Microbiological Risk Assessment of Meat: A Veterinary Perspective. Dr. Kathryn H. Christiansen, Epidemiology Department, Central Veterinary Laboratory, Woodham Lane, New Haw, Addlestone, Surrey KT15 3NB, UK
INTRODUCTION
The annual number of notified cases of food poisoning in the UK has risen fourfold over the last 15 years (7). Food animals are important reservoirs of many food poisoning organisms and these organisms may be transmitted to humans through the consumption of contaminated meat. Poultry meat is a significant source of the two most common causes of foodborne illness in the UK, Campylobacter and Salmonella, while beef products have been implicated as a source of Escherichia coli O157:H7 infection for humans. Traditional organoleptic meat inspection procedures have proved inadequate to detect or control these diseases, which are largely asymptomatic in animals.
New procedures are required to improve the microbiological safety of meat. The most promising is the hazard analysis critical control point system (HACCP). A HACCP system for meat involves assessing the hazards and risks associated with meat at each step of the food chain, identifying critical control points necessary to prevent or control the hazards, monitoring these points and taking corrective action if necessary. The critical control points where veterinarians have an interest are the farm and the slaughterhouse. If pathogen levels in food animals and the slaughterhouse environment can be reduced, the incidence of foodborne disease in the human population should fall.
A substantial body of research has accumulated on pathogen levels in slaughter animals and on carcasses and on the effectiveness of measures to reduce them. However, for economic, practical or social reasons, farmers and industry frequently fail to adopt such measures or to realise their full potential.
A public anxious about food safety has put considerable pressure on government to raise food safety standards. This concerns the food industry as higher standards generally mean increased costs. As governments seek to transfer the cost of assuring food safety to industry, food safety standards will be challenged more vigorously by industry and consumer groups. The World Trade Organisation has adopted health risk assessment as the main way to scientifically justify food safety standards. This paper examines the usefulness of a risk assessment approach to evaluate the impact of control measures at the farm and slaughterhouse levels on the risk to public health from microbial hazards in meat.
HEALTH RISK ASSESSMENT
Health risk assessment is the scientific process of estimating the risk of adverse effects on human health from exposure to hazardous substances or agents. In 1983, the United States National Academy of Sciences (NAS) published a framework for a health risk assessment of chemical hazards in the environment (4). This model has proved useful for quantifying the risk to human health from chemical hazards in food, water and the environment but its extension to microbial hazards has been problematic (3,6).
The NAS risk assessment model comprises four steps: hazard identification, dose-response assessment, exposure assessment and risk characterisation. In terms of microbiological risk assessment, hazard identification involves determining whether a particular microorganism may adversely affect human health. If a microorganism has no effect on health, there is no risk. If there is evidence for an association between an organism and ill-health, then the next step is to quantify the risk that it poses to human health.
The dose-response relationship is the key element of a health risk assessment. It is the probability or risk that an individual will become ill following exposure to a specified dose of the pathogen. Evaluation of the degree of human exposure likely to occur to various doses of a pathogen forms the exposure assessment. As people may be exposed to a pathogen in many different situations, this step usually involves the elaboration of various exposure scenarios.
With knowledge of exposure levels to a pathogen, the dose-response relationship can be used to determine the risk of infection or illness in the exposed population. This process of integrating the dose-response data with the exposure data comprises the risk characterisation step of a health risk assessment. It also includes a summation of the biological and statistical uncertainties revealed by the risk assessment process.
The problems associated with applying quantitative risk assessment techniques to microbial hazards in food derive from
Because of the complexity of the subject, health risk assessment of microbial hazards in food has remained largely qualitative. Rose and Sobsey (5) published a quantitative health risk assessment of viruses in shellfish in the USA in 1993. However, quantitative risk assessment of microbial hazards in meat has been described only in broad terms (1,2,3).
MICROBIOLOGICAL RISK ASSESSMENT OF MEAT
Hazard identification
In the UK, veterinarians are responsible for assuring the safety and wholesomeness of meat up to the time that the product leaves the slaughterhouse or meat packing plant. Therefore we have a particular interest in microorganisms that may enter the human food chain on or in food animals and infect humans through the consumption of, or contact with, meat.
Knowledge about the pathogenicity for humans of microorganisms found in food animals and their products has been growing rapidly, especially since the development of molecular techniques for detecting and characterising microorganisms in the late 1970s. Over the last twenty years, pathogens that have joined Salmonella as significant causes of meat-borne diseases include Campylobacter jejuni and Escherichia coli O157: H7. However, the role of food animals and meat in the transmission of many foodborne pathogens to humans remains unclear. Genetic typing of microorganisms is challenging established theories about host-parasite relationships, so hazard identification will be a moving target for some time to come.
Dose-response assessment
Because the pathogenicity of microorganisms differs markedly among host species, we cannot use experimental studies in animals as a proxy for dose-response relationships in humans as is done for chemical hazards. Published data from feeding studies using human volunteers do exist for some pathogens but ethical considerations limit the use of this method.
This means that we must use indirect ways to quantify the risk to human health from microbial hazards. One method is to obtain data from epidemiological investigations of foodborne disease outbreaks. With knowledge of attack rates associated with the consumption of different amounts of the implicated food and the concentration of the pathogen in the food, we can estimate the dose-response relationship, albeit with qualifications.
Exposure assessment
To estimate the risk that an individual will become ill from exposure to a microbial hazard in food, we need to know the average level of contamination of a single serving of the food at the time of consumption. However, as this is rarely known, contamination levels of the food at retail are used as a proxy for levels consumed. This works well for chemical hazards, which are stable in foods, but is not very satisfactory for microorganisms, which may multiply or die in food between purchase and consumption or be destroyed by cooking. To avoid overestimating the risk of exposure to microbial hazards in food, the effect of the consumer's treatment of the food on microbial load should be assessed.
Unfortunately for veterinary regulators, exposure assessment does not end with estimates of pathogen levels in retail meats. To evaluate the impact of interventions at the beginning of the meat food chain on the risk to public health from microbial hazards in meat, we must find a way to predict the effect of infection rates in slaughter animals or contamination rates in carcasses on pathogen levels in meat as it is finally consumed.
The method that holds most promise is to model potential concentrations of microorganisms in meat from "farm to fork" under different conditions. Data on the response of meat-borne pathogens to environmental conditions are available in the scientific literature. The difficulty comes from the complexity of the system being modelled. Carcasses are broken down into many different products, which are subjected to a wide range of conditions along the food chain. Therefore, to model such a system, a large number of exposure scenarios are required. We also have to find ways of accounting for the uneven distribution of microorganisms in carcasses and meat and the often unforeseen chain of events that leads to an outbreak of foodborne disease.
Risk characterisation
From the dose-response relationship for a specific pathogen and the exposure assessment, we can estimate the probability of illness from the consumption of a single serving of meat and the expected number of cases of meat-borne illness in the population under various scenarios. If we couple this with knowledge of the relationship between microbial loads in animals or carcasses and the risk of foodborne disease in the human population, we can assess the effect of a reduction in pathogen levels in animals or carcasses on the incidence of foodborne disease. Finally, if the effect of an intervention on microbial levels in animals or carcasses is known, the impact of the intervention on the number of human cases of foodborne disease in the population can be evaluated.
CONCLUSIONS
Risk assessment has considerable potential as a tool to evaluate the impact of changes in farm and slaughterhouse practices on public health. The main problems in carrying out a successful risk assessment of meat-borne microbial hazards are methodological, namely:
Until quantitative risk assessment techniques are applied to microbial hazards in meat, we cannot fully evaluate the usefulness of the methodology as a means of justifying standards and procedures for the safety of meat.
References