The evolution of microbiological risk assessment 17 HACCP The HACCP concept is a systematic approach to the identification, assessment and control of hazards in a particular food operation. It aims to identify problems before they occur and establish measures for their control at stages in production that are critical to ensuring the safety of food. Control is proactive, since remedial action is taken in advance of problems occurring In a review of the historical background, Barendsz(1995)and Untermann et al.(1996)described the development of the HACCP approach, which began in the 1960s. The concept arose from a collaboration between the pillsbury Company, the US Army Natick Research and Development Laboratories and the US National Aeronautics and Space Administration. The original purpose was to establish a system of safe food production for use in human space travel. At that me,the limitations of end-product testing were already appreciated and therefore more attention was given to controlling the processes involved in food production and handling. When first introduced at a meeting on food protection (Department of Health, Education and Welfare, 1972), the concept involved three principles: (i) hazard identification and characterisation; (ii)identification of critical control points(CCPs)and(iii)monitoring of the CCPs Many large food companies started to apply HACCP principles on a voluntary basis, and in 1985 the US National Academy of Science recommended that the system should be used. Further support came from the ICMSF (1988), which extended the concept to six principles. They added specification of criteria, corrective actions and verification(see Table 2.3). In 1989, the Us National Advisory Committee on Microbiological Criteria for Foods(NACMCF)added in a further principle: the establishment of documentation concerning all procedures and records appropriate to the principles and their application. Use of the HACCP system was given an international dimension by the Codex Alimentarius Commission(CAC) which published details of the principles involved in 1991 and their practical application(CAC, Committee on Food Hygiene, 1991). In 1997, the CAC laid down the final set of principles and clarified the precise meaning of the different terms(CAC, Committee on Food Hygiene, 1997) General principles of food hygiene(Alinorm 97/13, Appendix D) HACCP system and guidelines for its application(Alinorm 97/13A A Principles for the establishment and application of microbiological criteria for foods(Alinorm 97/13A, Appendix IlD) he full HACCP system, as described in Alinorm 97/13, is shown in Table 23. The document also gives guidelines for practical application of the HACCP ystem. By 1973, the FDA had made the use of HACCP principles mandatory for the production of low-acid canned foods(FDA, 1973) and, in 1993, the system became a legal requirement for all food products in the European Union ective 93/43) Despite widespread usage, the present HACCP concept still has some weak points. One of them is the definition of a hazard. This is not defined asan agent
HACCP The HACCP concept is a systematic approach to the identification, assessment and control of hazards in a particular food operation. It aims to identify problems before they occur and establish measures for their control at stages in production that are critical to ensuring the safety of food. Control is proactive, since remedial action is taken in advance of problems occurring. In a review of the historical background, Barendsz (1995) and Untermann et al. (1996) described the development of the HACCP approach, which began in the 1960s. The concept arose from a collaboration between the Pillsbury Company, the US Army Natick Research and Development Laboratories and the US National Aeronautics and Space Administration. The original purpose was to establish a system of safe food production for use in human space travel. At that time, the limitations of end-product testing were already appreciated and therefore more attention was given to controlling the processes involved in food production and handling. When first introduced at a meeting on food protection (Department of Health, Education and Welfare, 1972), the concept involved three principles: (i) hazard identification and characterisation; (ii) identification of critical control points (CCPs) and (iii) monitoring of the CCPs. Many large food companies started to apply HACCP principles on a voluntary basis, and in 1985 the US National Academy of Science recommended that the system should be used. Further support came from the ICMSF (1988), which extended the concept to six principles. They added specification of criteria, corrective actions and verification (see Table 2.3). In 1989, the US National Advisory Committee on Microbiological Criteria for Foods (NACMCF) added in a further principle: the establishment of documentation concerning all procedures and records appropriate to the principles and their application. Use of the HACCP system was given an international dimension by the Codex Alimentarius Commission (CAC) which published details of the principles involved in 1991 and their practical application (CAC, Committee on Food Hygiene, 1991). In 1997, the CAC laid down the ‘final’ set of principles and clarified the precise meaning of the different terms (CAC, Committee on Food Hygiene, 1997): • General principles of food hygiene (Alinorm 97/13, Appendix II). • HACCP system and guidelines for its application (Alinorm 97/13A, Appendix II). • Principles for the establishment and application of microbiological criteria for foods (Alinorm 97/13A, Appendix III). The full HACCP system, as described in Alinorm 97/13, is shown in Table 2.3. The document also gives guidelines for practical application of the HACCP system. By 1973, the FDA had made the use of HACCP principles mandatory for the production of low-acid canned foods (FDA, 1973) and, in 1993, the system became a legal requirement for all food products in the European Union (Directive 93/43). Despite widespread usage, the present HACCP concept still has some weak points. One of them is the definition of a hazard. This is not defined as ‘an agent The evolution of microbiological risk assessment 17
18 Microbiological risk assessment in food processing Table 2.3 The seven principles of the HACCP system(CAC, Committee on Food Hygiene, 1997) List all potential hazards associated with each step, conduct a hazard analysis, and consider any measures to control identified hazard Determine the critical control points Determine CCPs (CCPs Establish critical limit(s) Establish critical limits for each CCP 4 Establish a system to monitor control Establish a system of monitoring for each of the ccp CCP 5 Establish corrective actions Establish the corrective action to be taken hen monitoring indicates that a particular CCP is not under control 6 Establish verification procedures Establish procedures for verification to confirm that the HACCP system is 7 Establish documentation and record Establish documentation concerning al procedures and records appropriate to these principles and their application with the potential to cause an adverse health effect, as in risk assessment, but as an unacceptable contamination, growth and/or survival by microorganisms of concern(ICMSF, 1988), which is more restrictive and does not cover al ossible hazards. Another weakness arises from the definition of a ccp It is stated that a CCP is a location, practice, etc. where hazards can be minimised (CMSF, 1988; International Association of Milk, Food and Environmental Sanitarians(IAMFES), 1991)or reduced to an acceptable level (Bryan, 1992 Alinorm 97/13). In both cases, these are qualitative objectives and may lead to differing interpretations. It was Notermans et al.(1995)who first made a plea to use the principles of quantitative risk assessment for setting critical limits at the CCPs(process performance, product and storage criteria). It was their opinion that only when the critical limits are defined in quantitative terms can the level of control at the CCPs be expressed realistically. At the International Association of Food Protection (IAFP) meeting in 2001, Buchanan et al (2001)also favoured the use of these principles and suggested that food safety bjectives should encompass end-product criteria, which are related to the criteria used in processing 2.3.6 Predictive modelling Modelling in food microbiology began about 1920, when methods were leveloped for calculating thermal death times. These models revolutionised the anning industry (Pflug and Gould, 2000). Later, Monod (1949, 1950) developed a model that described the continuous, steady-state culture of
with the potential to cause an adverse health effect’, as in risk assessment, but as ‘an unacceptable contamination, growth and/or survival by microorganisms of concern’ (ICMSF, 1988), which is more restrictive and does not cover all possible hazards. Another weakness arises from the definition of a CCP. It is stated that a CCP is a location, practice, etc. where hazards can be minimised (ICMSF, 1988; International Association of Milk, Food and Environmental Sanitarians (IAMFES), 1991) or reduced to an acceptable level (Bryan, 1992; Alinorm 97/13). In both cases, these are qualitative objectives and may lead to differing interpretations. It was Notermans et al. (1995) who first made a plea to use the principles of quantitative risk assessment for setting critical limits at the CCPs (process performance, product and storage criteria). It was their opinion that only when the critical limits are defined in quantitative terms can the level of control at the CCPs be expressed realistically. At the International Association of Food Protection (IAFP) meeting in 2001, Buchanan et al. (2001) also favoured the use of these principles and suggested that food safety objectives should encompass end-product criteria, which are related to the criteria used in processing. 2.3.6 Predictive modelling Modelling in food microbiology began about 1920, when methods were developed for calculating thermal death times. These models revolutionised the canning industry (Pflug and Gould, 2000). Later, Monod (1949, 1950) developed a model that described the continuous, steady-state culture of Table 2.3 The seven principles of the HACCP system (CAC, Committee on Food Hygiene, 1997) Principle Activity 1 Conduct a hazard analysis List all potential hazards associated with each step, conduct a hazard analysis, and consider any measures to control identified hazards 2 Determine the critical control points (CCPs) Determine CCPs 3 Establish critical limit(s) Establish critical limits for each CCP 4 Establish a system to monitor control of the CCP Establish a system of monitoring for each CCP 5 Establish corrective actions Establish the corrective action to be taken when monitoring indicates that a particular CCP is not under control 6 Establish verification procedures Establish procedures for verification to confirm that the HACCP system is working effectively 7 Establish documentation and record keeping Establish documentation concerning all procedures and records appropriate to these principles and their application 18 Microbiological risk assessment in food processing
The evolution of microbiological risk assessment 19 microorganisms and became the basis for continuous fermentation processes. In principle, the model was analogous to that used for chemical processes. The recent resurgence of predictive modelling in relation to microbial growth in food originated in the 1960s and has been reviewed by ross and MacMeekin(1993) In contrast to the situation studied by Monod, the identities and concentrations of nutrients involved are unknown and the organisms of interest are initially present in low numbers, with growth conditions often being less than optimal For these reasons, initial attempts at mathematical modelling in food microbiology have been more empirical than was the case for fermentation rocesses. focusing on batch rather than continuous-culture kinetics. As shown y Whiting and Buchanan(1997), growth data are fitted to equations using interactive least-square computer algorithms. Assumptions about randomness normal distribution and stochastic specifications are the same as they would be for any statistical application of regression(Ratkowsky, 1993). All models are actually simplifications that represent the complex biochemical processes controlling microbial growth and are limited to the most important input parameters, such as temperature, time, water activity and pH. One of the reasons for simplifying the approach is that knowledge of the complex biochemical processes involved is far from complete. The major advantage is that the current models are easy to handle;, however, the outcome should always be used with caution and verification may be necessary in some cases Primarily, the development of predictive modelling was driven by the proliferation of refrigerated and limited shelf-life food products. It was recognised that (i)even so-called rapid' microbiological methods were too slow to show within an acceptable period of time, whether microbes in the product grew or died (Spencer and Bains, 1964);(i) testing of factors in a food product that affect microbial growth and toxin production, whether singularly or in combination, laborious and time-consuming and (iii) work had been done in Canada to draw together the results of numerous growth experiments on Clost. botulinum (Hauschild, 1982). The mathematical and statistical tools already existed prior to the expansion in modelling activity and the process was favoured by the introduction of powerful personal computers and the availability of user-friendly software In the review of Ross and MacMeekin (1993), the main reasons for developing predictive models were summarised as follows To permit predictions of product shelf-life and safety, and the consequences of changes in product formulation or composition; to facilitate a rational design for new processes, etc, to meet or to obtain an insight into requirements for product safety or shelf-life To allow objective evaluations to be made of processing operations and, fr this, an empowering of the HACCP approach To provide an objective evaluation of the consequences of any lapses in process control and subsequent storage of the end-product Now that MRA has become established in food microbiology, it is clear that the use of predictive models is essential in risk assessment. This is especially
microorganisms and became the basis for continuous fermentation processes. In principle, the model was analogous to that used for chemical processes. The recent resurgence of predictive modelling in relation to microbial growth in food originated in the 1960s and has been reviewed by Ross and MacMeekin (1993). In contrast to the situation studied by Monod, the identities and concentrations of nutrients involved are unknown and the organisms of interest are initially present in low numbers, with growth conditions often being less than optimal. For these reasons, initial attempts at mathematical modelling in food microbiology have been more empirical than was the case for fermentation processes, focusing on batch rather than continuous-culture kinetics. As shown by Whiting and Buchanan (1997), growth data are fitted to equations using interactive least-square computer algorithms. Assumptions about randomness, normal distribution and stochastic specifications are the same as they would be for any statistical application of regression (Ratkowsky, 1993). All models are actually simplifications that represent the complex biochemical processes controlling microbial growth and are limited to the most important input parameters, such as temperature, time, water activity and pH. One of the reasons for simplifying the approach is that knowledge of the complex biochemical processes involved is far from complete. The major advantage is that the current models are easy to handle; however, the outcome should always be used with caution and verification may be necessary in some cases. Primarily, the development of predictive modelling was driven by the proliferation of refrigerated and limited shelf-life food products. It was recognised that (i) even so-called ‘rapid’ microbiological methods were too slow to show, within an acceptable period of time, whether microbes in the product grew or died (Spencer and Bains, 1964); (ii) testing of factors in a food product that affect microbial growth and toxin production, whether singularly or in combination, is laborious and time-consuming and (iii) work had been done in Canada to draw together the results of numerous growth experiments on Clost. botulinum (Hauschild, 1982). The mathematical and statistical tools already existed prior to the expansion in modelling activity and the process was favoured by the introduction of powerful personal computers and the availability of user-friendly software. In the review of Ross and MacMeekin (1993), the main reasons for developing predictive models were summarised as follows: • To permit predictions of product shelf-life and safety, and the consequences of changes in product formulation or composition; to facilitate a rational design for new processes, etc.; to meet or to obtain an insight into requirements for product safety or shelf-life. • To allow objective evaluations to be made of processing operations and, from this, an empowering of the HACCP approach. • To provide an objective evaluation of the consequences of any lapses in process control and subsequent storage of the end-product. Now that MRA has become established in food microbiology, it is clear that the use of predictive models is essential in risk assessment. This is especially The evolution of microbiological risk assessment 19
20 Microbiological risk assessment in food processing true for exposure assessment. In many foods, particularly those that are fresh and have a short shelf-life, rapid changes in microbial populations can occur and the models are needed to determine, for example, the necessary storage conditions The models can also provide information about risk factors in handling the product, which have a considerable influence on human exposure to particular pathogens. They may also help to clarify the effects of different control options Thus, the modelling approach facilitates control of the most important facto that affect food safety. Without the use of predictive models, a quantitative MRA for assessing food safety would be virtually impossible 2.3.7 Introduction of Qra Systematic risk analysis approaches have been used by the Food and Agriculture Organisation of the United Nations(FAO)and the World Health Organisation (WHO) since 1955, when the evaluation of food additives at the international level was initiated as a result of a joint FAO/HO conference on food additives The conference recommended to the directors general of fao and who that one or more expert committees should be convened to address the technical and administrative aspects of chemical additives and their safety in food. This recommendation provided the basis for setting up the Joint FAO/WHO Expert Committee on Food Additives (JECFA). The JECFA started its meetings in 1956, initially to evaluate the safety of food additives Risk assessment has also evolved over the last decade within the cac. the Commission, which was established in 1962 under the parentage of the FAO and WHO, is an intergovernmental organisation with the responsibility for veloping international standards, guidelines or other recommendations fo food in order to protect the health of consumers and facilitate international trade In the course of time the cac has enlarged its activities and in addition to risk evaluation for food additives, chemical contaminants, pesticide residues and ternary drug residues in foods, the issue of biological hazards in foods is now lso being addressed. However, no clear MRA activities were undertaken prior to1995 The development of MRa was strongly stimulated when in 1995, at the GATT Uruguay Round, the WTO was established and a free trade in safe food was agreed. In the WTo Agreement on the Application of Sanitary and Phytosanitary Measures, the so-called SPS Agreement(Anon, 1995), requires that countries signatory to the agreement base their laws concerned with rotecting human, animal and plant health on a risk analytical basis. Thus, the SPS Agreement requires food safety legislation to be scientifically based and the process of risk assessment to be applied, for example, when introducing microbiological criteria for controlling imported foods. In the pursuance of harmonisation and to avoid the need for all countries and all food producers to carry out a risk assessment on each of their products, the WTO SPS Agreement has chosen the scientifically based international standards, guidelines and recommendations of three organisations, one of which is the CAC
true for exposure assessment. In many foods, particularly those that are fresh and have a short shelf-life, rapid changes in microbial populations can occur and the models are needed to determine, for example, the necessary storage conditions. The models can also provide information about risk factors in handling the product, which have a considerable influence on human exposure to particular pathogens. They may also help to clarify the effects of different control options. Thus, the modelling approach facilitates control of the most important factors that affect food safety. Without the use of predictive models, a quantitative MRA for assessing food safety would be virtually impossible. 2.3.7 Introduction of QRA Systematic risk analysis approaches have been used by the Food and Agriculture Organisation of the United Nations (FAO) and the World Health Organisation (WHO) since 1955, when the evaluation of food additives at the international level was initiated as a result of a joint FAO/WHO conference on food additives. The conference recommended to the Directors-General of FAO and WHO that one or more expert committees should be convened to address the technical and administrative aspects of chemical additives and their safety in food. This recommendation provided the basis for setting up the Joint FAO/WHO Expert Committee on Food Additives (JECFA). The JECFA started its meetings in 1956, initially to evaluate the safety of food additives. Risk assessment has also evolved over the last decade within the CAC. The Commission, which was established in 1962 under the parentage of the FAO and WHO, is an intergovernmental organisation with the responsibility for developing international standards, guidelines or other recommendations for food in order to protect the health of consumers and facilitate international trade. In the course of time, the CAC has enlarged its activities and, in addition to risk evaluation for food additives, chemical contaminants, pesticide residues and veterinary drug residues in foods, the issue of biological hazards in foods is now also being addressed. However, no clear MRA activities were undertaken prior to 1995. The development of MRA was strongly stimulated when in 1995, at the GATT Uruguay Round, the WTO was established and a free trade in safe food was agreed. In the WTO Agreement on the Application of Sanitary and Phytosanitary Measures, the so-called SPS Agreement (Anon., 1995), requires that countries signatory to the agreement base their laws concerned with protecting human, animal and plant health on a risk analytical basis. Thus, the SPS Agreement requires food safety legislation to be scientifically based and the process of risk assessment to be applied, for example, when introducing microbiological criteria for controlling imported foods. In the pursuance of harmonisation and to avoid the need for all countries and all food producers to carry out a risk assessment on each of their products, the WTO SPS Agreement has chosen the scientifically based international standards, guidelines and recommendations of three organisations, one of which is the CAC, as the 20 Microbiological risk assessment in food processing
The evolution of microbiological risk assessment 2 preferred measures for adoption by WTO members. In addition, the SPs Agreement states that countries should take into account the risk assessment technique developed by the relevant international organisations, when undertaking a risk assessment. As a result of this, the FAo and wHo began to organise expert consultations dealing with food safety risk assessment, with the purpose of providing member countries with principles and guidelines for undertaking such an assessment. An overview of the key documents produced is given in Table 2. 4 The first expert consultation was devoted to the application of risk analysis to food safety standards issues. The consultation was convened at the request of the Forty-first Session of the CAC Executive Committee, with the aim of promoting consistency in the use of risk analysis for food safety purposes. The main jective was to provide the FAO, WHO and CAC, as well as member countries with advice on practical approaches for the application of risk analysis to food standards issues. At that meeting, food safety risk analysis terms were defined. A model for risk assessment was also agreed upon. This comprises the four components:(i) hazard identification, (ii) hazard characterisation, (iii) exposure ssessment and(iv)risk characterisation. At that consultation, the estimation of risk from biological agents was considered in detail, since it was the general view of the experts that such risks are in many ways a much larger and more immediate problem to human health than risks associated with chemical contaminants in food At an expert consultation in 1997, a risk management framework was set up and general principles of food safety risk management were elaborated. In addition, key risk management terms were defined. The main elements of risk management were identified as (1 risk evaluation, (ii) assessment of risk management options, (iii)implementation of management decisions and(iv) monitoring and review. As far as the general principles are concerned, it was stated that risk management decisions should be transparent, primarily aimed at the protection of human health and should ensure that the scientific integrity of the risk assessment process is maintained As a logical continuation, a third expert consultation dealt with the application of risk communication. The main issues addressed at this meeting Table 2. 4 FAO/WHO documents dealing with food-related risk analysis Risk analysis documents References 1995 Application of risk analysis to food standards issues FAO/WHO (1995) 1997 Risk management and food safery FAO/WHO (1997) 1998 The application of risk communication to food FAO/WHO (1998) standards and safety matters 1999 Risk assessment of microbiological hazards in foods FAO/WHO (2000a) 2000 The interaction between assessors and managers of FAO/Ho (2000b microbiological hazards in food
preferred measures for adoption by WTO members. In addition, the SPS Agreement states that countries should take into account the risk assessment technique developed by the relevant international organisations, when undertaking a risk assessment. As a result of this, the FAO and WHO began to organise expert consultations dealing with food safety risk assessment, with the purpose of providing member countries with principles and guidelines for undertaking such an assessment. An overview of the key documents produced is given in Table 2.4. The first expert consultation was devoted to the application of risk analysis to food safety standards issues. The consultation was convened at the request of the Forty-first Session of the CAC Executive Committee, with the aim of promoting consistency in the use of risk analysis for food safety purposes. The main objective was to provide the FAO, WHO and CAC, as well as member countries, with advice on practical approaches for the application of risk analysis to food standards issues. At that meeting, food safety risk analysis terms were defined. A model for risk assessment was also agreed upon. This comprises the four components: (i) hazard identification, (ii) hazard characterisation, (iii) exposure assessment and (iv) risk characterisation. At that consultation, the estimation of risk from biological agents was considered in detail, since it was the general view of the experts that such risks are in many ways a much larger and more immediate problem to human health than risks associated with chemical contaminants in food. At an expert consultation in 1997, a risk management framework was set up and general principles of food safety risk management were elaborated. In addition, key risk management terms were defined. The main elements of risk management were identified as (i) risk evaluation, (ii) assessment of risk management options, (iii) implementation of management decisions and (iv) monitoring and review. As far as the general principles are concerned, it was stated that risk management decisions should be transparent, primarily aimed at the protection of human health and should ensure that the scientific integrity of the risk assessment process is maintained. As a logical continuation, a third expert consultation dealt with the application of risk communication. The main issues addressed at this meeting Table 2.4 FAO/WHO documents dealing with food-related risk analysis Year Risk analysis documents References 1995 Application of risk analysis to food standards issues FAO/WHO (1995) 1997 Risk management and food safety FAO/WHO (1997) 1998 The application of risk communication to food standards and safety matters FAO/WHO (1998) 1999 Risk assessment of microbiological hazards in foods FAO/WHO (2000a) 2000 The interaction between assessors and managers of microbiological hazards in food FAO/WHO (2000b) The evolution of microbiological risk assessment 21