Home style frying of steak and meat products: Survival of Escherichia coli related to dynamic temperature profiles.

Abstract

Microbial survival of heating and cross-contamination are the two transmission routes during food preparation in the consumers' kitchen that are relevant for QMRA (Quantitative Microbial Risk Assessment). The aim of the present study was to extend the limited amount of data on microbial survival during real-life preparation of meat and meat products and to obtain accessory temperature data that allow for a more general (product unspecific) approach. Therefore survival data were combined with extensive measurements of time- and location dependent temperature using an infrared camera for the surface and buttons for the inside of the product, supplemented with interpolation modelling. We investigated the survival of heating of Escherichia coli O111:H2 in beefsteak, hamburgers (beef and 50% beef 50% pork (HH)), meatballs (beef and HH) and crumbs (HH). For beefsteak, survival as a whole is dominated by the sides, giving a log reduction of 1-2 (rare), 3-4 (medium) and 6-7 (done). Limited measurements indicated that done preparation gave 5-6 log reduction for crumbs and at least 8-9 log for the other products. Medium preparation gave a higher reduction in hamburgers (2-4 log) than in meatballs (1-2 log) and in beef (3-4) than in HH (2-3) hamburgers. In general, our 'done' results give larger inactivation than found in literature, whereas 'rare' and 'medium' results are similar. The experiments resulted in two types of curves of D70/z-values, dependent on product, doneness and for beefsteaks sides vs. top/bottom. One type of curve agrees reasonably with literature D70/z estimates from isothermal temperature experiments, which supports using these estimates for home style cooking QMRA calculations. In case of the other type of curve, which is mainly found for (near) surface contamination in close contact with the pan, these literature estimates cannot be applied. We also applied a simplified approach, assuming thermal inactivation is dominated by the highest temperatures reached. The time duration of this highest temperature gives accessory D-values which prove to fit with isothermal temperature literature data, thus suggesting application of such data for QMRA is possible by this approach also, which is less labor intensive both in terms of measurements and modelling. In real life, variability in product properties and preparation styles is large. Further studies are needed to analyze the effect on survival, preferably focusing on determining the essential variables. More variation in heating time will allow for estimating D70/z point estimates rather than curves representing possible sets of D70/z-values.