Principles of Longitudinal and Integrated Food Safety Assurance

11 Principles of Longitudinal and Integrated Food Safety Assurance

11.1 LISA Concept and its Main Elements

General Framework of Modern Food Safety Assurance Systems

Historically, the main approaches to assuring that food is safe to eat included: (i) food inspection; and/or (ii) end product laboratory testing.

Food inspection – for example veterinary meat inspection – has contributed immensely to human health over the past 150 years through organoleptically detecting classical zoonotic diseases in slaughtered animals and eliminating them from the food supply. However, with the passage of time, the nature of meat safety problems has changed. Classical zoonotic diseases became eradicated or very infrequent. Unfortunately, microbial pathogens now causing the majority of food-borne diseases (e.g. Salmonella, Campylobacter, E. coli O157) can be shed by animals showing no clinical symptoms and these diseases are undetectable by conventional meat inspection. They may be detected by another approach, end product (carcass) laboratory testing, but this has been shown to be largely ineffective, the reasons for this including: not every food item can be tested for all pathogens; available testing methods are often insufficiently sensitive; the results obtained are too late to be of use; and such testing does not indicate the root of any problem. In one way or another, these two traditional food safety approaches are reactive, i.e. they deal with problems only after they have appeared.

The basis of our modern food safety assurance system is a novel approach designed to address potential food safety problems before they actually appear (proactively; preventatively), and at points of the food chain where they are expected to appear. Health hazards (harmful agents) enter the food chain at different, sometimes multiple, points; they have to be dealt with at each of those points. However, because events at one point affect the adjacent points of the chain (longitudinal effect), activities at any individual point cannot be effective if applied in isolation. Instead, hazards have to be controlled at relevant, multiple points in a coordinated (integrated) way. Where they cannot be totally eliminated, public health risks can be reduced; it is possible to achieve a ‘summation effect’ of risk reductions in such a longitudinal and integrated system that results in an ultimate risk reduction (i.e. at the moment of food consumption) that would be unachievable using other methods. Understandably, because participants in the food chain are numerous, diverse in profile and activities, the development and application of this ‘farm-to-fork’ system must be both multidisciplinary and science-based.

The commercial basis of the LISA concept lies in the fact that the final product of all individual producers in the food chain (feed producer, farmer, abattoir, processor, and retailer) is the same – food. Unless the produced food is safe, no participant can be economically viable. Therefore, the commercial frame of the LISA concept can be illustrated by using existing examples of longitudinal integration of production operations from farm to supermarket, e.g. the poultry meat chain and the milk/dairy chain; they are often driven by large retailer chains.

Operational Aspects of LISA Concept

The main operational aspect of, and tools for, application of LISA concept are summarized in Fig. 11.1.

To start dealing with public health hazards, it is first necessary to know whether, and where, they exist in the food chain. This information can normally be obtained through monitoring and surveillance programmes that target hazards with both local pre-history and potential newly introduced hazards. In the EU, Directive 2003/99/EC describes conditions and methods of monitoring and surveillance for: (i) zoonoses and zoonotic agents; (ii) antimicrobial resistent agents; (iii) investigation of food-borne outbreaks; and (iv) exchange of information on zoonoses and these agents.

Further, using risk assessment methodology, public health risks from hazards need to be quantified, which enables their ranking. Then, the largest proportion of available scientific and financial resources can be rationally directed towards development and implementation of control systems for hazards posing the highest risks.

Presently, the best available control systems are based on Good Hygiene Practice (GHP) and Hazard Analysis and Critical Control Points (HACCP) principles. These principles can be used globally, i.e. when considering the whole food chain, with identification of global control measures available along it (Fig. 11.1). Furthermore, specific controls applied at individual points are based on development and implementation of GHP and HACCP programmes specifically tailored for each individual producer. Food safety management is, along with food quality management, part of the Total Quality Management system (TQM; Fig. 11.2). The effectiveness of food safety systems, at both food chain- and individual point-level, need to be continuously evaluated after their implementation. This can be achieved, for example, through monitoring and surveillance of hazards targeted by the systems, so as to verify that public health risk reductions for targeted hazards have been achieved, as well as to note any emerging new hazard not yet targeted by existing systems.


Fig. 11.1. Operative aspects of LISA concept: example of the meat chain.

At each point of the food chain, information on the pre-history of the products (or components) entering that point (i.e. Food Chain Information (FCI)) needs to be available, so that these products can be grouped according to the level of risks from particular hazards which they pose, and then handled accordingly. For food safety systems to be effective, both at the global (i.e. food chain) and individual point (i.e. producer) level, product traceability along the whole food chain is a necessary prerequisite. For effective traceability, a product identification system is necessary, enabling correlation of all product components that enter and leave any point of the food chain, as well as between-points (both ‘forward’ and ‘backward’) product correlation. One of the key elements of the LISA concept is timely and two-directional (forward and feedback) flow of information on all relevant aspects of the product to be finally consumed, between all relevant participants in the food chain.


Fig. 11.2. Relationship between food safety management and food quality management systems.

11.2 Risk Assessment of Campylobacter in Poultry



Campylobacter is the leading cause of zoonotic enteric human infections in most developed countries (Anon., 2001). The human cases are usually caused by C. jejuni, and to a lesser extent by C. coli. The high prevalence rates in chicken meat at retail (Anon., 2001), and the fact that case-control studies conducted worldwide repeatedly have identified handling raw poultry and eating poultry products as important risk factors for sporadic campylobacteriosis, seem to support the fact that chickens play an important role in the transfer of Campylobacter to humans.

Quantitative risk assessment can be used as a tool to provide risk managers with information on the influence of different mitigation strategies on the number of human cases associated with thermophilic Campylobacter species in chickens. In this chapter we try, in broad lines, to illustrate the elements and applicability of a formal quantitative risk assessment of human campylobacteriosis caused by chickens.

Risk Assessment Framework

A formal risk assessment includes the steps: (i) hazard identification, which aims to identify the risk of campylobacteriosis associated with thermophilic Campylobacter in chickens; (ii) hazard characterization, which focuses on evaluating the nature of adverse health effects associated with food-borne Campylobacter spp. and the dose–response relationships; (iii) exposure assessment, in which the likelihood and magnitude of exposures to Campylobacter as a result of consumption of a chicken meal are estimated; and (iv) risk characterization, which estimates the risk of campylobacteriosis in a given population for a given set of input data.

A risk model, based on a farm-to-fork approach, was developed to estimate the exposure to Campylobacter from chickens and the number of human cases associated with this exposure (see Fig. 11.3; Rosenquist et al., 2003).

This model details the changes in prevalence and number of Campylobacter on chickens throughout the production line from slaughter to consumption. Module 1 models the transfer and spread of Campylobacter through a chicken slaughterhouse. Module 2 describes the transfer and spread of Campylobacter during food handling in private kitchens and the different consumption patterns. Output distributions from Module 1 were used as input to Module 2, and output distributions from Module 2 were then integrated with the dose-response relationship to estimate the number of human cases associated with thermophilic Campylobacter species in chickens.


Fig. 11.3. Framework of the risk model. Concentration and number of Campylobacter in chickens or in chicken meals.

An overview of the different steps from farm to fork in Danish broiler production can be seen from Fig. 11.4. In the quantitative risk assessment of Campylobacter presented here, only the non-shaded areas are taken into account.

Hazard identification and hazard characterization

As described in the introduction, eating poultry products has been identified as an important risk factor for campylobacteriosis in humans. Thus, having identified the hazard, the next step is to characterize this hazard. Hazard characterization focuses on evaluating the nature of adverse health effects associated with food-borne Campylobacter spp. and on describing the dose–response relationships.

Enteropathogenic Campylobacter may cause an acute enterocolitis, the main symptoms being malaise, fever, severe abdominal pain and watery to bloody diarrhoea. The incubation period varies from 1 to 11 days, typically 1–3 days. In most cases the diarrhoea is self-limiting and may persist for up to a week (Allos and Blaser, 1995). Campylobacter infections may be followed by rare, but severe, non-gastrointestinal sequelae: (i) reactive arthritis, a sterile post-infectious process affecting multiple joints, which is often associated with the tissue phenotype HLA-B27; (ii) the Guillain-Barré syndrome, a demyelinating disorder of the peripheral nervous system resulting in weakness – usually symmetrical – of the limbs, weakness of the respiratory muscles and loss of reflexes, that may become chronic or even mortal; and (iii) the Miller Fisher Syndrome, a variant of the Guillain-Barré syndrome characterized by opthalmoplegia, ataxia and areflexia. Development of antimicrobial resistance, such as the emergence of fluoroquinolone-resistant C. jejuni in humans, may compromise treatment of patients in severe cases where drug treatment is required. In severe cases the drug of choice is usually erythromycin, though fluoroquinolones such as ciprofloxacin and norfloxacin are also used.


Fig. 11.4. Overview of the steps describing the flow of Danish broilers/chickens from farm to fork. Shaded areas are not included in the QRA model. The numbers are the amount of whole chickens in tons in 1998.

Only a few studies describing the human response to a known dose of Campylobacter exist. In one experiment a dose of 500 organisms ingested with milk caused illness in one volunteer. In another experiment involving 111 healthy young adults from Baltimore, Ohio, doses ranging from 800 to 20,000,000 organisms caused diarrhoeal illness (Black et al., 1988). Rates of infection increased with dose, but development of illness did not show a clear dose relation. In an outbreak at a restaurant, the number of C. jejuni in the causative chicken meal was estimated to range from 53 to 750/g. These few investigations indicate that the infective dose of C. jejuni may be relatively low. The data generated by Black et al. (1988) have formed the basis of a dose–response model, which translates the number of organisms an individual is exposed to into an estimate of the individual’s probability of acquiring infection and illness. This estimate is dependent on: (i) the numbers of organisms ingested; (ii) the probability of each individual organism of surviving and infecting the host once it is ingested; and (iii) the probability that the host will become ill once infected. The estimate is also influenced by the virulence of the ingested strain, the vehicle with which it is ingested (Black et al., 1988) and the susceptibility of the individual, e.g. immune status, age and stomach contents.

Exposure assessment

Data on flock prevalence and number of Campylobacter on skin surface throughout the processes of scalding, defeathering, evisceration, washing and chilling are available. In addition, data on the prevalence of Campylobacter-contaminated broilers and on the number of Campylobacter on either chilled or frozen whole carcasses are available.

Two successive mathematical models (Module 1 and Module 2 in Fig. 11.3) were developed to estimate the likelihood and magnitude of exposures to Campylobacter as a result of consumption of a chicken meal. These detailed the prevalence and the number of Campylobacter on chickens throughout the production line, from slaughter to consumption, and the consumption patterns of the consumers. No growth models were included in the exposure assessment, as thermophilic Campylobacter species do not multiply below 32°C (ICMSF, 1996).

After processing, carcasses for sale are stored as either chilled or frozen products. While chilled storage (at 4°C) does not seem to affect the number of Campylobacter considerably, the number of Campylobacter will be reduced due to freezing at −20°C (approximately 0.5–2.5 log units) (Yogasundram and Shane, 1986). No further changes in the number of Campylobacter during transport and storage were considered in the model. The ratio of chilled compared to frozen chicken products sold in retail stores was included.

The transfer of Campylobacter from a Campylobacter-contaminated chicken to the consumer may occur via several contamination routes. Humans may become infected by direct contact, i.e. by licking hands that have been in contact with a chicken or, indirectly, by consuming an undercooked chicken meal or a food item, e.g. salad or prepared chicken, which has been cross-contaminated during handling or preparation of a raw chicken. Since Campylobacter is rather sensitive to heat, the transfer of Campylobacter to humans due to undercooking is assumed to be a rather insignificant event. To simplify the process, only the transfer caused by cross-contamination via unwashed cutting boards was included in the module, as this pathway was assumed to be the most important route of transfer. Hence, Module 2 in the risk model quantifies the transfer of Campylobacter from a contaminated raw chicken to preparation surfaces, and subsequently from these surfaces to ready-to-eat food (salad and prepared chicken). It was assumed that washing the cutting boards, immediately after handling of the raw chicken, would eliminate the risk of cross-contamination. In contrast, if the cutting boards were not washed, there would be a risk of transferring Campylobacter.

Risk characterization

In the risk characterization part, the estimated exposure is integrated with the dose–response model to provide a risk estimate. In most cases the risk estimate itself is not very interesting. Also, there will often be major uncertainties concerning the estimate. However, the ability to run simulations and to observe how the risk estimate changes when different mitigation strategies are applied is a very useful exercise in establishing efficient risk management strategies.

Four different mitigation strategies to reduce the incidence of campylobacteriosis associated with the consumption of chicken meals have been compared, by running Monte Carlo simulations on the quantitative risk model developed to detail the probability of exposure to Campylobacter and the likelihood of campylobacteriosis associated with this exposure.

The simulations indicated that the incidence of campylobacteriosis associated with consumption of chicken meals could be reduced 30 times by introducing a 2-log reduction of the number of Campylobacter on the chicken carcasses. To obtain a similar reduction of the incidence of campylobacteriosis, the flock prevalence should be reduced approximately 30 times (e.g. from 60% to 2%) or the kitchen hygiene improved approximately 30 times (e.g. from 21% not washing the cutting board to 0.7%).

Risk management options

Several countries have implemented, or are at the point of implementing, strategies to reduce the number of Campylobacter -contaminated broiler flocks. Until now establishment of ‘strict hygienic barriers’ or ‘biosecurity zones’ at each poultry house seemed to be the only preventive option shown to work in practice (Reiersen et al., 2001).

The numbers of Campylobacter on chickens may be reduced by introducing different techniques during processing. It is well known that, for example, freezing meat leads to a drop in the concentration of approximately 2 log units (Yogasundram and Shane, 1986). If broiler flocks are examined for Campylobacter prior to delivery to the slaughterhouse, and if a flock is tested positive, then such meat could be sold as a frozen product while Campylobacter-negative flocks could be sold as fresh chicken. This intervention would, according to the risk assessment, be very efficient in lowering the number of human cases of campylobacteriosis.

Other techniques that might have a positive effect on removal or inactivation of Campylobacter are: (i) increasing the scalding temperature; (ii) improving evisceration techniques (to avoid faecal contamination of the meat); (iii) using more water throughout the entire slaughter line; (iv) using forced air-chilling; and (v) introducing disinfectants.

Education of consumers to obtain a reduction of cross-contamination during food handling was included in the model by changing the number of people who did not wash their cutting board during food handling. From the simulations it was obvious that an improvement of the hygiene level in private kitchens (i.e. by washing the cutting board) could reduce the incidence of campylobacteriosis. There was a linear one-to-one relationship between the occurrence of not washing the cutting board and the number of human cases. This means that efforts, directed at improving the frequency of washing the cutting board, for example by a factor of two, would result in a reduction of the incidence of campylobacteriosis by a factor of two.


Allos, B.M. and Blaser, M.J. (1995) Campylobacter jejuni and the expanding spectrum of related infections. Clinical Infectious Diseases 20, 1092–1101.

Anon. (2001) Trends and Sources of Zoonotic Agents in Animals, Feeding Stuff, Food and Man in the European Union and Norway in 1999. Part 1. Document No. SANCO/1069/2001 of the European Commission, Community Reference Laboratory on the Epidemiology of Zoonoses, BgVV, Berlin, Germany.

Black, R.E., Levine, M.M., Clements, M.L., Hughes, T.P. and Blaser, M. (1988) Experimental Campylobacter jejuni infection in humans. Journal of Infectious Diseases 157, 472–479.

ICMSF (1996) Micro-organisms in Foods 5. Characteristics of Microbial Pathogens. Blackie Academic and Professional, London, pp. 45–65.

Reiersen, J., Briem, H., Hardardottir, H., Gunnarsson, E., Georgsson, F. and Kristinsson, K.G. (2001) Human campylobacteriosis epidemic in Iceland 1998–2000 and effect of interventions aimed at poultry and humans. International Journal of Medical Microbiology 291 (Suppl. 31), 153.

Rosenquist, H., Sommer, H., Nielsen, N.L., Nørrung, B. and Christensen, B. (2003) Risk assessment of human illness related to Campylobacter jejuni in Chicken. International Journal of Food Microbiology 83, 87–103.

Yogasundram, K. and Shane, S.M. (1986) The viability of Campylobacter jejuni on refrigerated chicken drumsticks. Veterinary Research Communications 10, 479–486.

11.3 Risk Assessment of Salmonella in Pigs



A marked increase in human cases caused by Salmonella typhimurium DT104 resistant to ampicillin, chloramphenicol (florfenicol), streptomycin (spectinomycin), sulphonamide and tetracycline (R-type ACSSuT) (MRDT104) was recognized from the early 1990s in England (Threlfall et al., 1997). Since, it has spread in animal production in many countries and has become a significant food-borne pathogen internationally (Threlfall et al., 1996; Wall et al., 1997; Tauxe, 1999). In 1996 MRDT104 was isolated for the first time in Denmark from an infected Danish pig herd. The Danish Bacon and Meat Council (DBMC) reacted by deciding to stamp out MRDT104-infected pig herds. The Danish Veterinary and Food Administration (DVFA) followed with a ‘DT104 order’ in 1997, which made the detection of MRDT104 in food animals notifiable and introduced, for the first time, zero tolerance for a pathogen in primary production and in food–including raw meat.

A significant increase in the number of MRDT104-infected herds in Denmark in 1999 and 2000 forced DBMC to stop the destruction strategy for economic reasons. Following this, surveillance was intensified and a Zoonosis Restriction Order was implemented. An important part of the order was the restriction on trade with live animals coming from MRDT104-infected herds. Also, all carcasses from MRDT104-infected slaughter herds should be showered with 80°C hot water for 15 s (called hot water treatment: HWT), which allowed the fresh meat to be distributed for retail. Otherwise, all the meat should be heat processed or condemned. In Fig. 11.5, a schematic presentation is given of the allocation of pigs with Salmonella and MRDT104-infected herds for slaughter in Denmark up to 2003.

In 2002, DBMC applied for a change in the MRDT104 management strategy in primary pig production. The most critical change was a lifting of the trade restrictions for MRDT104-infected herds. This would have been most likely to lead to an increased spread of MRDT104 through piglets produced by MRDT104-infected sow herds. In Denmark, more than 10,000,000 live pigs, mostly piglets, are traded each year. Trade, as such, is a risk factor for spread of pathogens between herds. More than 23,000,000 Danish pigs are slaughtered each year.

The occurrence of MRDT104 is still rather restricted in Denmark. Since the introduction of MRDT104 in Denmark in 1996, very few poultry flocks and cattle herds have been positive for MRDT104. In contrast, more than 100 pig herds have been recognized as MRDT104 infected. In a 1-year period from mid-2001 to mid-2002, 35 pig herds were detected positive for MRDT104, and for the same period The Danish Zoonosis Centre estimated that Danish pork was responsible for approximately 200 human cases of salmonellosis, with MRDT104 from pork being estimated to be responsible for only two human registered cases.


Stay updated, free articles. Join our Telegram channel

Dec 15, 2017 | Posted by in GENERAL | Comments Off on Principles of Longitudinal and Integrated Food Safety Assurance

Full access? Get Clinical Tree

Get Clinical Tree app for offline access