Master of Science / Food Science Institute -- Animal Science & Industry / Daniel Y.C. Fung / In February 2008, the FDA released a draft Compliance Policy Guide (CPG) on Listeria monocytogenes and proposed that ready-to-eat (RTE) foods that do not support the growth of L. monocytogenes may contain up to 100 CFU/g of this pathogen. Frozen foods such as ice cream fall in that category since they are consumed in the frozen state. However, other frozen foods, such as vegetables and seafood that are thawed and served at salad and food bars, may support the growth of Listeria and would not be allowed to contain 100 CFU/g according to the draft CPG. In the current study, growth curves were generated for L. monocytogenes inoculated onto four thawed frozen foods - corn, green peas, crabmeat, and shrimp - stored at 4, 8, 12, and 20ºC. Growth parameters, lag phase duration (LPD), and exponential growth rate (EGR) were determined using a two-phase linear growth model and the Square Root Model. The results demonstrated that L. monocytogenes has a very short LPD on these thawed frozen foods during refrigerated storage and that there would be several orders of magnitude of growth (i.e., more than 1.7 log increase at 4 ºC) of the organism before the product is found to be organoleptically unacceptable. Although it would not be possible to take advantage of any extended lag phase duration caused by freeze injury to the organism, frozen foods containing less than 100 CFU/g of L. monocytogenes that are thawed, or thawed and cooked, and then consumed immediately, should not represent a public health hazard.

Bacterial responses to environmental stresses may be easily observed and predicted under controlled laboratory conditions. However, realistic conditions encountered during manufacturing, in retail or in households may cause unpredicted responses of spoilage or pathogenic bacteria. Therefore it is essential to identify and understand the microbial dynamics under such conditions. The overall aim of the present study was to simulate the most common environmental conditions and consumer-style practices during storage or preparation of Ready-to-Eat (RTE) and Ready-to-Cook (RTC) products in the domestic environment, and predict the microbial dynamics which may deteriorate the quality or compromise the safety of these foods. Aiming to develop a unified mathematical model for the prediction of the growth of the specific spoilage microorganisms (SSOs), the spoilage pattern of three RTE acidic spreads of low pH was described in relation to microbial, physicochemical and molecular changes during storage. Results showed that the spoilage profile of the products was primarily affected by the initial pH and the storage temperature, despite the differences in their formulation. These findings enabled the assessment of two unified models (polynomial and Ratkowsky) for the prediction of the growth of lactic acid bacteria (LAB; SSOs) in such acidic spreads, using only the initial pH, the concentration of undissociated acetic acid and the storage temperature. The models were validated under realistic conditions in household refrigerators. Despite the abrupt fluctuations of the temperature during validation procedure, they both were able to adequately predict the growth of LAB in the spreads. However, the initial contamination level was proved to be necessary and crucial for the accurate prediction of microbial dynamics. The time-temperature profiles of the validation procedure revealed that the suggested storage conditions were not followed promptly and, therefore, concerns were raised on the effect of such consumer mishandlings on the safety of foods. Therefore, the responses of Salmonella spp. and Escherichia coli O157:H7 to the stresses encountered during frozen storage, thawing and cooking of ground beef, simulating typical scenarios followed by the consumers, were evaluated. The results revealed that the guidelines issued by the food safety authorities lack of some specific points that may affect the safety of the final product, such as the duration of frozen storage and the method of cooking. In particular, it was found that the heat resistance of E. coli O157:H7 was likely increased after long term frozen storage, while cooking in pan-grill did not ensure the safety of the final product, even when cooked at the suggested temperature. As shown in the first study, the initial contamination level played a significant role on the predictions of the models and further on the shelf-life of the products. Therefore, the dynamics of realistically low initial populations of Listeria monocytogenes and Salmonella Typhimurium versus higher levels of the pathogens (such those used during in vitro trials) in RTE fresh-cut salads were compared. In addition, any potential uncertainty sources for the growth potential of the pathogens in broth-based simulations were investigated. Results showed that the growth variability of low inocula is highly affected by the marginal storage temperatures, the indigenous microflora and the availability of nutrients. Because of this, growth from low populations showed the likelihood to exceed the growth derived from unrealistically high inocula, suggesting that ―fail-dangerous‖ implications may derive from such challenge tests. Data derived from this part were compared with broth-based simulations and the results showed that high uncertainty should be expected when extrapolating such predictions from low initial populations in fresh-cut salads, due to the various factors affecting the microbial growth on a real food, which are (inevitably) ignored by broth-based models. Overall, the present Thesis highlights the significant impact of consumer mishandlings on the food safety and quality of foods and contributes to the identification of unpredicted potential risk origins in the domestic environment.