Food Protection Trends - August 2009 - 498

the basis of a hazard analysis that includes historical association of ingredients with Salmonella, prevalence and extent of contamination (i.e., the incoming load of Salmonella), and the intended use of the final product. The selected log reduction should include a margin of safety, e.g., an additional 2-log reduction beyond the extent or levels of contamination expected to occur in the ingredients (21, 25, 41, 42). – Where regulatory or industry standards for log reduction have been established, these should be applied. For example, based on a comprehensive risk assessment a 4-log reduction of Salmonella in raw almonds has been established in the uS to ensure safety of the finished product. ❏ determine the adequacy of the selected control measure and associated critical limits for processing. – Critical limits should be developed on the basis of thermal parameters (e.g., d- and zvalues, thermal death times) or non-thermal parameters of the most resistant and pertinent Salmonella serotype, based on occurrence in the product ingredients, processing environment, and/or association with an outbreak involving the product or similar products. – In many cases, processing conditions are initially driven by quality attributes, and it is essential to determine whether these conditions can deliver the target log reduction (several quick trials in the lab can be done for a feasibility assessment; literature data can also be used). Working with process engineers to optimize the process to deliver the target log reduction while still maintaining product quality is a common approach used in the industry. – In practice, several approaches can be used for validating the adequacy of process parameters. as noted previously, if the process can be mimicked reasonably well in a laboratory (e.g., for oil roasting), then Salmonella can be used in process validation in a laboratory setting to confirm that the critical limits, when achieved, consistently result in the target Salmonella log reduction. if the process is too complex to mimic in a lab setting (e.g., heat extrusion), other approaches for validation may be used, such as determining lethality based on the processing conditions (e.g., integrated lethality based on time and temperature profiles) or using a suitable surrogate for validation on the processing line. in addition to process parameters, other critical factors such as the initial temperature and initial moisture level of the ingredient(s) should also be considered in lethality validation studies. – A non-pathogenic microbial surrogate or a non-microbial surrogate such as an enzyme can be used after appropriate validation. For example, E. faecium nrrl b-2354 has been determined to be an appropriate surrogate for Salmonella in the validation of processing methods for almonds (1). ❏ use published data to guide the determination of whether a challenge study is needed for control measure validation. – The utility of literature data depends on the food or model matrix and the design used in the study to generate the data. according to the rationale outlined by naCmCF (44), the value of a particular set of literature data will be enhanced if the matrix and conditions used to generate the data are similar to the product and process to which the data are being applied. – Available heat resistance data may be used to estimate log reduction by thermal processing in a low-moisture product. the ideal approach is to use available heat resistance data collected in the same food matrix, such as using d- and z-values obtained in wheat flour to calculate log reduction in wheat flour during heat processing. Care should be taken when using d- and z-values, as inactivation may not be linear. in some cases a non-linear heat resistance model may have been developed for a product (e.g., peanut butter, almonds) and this can also be used. When d- and z-values are not available for the food at the water activity under consideration, data obtained with a product of similar composition may be used, e.g., data obtained in wheat flour or corn flour for cereal products. When data for a food matrix are not available, data obtained in a model system (e.g., sucrose solution) with similar aw may be used to estimate lethality. When using this approach, it is important to keep in mind uncertainties inherent in applying available data and assumptions made. – In most cases, literature data are used to guide efforts in identifying parameters specific to a product of interest, whether a challenge study is needed, and how a challenge study may be designed. Whether published data are sufficient to support the adequacy of the lethality of a chosen control measure and associated critical limits depends on several factors. according to the rationale developed from industry experience (49), if an evaluation based on literature data shows that survival of Salmonella is not likely to occur, with a reasonable margin of safety, challenge studies would not be needed. For example, analysis of the time and temperature profiles for a heat extrusion process may indicate that, based on the aw of the ingredients and the product, the process is expected to deliver Salmonella inactivation that would greatly exceed 5 log. on the other hand, if there is less confidence in using published data, then limited challenge studies may be needed to verify estimated log reduction based on literature data. if the evaluation shows that there is limited lethality for the product/process based 498 FOOD PROTECTION TRENDS | AUGUST 2009

Food Protection Trends - August 2009

Table of Contents for the Digital Edition of Food Protection Trends - August 2009