Meta-analysis of antimicrobial activity of Allium, Ocimum, and Thymus spp. confirms their promising application for increasing food safety.

Biopreservation strategies such as the use of Mediterranean plant extracts to ensure food safety are promising to deal with the emergence of antimicrobial resistances and the overreliance on food chemical additives. In the last few decades, antimicrobial susceptibility testing (AST) for evaluating the in vitro antibacterial potential of plant extracts against the most relevant foodborne pathogens has been widely reported in the literature. The current meta-analysis aimed to summarise and analyse the extensive evidence available in the literature regarding the in vitro antimicrobial capability of Allium, Ocimum and Thymus spp. extracts against foodborne pathogens. A systematic review was carried out to gather data on AST results of these extracts against Listeria monocytogenes, Staphylococcus aureus, Salmonella spp., Escherichia coli and Bacillus cereus, including inhibition diameters (ID) and minimum inhibitory concentrations (MIC). A total of 742 records were gathered from a raw collection of 2,065 articles. Weighted mixed-effect linear models were adjusted to data to obtain pooled ID, pooled MIC and the relationship between both model estimations and observations. The pooled results revealed B. cereus as the most susceptible bacteria to Allium sativum (pooled ID = 20.64 ± 0.61 mm) by diffusion methods and S. aureus (pooled MIC = 0.146 mg/mL) by dilution methods. Diffusion methods did not yield conclusive results for Ocimum spp. extracts; however, the lowest pooled MIC was obtained for S. aureus (0.263 mg/mL). Among the foodborne pathogens evaluated, B. cereus showed the highest sensitivity to Thymus spp. extracts by both diffusion and dilution methods (pooled ID = 28.90 ± 2.34 mm and MIC = 0.075 mg/mL). The methodology used for plant extraction was found to not significantly affect MIC values (p > 0.05). Overall, the antimicrobial effectiveness of the studied extracts against Gram-positive and Gram-negative bacteria was demonstrated. Finally, the robustness of the meta-regression model was confirmed, also revealing an inversely proportional correlation between the ID and MIC measurements (p < 0.0001). These results provide a robust scientific basis on the factors affecting the in vitro antimicrobial efficacy of extracts from Mediterranean plants. They also provide valuable information for stakeholders involved in their industrial application in food, including producers, regulatory agencies and consumers which demand green-labelled foods.

Cardinal parameter meta-regression models describing Listeria monocytogenes growth in broth.

Since Listeria monocytogenes has a high case-fatality rate, substantial research has been devoted to estimate its growth rate under different conditions of temperature, pH and water activity (aw). In this study, published findings on L. monocytogenes growth in broth were extracted and unified by constructing meta-regression models based on cardinal models for (i) temperature (CM[T]), (ii) temperature and pH (CM[T][pH]), and (iii) temperature, pH and aw (CM[T][pH][aw]). After assessing all the sources retrieved between 1988 until 2017, forty-nine primary studies were considered appropriate for inclusion. Apart from the modelling variables, study characteristics such as: type of broth (BHI, TSB, TPB), reading method (colony-forming-units, CFU; or binary-dilution optical density methods, OD), inoculum concentration and strain serotype, were also extracted. Meta-regressions based on CM[T] and CM[T][pH] were fitted on subsets of the 2009 growth rate measures and revealed that type of broth and reading method significantly modulated the cardinal parameter estimates. In the most parsimonious CM[T][pH][aw] meta-regression model, whereby the variability due to type of broth was extracted in a nested random-effects structure, the optimum growth rate µopt of L. monocytogenes was found to be lower when measured as CFU (0.947 log CFU/h; SE = 0.094 log CFU/h) than when measured as OD (1.289 log CFU/h; SE = 0.092 log CFU/h). Such a model produced the following cardinal estimates: Tmin = -1.273 °C (SE = 0.179 °C), Topt = 37.26 °C (SE = 0.688 °C), Tmax = 45.12 °C (SE = 0.013 °C), pHmin = 4.303 (SE = 0.014), pHopt = 7.085 (SE = 0.080), pHmax = 9.483 (SE = 0.080), awmin = 0.894 (SE = 0.002) and awopt = 0.995 (SE = 0.001). Integrating the outcomes from numerous L. monocytogenes growth experiments, this meta-analysis has estimated pooled cardinal parameters that can be used as reference values in quantitative risk assessment studies.

Effects of Essential Oils on Escherichia coli Inactivation in Cheese as Described by Meta-Regression Modelling.

The growing intention to replace chemical food preservatives with plant-based antimicrobials that pose lower risks to human health has produced numerous studies describing the bactericidal properties of biopreservatives such as essential oils (EOs) in a variety of products, including cheese. This study aimed to perform a meta-analysis of literature data that could summarize the inactivation of Escherichia coli in cheese achieved by added EOs; and compare its inhibitory effectiveness by application method, antimicrobial concentration, and specific antimicrobials. After a systematic review, 362 observations on log reduction data and study characteristics were extracted from 16 studies. The meta-regression model suggested that pathogenic E. coli is more resistant to EO action than the non-pathogenic type (p < 0.0001), although in both cases the higher the EO dose, the greater the mean log reduction achieved (p < 0.0001). It also showed that, among the factual application methods, EOs' incorporation in films render a steadier inactivation (p < 0.0001) than when directly applied to milk or smeared on cheese surface. Lemon balm, sage, shallot, and anise EOs showed the best inhibitory outcomes against the pathogen. The model also revealed the inadequacy of inoculating antimicrobials in cheese purposely grated for performing challenge studies, as this non-realistic application overestimates (p < 0.0001) the inhibitory effects of EOs.