Predictive Food Microbiology - new models for safety and quality assessment of a broad range of dairy products

This PhD-thesis focused on development of mathematical models to predict growth of spoilage and pathogenic bacteria in a broad range of dairy products. The studied products included milk and smear-, veined-, ripened-, brined-, fresh-, cream-, processed- and chemically acidified cheeses. These products are of significant economic importance to the Danish dairy sector, therefore, management of psychrotolerant pseudomonads and Listeria monocytogenes in the studied products is important as their growth affects product shelf-life and safety. Furthermore, prevalence of L. monocytogenes in different types of cheeses was studied by a systematic review and meta-analysis.
The systematic review of the literature showed that L. monocytogenes primarily is involved in outbreaks related to smear- or fresh cheese. The data collected for prevalence of L. monocytogenes in different types of European cheeses revealed that the highest mean prevalence was observed for smear cheese. Meta-analysis of the results demonstrated that prevalence of L. monocytogenes in cheeses produced with un-pasteurized milk was similar to those produced with pasteurized milk highlighting the importance of post-pasteurization contamination.
Existing predictive models for L. monocytogenes were evaluated for their ability to predict growth in different types of cheeses including smear, veined, ripened, fresh and brined cheeses. Predictions were compared with growth responses of L. monocytogenes extracted from literature studies where constant storage temperature and constant product characteristics were assumed. Two models were identified as being able to accurately predict growth of L. monocytogenes in smear cheese and brined cheese, respectively. Challenge tests were performed to collect L. monocytogenes growth data and dynamic product characteristics for smear cheese during production, ripening and storage. Growth of L. monocytognes in smear cheese was correctly predicted from changes in storage temperature and changes in product characteristics by using an existing mathematical model including the effect of temperature, pH, lactic acid concentration in water phase and water activity.
The effect of temperature on the minimum pH that supports growth of L. monocytogenes (pHmin) was quantified in broth studies and by using literature data obtained with different pH-values and different constant temperatures. A model was developed to describe the effect of temperature on the minimum pH for L. monocytogenes growth. A growth and growth boundary model was developed by substituting the constant pHmin-value present in an existing model by the new pHmin-term. Challenge tests where L. monocytogenes was inoculated in chemically acidified cheese (glucono-delta-lactone; GDL) and cream cheese were performed to collect growth responses in low pH foods. In addition, literature data for growth of L. monocytogenes in products with or without GDL were collected. Growth rates of L. monocytogenes were accurately predicted by the new model in a broad range of foods. Growth and no-growth responses of L. monocytogenes in seafood, meat, non-fermented dairy products as well as fermented cream cheese were 90.3% correctly predicted with the incorrect predictions being 5.3% fail-safe and 4.4% fail-dangerous. The new model can support product development, reformulation or risk assessment of a wide variety of foods including meat, seafood and different dairy products (milk, cream, desserts, chemically acidified cheese and cream cheese) with pH-values as low as 4.6.
The antimicrobial effect of phosphate salts necessary to produce spreadable processed cheese was examined in broth studies and their inhibiting effect on L. monocytogenes growth was modelled. It was concluded that emulsifying salts can be used as additional growth hurdle in spreadable processed cheese in order to prevent growth if the pathogen e.g. is introduced by consumer handling after opening the hot-filled packaged. A mathematical model was developed to predict the effect of phosphate salts, lactic-, acetic-, citric acid, pH, aw, temperature and interaction amongst all these factors on growth and the growth boundary of L. monocytogenes in spreadable processed cheese. Challenge tests showed that both growth and the growth boundary were accurately predicted by the developed model. The growth and growth boundary model correctly predicted 54 of 60 growth and no growth responses of L. monocytogenes in spreadable processed cheese. The developed model can be used by the dairy sector to facilitate formulation of safe recipes and this approach seems faster and more cost effective than the traditional challenge testing.
The model described in the previous paragraph was further expanded to contain a term to account for the inhibiting effect of nisin A added as preservative to processed cheese. The antimicrobial activity of nisin A against L. monocytogenes was quantified in broth studies and additional antimicrobial data, obtained at different pH-values, were collected from the literature. A nisin-term was developed to describe the effect of pH on the antimicrobial activity of nisin A. Furthermore, a liquid chromatography/mass spectrometry (LC-MS/MS) method was developed and validated to quantify nisin A and Z present in cheese. Challenge tests were performed to generate data for model evaluation. When the quantified residual nisin A concentrations measured by LC-MS/MS in cheese were used as model input this resulted in accurate predictions of growth for L. monocytogenes. The model can support risk assessment and product development, but further studies with higher residual concentrations of nisin A in cheeses will be beneficial for model validation.
A growth boundary model for psychrotolerant pseudomonas in cottage cheese with cultured cream dressing, raw milk and heat treated milk was developed and validated. The model included terms for the effect of temperature, pH, NaCl/aw, lactic-, sorbic acid and interaction among all factors can be used e.g. to optimized cottage cheese formulations to inhibit growth of psychrotolerant pseudomonads.
The present PhD-project has developed/validated five new predictive growth and growth boundary models that are ready to be applied by the dairy sector. This represents important progress for the use of predictive food microbiology models with dairy products. For further progress in this area there is a need for the dairy sector to increase attention on more detailed chemical product characterization of those dairy products where it is relevant to predict microbial responses.