Safety of Sous-vide Foods30 June 2013
A new generation of foods - often called sous vide foods - that are cooked at low temperatures, is being developed.
Research for the Food Standards Agency in the UK, found the most common uses of sub 60°C cooking temperatures were for steaks, fish and seafood
There is a gap in data for models for food pathogens at these temperatures. The project informed the FSA on this issue and proposed a way forward to fill the knowledge gap.
The research by Dr S.C. Stringer, M.A. Fernandes and Dr A. Metris from the
Gut Health and Food Safety Programme at the Institute of Food Research Enterprises looked at the feasibility of extending the ComBase Predictor database to describe the growth/survival/death response bacterial foodborne pathogens between 40º and 60ºC.
It allowed the calculation of the lethal effect, normally associated with heat treatments at relatively low temperatures, on the reduction of food pathogens.
Traditionally food has been cooked at high temperatures. Sous - vide is a method of cooking food vacuum packed in a sealed plastic pouch, then cooked by submersion in a water bath. Recent years have seen an increase in the number of sous vide foods being cooked at low temperatures (e.g. 42°C to 70°C).
At present growth models for vegetative pathogens go to about 40°C while thermal death models are from about 55 to 60°C upwards.
The research found that there is a lack of information in the range of about 40 to 60°C . This makes it difficult to assess the safety of these new sous vide foods that are cooked in a water-bath at around 50 to 70 °C.
The work gathered information that could be used to assess the hazard associated with lower temperature cooking. It also investigated how much work is needed to upgrade the database used to model the hazards. This helped ensure that the safety assessment for such sous vide foods is consistent, effective and commensurate with any risk to public health.
ComBase (www.combase.cc) is the primary web-based resource for quantitative and predictive food microbiology.
The study was carried out on the feasibility of extending ComBase models to describe the behaviour of E. coli, salmonella and L. monocytogenes between around 40 and 60°C.
This involved assessing the extent and quality of current data, analysis of current data and identification of how much additional data would be required to generate a robust thermal death model for these pathogens. The modelling techniques required to extend the death models in ComBase were assessed and the amount of additional data and work to extend models for ComBase Predictor was identified.
The project explored the feasibility of extending models, such as those in ComBase, to cover low temperature heat treatments. The aim was to help assess pathogen survival, which may contribute to providing improved assessment of sous vide safety.
Data on the heating time/temperatures used in sous - vide cooking were acquired from the internet, cook books, recipes provided from suppliers of sous - vide equipment, directly from chefs and from information elicited from EHOs on sous - vide processed they had observed. Details of 979 cooking times/temperature were gathered for a variety of food types.
This data confirmed that low temperature cooking was being used for several foods. The most common uses of sub 60°C cooking temperatures were for steaks, fish and seafood. The heat treatment used on some of these foods, particularly fish, meant the food was not fully cooked.
Data on thermal death, survival and growth of L. monocytogenes, STEC E. coli and Salmonella at temperatures between 40°C and 60°C were collected from Web of Science, Pubmed, Agricola and Google Scholar data bases. Searching these data sources identified 2,171 potential manuscrips and this was manually reduced to 305 on concideration of the title and abstract. Relevant kinetic data was extracted from 87 of the manuscripts and 1,935 records were added to the ComBase database. The data collected clearly showed that the majority of kinetic thermal death data has been collected at 55°C and above and there was limited kinetic data on growth at temperatures above 40°C.
Although many data on the kinetic behaviours of pathogens between 40°C and 60°C were available, they were not sufficient to create a model that would cover the entire 40°C to 60°C range, the research team said.
This is because there are the fewest data around the growth/no growth region. This is the area where most data are needed as here the kinetics are most variable and uncertain. For example at a single temperature in this region growth, survival or death may be observed depending on the strain, the medium or any pre treatment of the cells.
The observation times used in inactivation experiments were much shorter than the recommended cooking time for sous - vide products potentially neglecting tailing effects that would affect the food safety.
The effect of vacuum was seldom studied and it may be a significant factor. Previous studies have shown that thermal inactivation appears higher for cells in the presence of oxygen than in low oxidising conditions.
It is likely that creation of a model to predict bacterial behaviour between 40°C and 60°C will require the acquisition of new data.
There are many types of models that could be created to compare the behaviour of bacteria in sous – vide systems to those in conventional foods. These vary in the work needed to create them, complexity of use and the quantity of new information that would have to be acquired.
ComBase currently contains separate growth and thermal inactivation models for each pathogen which predict behaviour at a single (static) temperature.
The simplest solution would be to extend the thermal inactivation models to lower temperatures. Extending models to 55°C would be relatively straight forward. Lower temperatures would require more data. It should also be noted that it is not be possible to extend such a model all the way to the growth/no-growth region.
Using kinetic data based on behaviour in standard culture medium could be used to generate a general model for all food. Alternatively specific models could be generated focusing on individual food groups, such as red meat, chicken or pate.
Specific food models would require considerable data collection on the behaviour of bacteria in the relevant foods.
During cooking a food is not at constant temperature: after being placed in the water bath the temperature will gradually increase to the maximum and will gradually decrease during cooling.
The time spent in heating and cooling will contribute to bacterial growth and inactivation. The use of variable temperature profiles instead of single temperatures requires the creation of dynamic models.
These have the advantages that they are the most realistic scenario and they can be made to take into account the possibility of growth following heat treatment.
The problem is that such models cross the growth/no-growth boundary and therefore involve probability of growth in addition to kinetic modelling. A full probabilistic model would require extensive data collection and intensive modelling effort.
Advice on the safety of sous – vide foods should be simple and accessible the report says.
It was found that it is already possible to download apps that calculate the temperature in sous – vide foods and estimate pathogen prediction.
A similar, but slightly more refined, user friendly system could be developed by combining similar temperature calculation methods with the predictive power of ComBase and aimed at UK safety criteria.
This type of system could be used to divide sous – vide products into categories such as pasteurised, unpasteurised or suitable for chilled storage. Suitable safety messages could then be assigned to different food products in each group.
The report says that more guidance on the safe production of sous – vide foods in service settings is required.
A suitable HACCP plan including control points relating to temperature control and guidance on appropriate pathogen inactivation levels combined with user friendly models of temperature and pathogen death could be used to assess the safety of sous - vide products.