cheese

Abstract

Biopreservatives such as plant-based antimicrobials and bacteriocinogenic starter cultures have been proposed as hurdles to increase microbiological safety of cheeses. In this study, meta-regression models were built to summarise the effectiveness of essential oils (EO) and lactic acid bacteria (LAB) on L. monocytogenes (LM) inactivation in cheese, and to evaluate other affecting factors and possible interactions. Suitable primary studies were identified through systematic literature search. From twenty-three studies reporting data on LAB and EO effects on LM counts in cheese, 282 and 322 entries were collected, respectively. The following information was obtained: study ID, antimicrobial class (EO or LAB) and name, LM mean log reduction, storage temperature, exposure time, application type (in milk or on cheese surface), antimicrobial concentration and pathogen inoculum level. Then, mixed-effects linear models with weights were separately adjusted to the LAB and EO data sets, with exposure time and antimicrobial concentration as nested fixed effects in application type. The results of the meta-regression model for the LAB data set revealed the significant impact of the application type (p=0.001), pathogen inoculum level (p<.0001) and storage temperature (p=0.001) on LM inactivation in cheese; while LAB concentration applied showed no significant effect (p=0.3688). An interaction between exposure time and type of application was also observed (p<.0001), meaning that the treatment duration, for the same LM reduction, depends if the antimicrobial is added to the milk or to cheese surface. Regarding the EO-LM meta-model, the results showed, again, the significant effects of pathogen inoculum level (p<.0001), storage temperature (p=0.0004) and application type (p<.0001), meaning that microbial reduction is faster when the EO is added to the milk (b=0.488) than onto the surface (b=0.334). Interactions between application type and exposure time (p<.0001) were also observed. Overall, the anti-listerial effect of EOs depends on its origin, yet, seemingly, those extracted from mint, oregano, salvia and basil present the greatest bactericidal effects in cheese matrix, as per analysis of random-effect marginal intercepts and concentration slopes. Globally, the effect of antimicrobials on LM reduction differs when applied in milk or on the cheese surface, and it is affected by antimicrobial concentration, storage temperature and time. The fact that pathogens inoculum level consistently appeared as a moderator driving the measured reductions should be further investigated.