VALIDATION OF PREDICTIVE MODELS OF MICROBIAL GROWTH IN FOODS

Ziel

General Background

There are strong, and increasing, demands from consumers for foods that are more convenient, fresher, more natural, less heavily processed, less heavily preserved (e.g. containing less salt, less sugar) and less reliant on additives and preservatives. It is essential that the growth and/or survival of micro-organisms that can spoil, or limit the shelf life of, these, as well as traditional foods, are controlled. At the same time, there is considerable public concern at the high level of foodborne disease in Europe. That concern includes the morbidity and mortality resulting from foodborne illness, and the substantial economic costs involved in its treatment and control.

Improved control of microbial growth and survival are essential if European States are to improve the quality and safety of the food supply with respect to microbial contamination, and to export with confidence, avoiding problems of microbiological hygiene. With the changes in food processing technology, the traditional inspectional approaches to control are proving relatively ineffective. End-product testing is costly, slow and, because of the enormous numbers of units produced and the relatively few that can be tested, offers little assurance of safety, taking into account the statistics of sampling and the sporadic occurrence of the microbes of concern. "Predictive Microbiology" is a powerful tool that can underpin improved control in food processing and distribution, including for new processes where no history is available e.g. high hydrostatic pressure as a means of food preservation. Approaches to modelling microbial responses have developed so much in recent years that a unified database for the food industries, regulatory authorities and other interested parties is conceivable. This would provide a safer food supply for consumers, improved control of shelf life/stability of products for consumers, manufacturers and distributors, and greater assurance of safety for regulatory authorities.

Concerted Action No 5 (Cost 905) "Predictive modelling of microbial growth and survival in foods", which terminated on 31 December 1993 (extended to 31 March 1994), introduced young scientists across Member States to the concept of modelling microbial growth responses and began to develop databases of the responses of key micro-organisms to the processes and conditions to which they are exposed in food processing. Accumulation of relevant microbiological data was standardized and co-ordinated and new approaches to modelling those microbial responses were developed, so that microbial death, survival or growth could be "predicted" using simple computer software. Widespread use of that software is leading to improved control of shelf-life of perishable products and greater assurance of food safety. The Concerted Action initially represented 33 laboratories from the following Member States: Belgium, Denmark, France, Germany, Ireland, Italy, Netherlands, Portugal, Spain, UK. There has also been participation by experts or interested parties from Hungary, Iceland, Norway, Sweden and Switzerland. Many of the participants in these programmes will continue modelling work. In addition, complementary research is underway in labs in the USA, Canada and Australasia.

Progress towards a single European database was slow and difficult, largely because participants were funded by a variety of sources, each with specific research requirements. Consequently, progress was uneven, and the direction of research could rarely be changed. A co-ordinated programme in the UK funded by the Ministry of Agriculture, Fisheries and Food has resulted in a computerized model-base, "Food MicroModel", intended for use by the food industry and other interested parties. "Food MicroModel" includes only foodborne pathogenic bacteria, while the FLAIR Concerted Action covered a wide range of representative food spoilage bacteria, yeasts and moulds. Data generated might ultimately be used in a single system.

Nevertheless, Concerted Action No 5 (COST 905) successfully demonstrated to both research and food industry interests the potential of modelling to improve food safety and shelf-life. The sharing of knowledge and skills has particularly benefited those with no previous experience of modelling in the context of food microbiology. Doubt and scepticism has been replaced by acknowledgement of the benefits and potential of modelling and by enthusiasm to extend the approach. There is gradual recognition that the errors of predictions are of the same order of magnitude as the errors of microbiological experimentation in foods.

Models developed in laboratory media, and taking account of the pH value, the water activity and the storage temperature, have proved appropriate to many foods i.e. the predictions of growth responses from the models are very similar to actual measures of growth or survival published in the scientific literature, thereby "validating" the model.

There is an urgent need to extend those comparisons to broad ranges of foods, representative of European practices and tastes. If foods are identified where the models appear not to be appropriate, details must be systematically recorded, and the reasons for the differences identified. Some food structures may play a role in determining the microbial response and additional research may be needed.

It is crucial that effort be co-ordinated, with carefully planned international collaboration to minimize unnecessary duplication and maximize returns in the shortest reasonable time.

The relevance of mathematical modelling to food microbiology to improve control of food safety and quality has been demonstrated clearly. It is hoped that consideration may be given to continued collaboration within available funding to develop a facility that would benefit both large companies and small and medium-sized enterprises in the food industry in Europe.

A final authoritative report on modelling in the Concerted Action will be published in 1994, promoting modelling as a user-friendly, cost-effective means of improving control of shelf-life and safety of foods.

A major success of the collaborative ventures undertaken so far has been the introduction of similarities of approach (e.g. the use of standard "protocols"), the sharing of information and even the initiation of "inter-laboratory modelling", such that output has been maximized and overlap and repetition minimized. With the introduction in 1994 of the first database(s) and model bases that can be used by the food industries and other interested parties it is most important that this international collaboration is not lost, and that a network of scientists working in Europe and other parts of the world, continue to collaborate and share common approaches, so that output and progress are maximized.

We therefore propose a COST Action to ensure that effect collaboration of laboratories undertaking microbiological modelling work in Europe (list) and would also plan to invite key colleagues in non-European countries (list) to be associated with the initiative. An ultimate objective would be to foster the development of a truly international data- and model-base.

Objectives of the proposed action

There are four key, but inter-related, objectives:

- Validate predictive microbiology models in a wide range of European foods;
- Evaluate instrumental methods for data capture for advanced predictive microbiology;
- Explore mixed population effects in microbial predictive modelling;
- Develop modelling of microbial survival to effectively eliminate pathogens from foods while maintaining quality.

Scientific content

- Validate predictive microbiology models in a wide range of European foods

Predictive microbiology models are being incorporated into substantial databases, and used to help industry and other bodies to ensure with greater certainty the safety and keepability of foods. Generally three or four factors (pH, water activity, temperature) are sufficient to explain microbial growth or survival responses. However, unknown factors in foods occasionally influence those responses, resulting in predictions that are over-cautious. This is an unexpected and important additional outcome of the modelling programme undertaken so far, and suggests a new, original avenue for research.

It is proposed therefore, to target those unknown factors, and to concentrate effort on foods where mis-matches between predicted and observed effects occur. The derived second generation models will contribute to improvements not only in safety assurance, but also shelf-life control, hazard analysis (HACCP) procedures etc, and will lead to greater confidence in new and competitive product innovations.

Foods used in challenge tests will be classified by their agreement with, or difference from, models developed in laboratory media. Physical or chemical factors will be identified which make the identified foods, or food product groups, less conducive to bacterial growth than predicted by models derived from media, and these factors explored to determine if their manipulation could improve the safety and stability of foods.

- Evaluate instrumental methods for data capture for advanced predictive microbiology

Databases used in predictive microbiology have predominantly used viable counts of microbes. Although accurate, these are slow and expensive. Alternative techniques will be explored (e.g. absorbence; turbidimetric methods; impedence; ATP-luminescence; detection-quantification of metabolic products e.g. by GLC; microscopic image analysis techniques; use of "microreactors"; epifluorescence; specific microscopical techniques for rapidly quantifying fungal growth). Some of these techniques are rapid and cheap, and can deal with the large numbers of samples that are necessary for effective model-building, but there are doubts about the relationships of many of the measurements to viable count techniques, which still therefore form the basis for confidence in current models.

We propose to validate rigorously a wide range of modern rapid techniques in media and food systems and to validate and ensure confident "data conversion" schemes. This would promote the use and acceptance of appropriate rapid instrumental methods in future modelling studies to develop models containing all the relevant factors that control microbial responses.

- Explore mixed population effects in microbial predictive modelling

Several classes of foods, including those meat-, dairy- and vegetable-based foods that are deliberately fermented, and also some foods that are not deliberately fermented, carry high numbers of innocuous or beneficial micro-organisms. These influence the potential for growth and survival of others, e.g. food-poisoning or spoilage micro-organisms, in ways which are not yet predictable from current microbiological modelling approaches.

We propose to develop techniques for studying mixed microbial populations, then to generate data and derive new models that will take account of the micro-microbe interactions that occur in foods. This will build on, and improve the value of, models that have already been developed.

- Develop modelling of microbial survival to effectively eliminate pathogens from food while maintaining quality

Whilst predictive modelling so far has concentrated on factors affecting microbial growth, the whole area of inactivation/survival has received far less attention e.g.: in stored foods; in foods during processing by heat (microwaves/ohmic), by irradiation, by surface sprays and "dips"; and by new techniques such as the application of high hydrostatic pressure and high voltage discharges. Survival/inactivation is already very important in ensuring food safety and stability and, as the elimination or control of food poisoning micro-organisms from the most-often contaminated foods becomes an increasingly achievable target, predictive survival/inactivation modelling will increase in significance and value.

We propose to develop suitable methodologies and to model the survival/inactivation of key organisms of food poisoning and spoilage significance in a range of conditions relevant to the main European food commodities and product types, and to major existing, new and "emerging" processing technologies.

Timetable

The planned duration of the project is four years. The detailed time schedule and lists of targets (tasks, milestones), including the definition and creation of working groups to study specific topics, will be established at the first meeting of the Management Committee. Progress will be reviewed annually, achievements documented, and targets for the following year defined.

First year:
The Working Groups (defined for example by interest in particular micro-organisms, food categories or measurements techniques) will develop a realistic working plan, leading to a plan of action for each year, covering both the overall objective and specific topics.

Second year:
After the annual evaluation, further meetings will be organized to optimize the research effort.

Third and fourth years:
The direction of research will depend on the annual evaluations by the Management Committee. A final meeting summarizing the achievements will conclude the Action.

Organization and Management

A Management Committee will be established following the signing of the Memorandum of Understanding by the appropriate number of signatories. At its first formal meeting, this Committee will define its rules of operating in accordance with existing COST regulations. The Management Committee will meet not more than three times per year.

The Action will be co-ordinated by a Chairperson in collaboration with the Action (secretary) scientific officer, who will be responsible for co-ordinating activities and ensuring that the Action meets its overall objectives. Leaders of Working Groups will be selected for each theme within the scientific programme.

The Action co-ordinator and leaders of Working Groups will form a Steering Group which will meet at six monthly intervals to review progress and take action as required.

Information exchange between participants will be facilitated by use of fax or electronic mail.

Original results from the research will be disseminated through joint communications in recognized scientific journals.

It is anticipated that 2 or 3 Workshops per year will be arranged during the COST Action, ideally at one of the participating laboratories, and from which Proceedings will be produced.

A final meeting will conclude the Action.

Annual Reports will be produced for the COST senior officials and a detailed Final Report, based on reports produced throughout the Action, will be compiled.

Economic dimension of the Action

Participation in the FLAIR/COST Action comprised 33 laboratories, with interest expressed by 5 others. Relevant research is in progress in a further 4 European network laboratories. Consequently the estimated effort in the joint European network is of the order of 75 man-years per annum, at an estimated total cost of 3 750 000 ECU/year.
This budget is covered by national sources in participating countries.

Current status
Within COST 914 four Working Groups have been formed to address four key, but inter-related, objectives :
Validate predictive microbiology models in a wide range of European foods;
Evaluate instrumental methods for data capture for advanced predictive microbiology;
Explore mixed population effects in microbial predictive modelling;
Develop modelling of microbial survival to effectively eliminate pathogens from foods while maintaining quality.
WG1. Validate predictive microbiology models in a wide range of European foods
Predictive models are being incorporated into substantial databases and used to help industry and other bodies to ensure the safety and keepability of foods. Generally, such factors as pH, water activity, temperature and the presence of preservatives are sufficient to explain microbial growth or survival responses. However, unknown factors in foods occasionally influence those responses, resulting in predictions that are over-cautious. This unexpected and important additional outcome of the modelling programme suggests a new, original avenue for research.
Foods used in challenge tests are being classified by their agreement with, or difference from, models developed in laboratory media. Physical or chemical factors will be identified which make the identified foods, or food product groups, less conducive to bacterial growth than predicted by models derived from media, and those factors explored to determine if their manipulation could improve the safety and stability of foods.
WG2. Evaluate instrumental methods for data capture for advanced predictive microbiology
Databases used in predictive microbiology have predominantly used counts of viable microbes : accurate, but slow and expensive. Alternative techniques are being explored (e.g. absorbance; turbidimetric methods; impedance; ATP-luminescence; detection-quantification of metabolic products e.g. by GLC; microscopic image analysis techniques; use of `microreactors'; epifluorescence; specific microscopical techniques for rapidly quantifying fungal growth, FTIR, GC-MS).
A wide range of modern rapid techniques in media and food systems are being explored to promote the use and acceptance of appropriate rapid instrumental methods in future modelling studies and to develop models containing all the relevant factors that control microbial responses.
WG3. Explore mixed population effects in microbial predictive modelling
Several classes of foods, including those meat-, dairy- and vegetable-based foods that are deliberately fermented, and also some foods that are not deliberately fermented, carry high numbers of innocuous or beneficial micro-organisms. These organisms sometimes influence the potential for growth and survival of others (e.g. food-poisoning or spoilage micro-organisms) in ways which, using current microbiological approaches, are not yet predictable. Techniques for studying mixed microbial populations are being developed to generate data and derive new models to take account of the microbe-microbe interactions that occur in foods. This will extend and improve models that have already been developed.
WG4. Develop modelling of microbial survival to effectively eliminate pathogens from food while maintaining quality
While predictive modelling has so far concentrated on factors affecting microbial growth, the whole area of inactivation/survival has received far less attention (e.g. in stored foods;in foods during processing by heat -microwave/ohmic-, by irradiation, by surface sprays and `dips'; and by new techniques such as the application of high hydrostatic pressure and high voltage discharges).
Methodologies are being developed to model the survival/inactivation of key organisms of significance in food poisoning and spoilage in a range of conditions relevant to the main European food commodities and product types, and to major existing, new and`emerging' processing technologies.
Widespread use of available software is improving control of shelf-life of perishable products and providing greater assurance of food safety. Doubt and scepticism have been replaced by acknowledgement of the benefits and potential of modelling and by enthusiasm to extend the approach. The errors of predictions are now recognised to be of the same order of magnitude as the errors of microbiological assays in foods.

Programm/Programme

IC-COST - European cooperation in the field of scientific and technical research (COST), 1971-

Thema/Themen

5 - Food Technology

Aufforderung zur Vorschlagseinreichung

Data not available

Finanzierungsplan

Data not available

Koordinator

CEC

EU-Beitrag

Keine Daten

Adresse

Rue de la Loi, 200
1049 Brussels
Belgien

Gesamtkosten