Microbiological Risk Assessment Associated with the Food Processing and Distribution Chain
they correspond to the Clostridium type. Identification involved partial sequencing of the coding region of 16S rRNA. Finally, for all isolates identified as C. sporogenes or C. botulinum, specific identification of the C. botulinum genes was carried out according to a variety of protocols (Raphael and Andreadis 2007; Fach et al. 2009; Woudstra et al. 2012).
To estimate the level of risk affecting the French population and to use that to validate the estimated level of risk during the study, epidemiological data covering the period 2001–2012 were collected and analyzed. In France, information on botulism epidemics is compiled each year by the National Institute for Public Health Surveillance (INVS 2013). The data for the 12 years were extracted and then analyzed to estimate the number of cases per million inhabitants in France.
The quantitative risk assessment model estimated, with current consumption, an average probability of illness of 8.0 x 10 -10 per inhabitant and per year. This value is very low, which led to the conclusion that current practices in terms of the thermal processing of canned foie gras were sufficient to control the risk of the presence of C. botulinum in foie gras in France.
It is worth noting here that this study was supported by the Comité Interprofessionnel des Palmipèdes à Foie Gras, a body that brings together all professional stakeholders in the sector. It was carried out by scientists from the academic and professional spheres, not by health authorities. This supports what was mentioned in section 1.2: hazard identification and risk assessment are tools that can be used by a wide-ranging scientific community.
1.5. The hazard identification process
The example developed with foie gras is quite revealing of the approach to hazard identification, to which we now return in more detail.
First of all, precise information must be gathered on the product, its ingredients, its manufacture (decontamination steps, possible recontamination in the factory), its storage method (frozen, refrigerated, ambient) and its method of use (for immediate consumption after opening or for use over a period of time).
Then, attention must be focused on the barrier effect put in place to make it fit for consumption (healthy and wholesome). If, for example, the product is canned and sterilized after packaging, there is no need to focus on vegetative cells that are incapable of withstanding the high-temperature thermal processing (sterilization means a scale of 121.1°C for at least three minutes, or its equivalent). Likewise, if the product is derived from lactic fermentation then for foods with an acidic pH distributed under refrigeration for consumption within a few weeks, such as natural yogurt in France, there is no need to focus on pathogenic bacteria such as Salmonella, the proliferation of which (if present initially) is inhibited by the double-barrier effect of acidity and storage at refrigerated temperature. These two products being, moreover, logically consumed at one sitting, the risk of recontamination on the consumer’s premises is very low. This first rapid analysis, based on the barrier effects put in place, can then be cross-referenced with epidemiological data (Table 1.1) before reaching a firm conclusion on the list of hazards to be retained for the assessment of the potential risk associated with these two types of products.
We mentioned above the use of the end product by the consumer. Unless the epidemiological data (literature or surveillance) indicate the contrary, it is logical to carry out hazard identification taking into consideration consumer practices as stated on the packaging or as reported from consumer surveys. It is futile to consider the “worst-case scenario” as the central point of analysis. For example, in the case of yogurt, it is reasonable to assume its immediate consumption in the minutes after opening. On the other hand, the consumer is not one entity: there are many consumers and many behaviors. This variability in behavior must therefore be taken into consideration when carrying out hazard identification, let alone risk assessment. To illustrate this last point, we will take the example of mayonnaise.
The mayonnaise sold in supermarkets and hypermarkets (homemade mayonnaise is excluded from this) is packaged in France either in a tube or in a jar and stored at the point of sale at room temperature in the grocery section. Before its purchase, this product is stable, and this stability is due to its processing (mainly pasteurized ingredients), its formulation (acetic acid, NaCl, etc.) and its very nature, the emulsion reducing the amount of free water (ICMSF 2005). Before purchasing the product, the barrier effect is therefore threefold: the processes, formulations and nature of the product. On the other hand, once the product is purchased, it is not consumed all in one go but, on the contrary, in several goes. It says on the packaging “to be kept in a cool place and consumed within the month”, but mayonnaise is typically a product that the consumer takes a few liberties with: some consumers keep it in a cool place, others at room temperature (put away with the oil, vinegar and condiments because it is associated with this type of product), some consume it in the week or the month but others consume the whole jar or tube regardless of its initial date of opening (which they have in any case forgotten). None of these behaviors, whether occurring frequently or less frequently, can be ruled out right from the start during the hazard identification stage. The risk of recontamination on the consumer’s premises exists and is even fairly high when the mayonnaise is packaged in a jar: in the middle of the table during the meal, the diners serve and re-serve themselves as they wish. This is thus the reason why in a so-called “industrial” mayonnaise, pathogenic bacteria of the Salmonella type, pathogenic Escherichia coli and Listeria monocytogenes must be taken into consideration. The barrier effect must be sufficient to guarantee their inactivation in the event of recontamination; this effect is based on the emulsion (quantity of oil of at least 65% in the United States) and the formulation: a pH less than 4 in the United States, acetic acid (which, beyond its effect on the pH, has an inhibitory effect, like most weak acids) and addition of preservatives, such as sorbic acid (ICMSF 2005). There is no doubt that consumer demand for natural products without preservatives and/or reduced-fat products will change the formulation of this product, and therefore hazard identification of hazards should be reviewed in line with these changes.
It has been said above that hazard identification requires precise information be gathered on the product, its ingredients, its manufacture, its method of storage and its mode of use. This information must be cross-referenced with epidemiological data, whether sourced from the literature or from national surveillance plans, and with knowledge of the behavior of the hazard (Figure 1.1). Sliced ham, sold in supermarkets, is a very good illustration of the cross-referencing of these three types of information. In fact, the production diagram (set out in the Appendices) shows that there is a cooking process that is sufficient to eliminate the vegetative forms of pathogenic bacteria. On the other hand, this same diagram shows that there is a handling stage (namely slicing) after cooking, which is a possible source of recontamination. At the same time, epidemiological data (literature and surveillance) point to a significant percentage of ham slices contaminated with Listeria monocytogenes during marketing. Finally, knowledge of the behavior of the hazard tells us that L. monocytogenes is quite capable of multiplying in cold conditions. Consequently, in the hazard identification process, L. monocytogenes must be taken into consideration. This example, developed in the Appendices, is taken from an article published in 2016 based on data analyzed in 2015 (Zwietering et al. 2016). The data have not been updated, so there may be a discrepancy between the figures presented and current figures, but the reasoning remains the same.
One final example of hazard identification is linked to refrigerated, vacuum-packed foods that receive heat treatment sufficient to give them a SL of four weeks or more, with no added preservatives. However, this thermal processing is still relatively mild in order to allow these products to keep all their qualities of texture and flavor. This balance is maintained by using knowledge of the main hazard (psychrotrophic Clostridium botulinum) and a specific barrier effect called the degree of protection (DoP). An example concerning vacuum-packed refrigerated products and the DoP principle is presented here. More details are provided in the Appendices. This example is sourced from the