Previous Page Table of Contents Next Page



Salmonellosis is a leading cause of foodborne illness in many countries with eggs and poultry being important vehicles of transmission. During the past two decades Salmonella Enteritidis has become a leading serovar causing human infections, with hen eggs being a principal source of the pathogen. The emergence of Salmonella Enteritidis as the leading cause of human salmonellosis in many countries was attributed to this serovars unusual ability to colonise the ovarian tissue of hens and be present within the contents of intact shell eggs.

Broiler chicken is the main type of chicken consumed as poultry in many countries. A large percentage are colonised by salmonellae during grow-out and the skin and meat of carcasses are frequently contaminated by the pathogen during slaughter and processing.

Considering the major role eggs and poultry have as vehicles of transmission in human cases of salmonellosis, an assessment of different factors affecting the prevalence, growth, and transmission of Salmonella in eggs and on broiler chicken carcasses on the risk of human illness would be useful to risk managers in identifying the intervention strategies that would have the greatest impact on reducing human infections.

A risk characterization study attempting to quantify the human-health risk attributable to Salmonella Enteritidis in eggs and Salmonella spp. in broiler chickens was described in the working documents provided to the expert consultation. The risk characterizations involved amalgamating earlier work commissioned by FAO and WHO (Hazard Identification and Hazard Characterization of Salmonellae in Broilers and Eggs - FAO/WHO MRA 00/03, Exposure Assessment of Salmonella Enteritidis in Eggs - FAO/WHO MRA 00/04, and Exposure Assessment of Salmonella spp. in Broilers - FAO/WHO MRA 00/05) with new material developed since the preceding consultation. To enable the characterization of risk, both the Salmonella in eggs and Salmonella in broiler chickens expert drafting groups represented elements of the food production and consumption system as computer-based models. Outputs from these models were obtained using Monte Carlo[1] simulation to enable uncertainty and variability in input variables to be expressed as probability distributions, with the uncertainty and variability propagated through the model at each iteration and finally reflected in the predicted incidence of human disease.


The risk characterizations initially set out to understand how the incidence of human salmonellosis is influenced by various factors during the agricultural phase of chicken meat and egg production, marketing, processing, distribution, retail storage, consumer storage, meal preparation and finally consumption. Such models are appealing because they enable the study of the broadest range of intervention strategies. However, as the work progressed it became evident that the quantity and quality of information available from all sources was not sufficient to allow the construction of a full and expansive model. Thus, the final scope of the two Salmonella risk characterizations, and the components of the food production and consumption continuum that they each consider can now be described as follows:

A. Salmonella Enteritidis in eggs. This risk characterization estimates the probability of human illness due to Salmonella Enteritidis following the ingestion of a single food serving of internally contaminated shell eggs either consumed as whole eggs, egg meals, or as ingredients in more complex food (e.g. cake). This work addressed selected aspects of egg production on farms, further processing of eggs into egg products, retail and consumer egg handling and meal preparation practices.

B. Salmonella spp. in broiler chickens. This risk characterization estimates the probability of illness in a year due to the ingestion of Salmonella spp. on fresh whole broiler chicken carcasses with the skin intact and which are cooked in the domestic kitchen for immediate consumption. This work commenced at the conclusion of slaughterhouse processing and considers in-home handling and cooking practices. The effects of pre-slaughter interventions and the slaughter process are not presently included in this model.

In conducting the above risk assessments a common hazard characterization module was used. Within the hazard characterization step the objectives were to produce one or more curves describing the probability that an individual becomes ill due to Salmonella versus the dose of Salmonella ingested within food.


5.3.1 Hazard identification and hazard characterization

Information was compiled from published literature and from unpublished data submitted to FAO and WHO by various interested parties. Initially this was used to provide a description of the public health outcomes, pathogen characteristics, host characteristics, and food-related factors that may affect the survival of Salmonella through the stomach. The hazard characterization then presented a review of information on relevant dose-response models describing the mathematical relationship between an ingested dose of Salmonella and the probability of human illness. An extensive review of available outbreak data was also conducted. From these data a new dose-response model was derived using a re-sampling approach and this was used in both risk characterizations in preference to models defined in earlier work (Food Safety Inspection Service (FSIS) - United States Department of Agriculture (USDA) / Food and Drug Administration (FDA) Salmonella Enteritidis model, Health Canada Salmonella Enteritidis model, and the model produced by Fazil from human feeding trials). Finally, an attempt was made to discern whether separate dose-response curves could be justified for different sub-populations defined on the basis of age and ‘susceptibility’ and whether Salmonella Enteritidis had a dose-response distinguishable from other Salmonella spp..

5.3.2 Risk characterization for Salmonella Enteritidis in eggs

The model developed for this risk characterization combines the FSIS-USDA/FDA and Health Canada models. Generally, where input types were similar, the Health Canada parameters were used. If an input type was missing in one model, then the other model’s parameters were used (e.g., the Health Canada model did not consider cooling constants, so these were specified by the FSIS-USDA/FDA model). The model structure was generally based on the FSIS-USDA/FDA model.

The exposure assessment comprises a production module, a module for the processing and distribution of shell eggs, a module for the processing of egg products, and a module for preparation and consumption. The production module predicts the probability of a Salmonella Enteriditis-contaminated egg occurring. The shell egg processing and distribution, and preparation and consumption modules predict the probability of human exposures to various doses of Salmonella Enteriditis from contaminated eggs. As the diagram below shows (Figure 5.1), the output of the exposure assessment, in general, feeds into the hazard characterization to produce the risk characterization output. This output is the probability of human illness per serving of an egg-containing meal.

FIGURE 5.1 Schematic diagram of the Salmonella Enteriditis in eggs risk assessment

5.3.3 Risk characterization of Salmonella spp. in broiler chickens

The exposure assessment component of the risk characterization mimics the movement of Salmonella-contaminated chickens through the food chain commencing at the point of completion of the slaughter process. For each iteration of the model a chicken carcass was randomly allocated an infection status and those carcasses identified as contaminated were randomly assigned a number of Salmonella organisms. From this point until consumption, changes in the size of the Salmonella population on each contaminated chicken were modelled using equations for growth and death. The growth of Salmonella was predicted using random inputs for storage time at retail stores, transport time, storage time in homes, and the temperatures the carcass was exposed to during each of these periods. Death of Salmonella during cooking was predicted using random inputs describing the probability that a carcass was not adequately cooked, the proportion of Salmonella organisms attached to areas of the carcass that were protected from heat, the temperature of exposure of protected bacteria and the time for which such exposure occurs. The number of Salmonella consumed were then derived using a random input defining the weight of chicken meat consumed per serving and the numbers of Salmonella cells in meat as defined from the various growth and death processes. Finally, in the risk characterization the probability of illness was derived by combining the number of organisms ingested (from the exposure assessment) with information on the dose-response relationship (hazard characterization).


5.4.1 Hazard characterization for Salmonella

5.4.2 Risk characterization for Salmonella Enteritidis in eggs

FIGURE 5.2 Dose-response curve derived from outbreak data based on re-sampling from uncertainty distributions for outbreak variables.

Grey lines

Dose-response curves generated by fitting to uncertain data

Dose-response observed from outbreak data


Upper, lower, expected value, 2.5th percentile and 97.5th percentile representing the uncertainty ranges in the dose-response relationship.

TABLE 5.1 Predicted probabilities of illness per egg serving for different flock prevalence settings, and assuming the baseline egg storage time and temperature scenario. 5th and 95th percentiles of the uncertainty distribution about the risk of illness per egg serving are also shown.

Flock prevalence
























5.4.3 Risk characterization of Salmonella spp. in broiler chickens

In addition the report highlighted that:


The expert consultation strived to identify features of the work that have an impact on the acceptability of findings and the appropriateness of extrapolating findings to scenarios not explicitly investigated in the risk assessments. They are listed here together for the attention of readers to facilitate interpretation of the documents.

5.5.1 Hazard characterization for Salmonella spp.

The hazard characterization did not attempt to quantify secondary transmission (person-to-person), or the impact of chronic or severe outcomes resulting from humans being infected with Salmonella.

The available data did not make it possible to discern age-dependent dose-response relationships or separate dose-response relationships for those individuals that might be expected to be more susceptible to illness.

The impact of the food matrix was not incorporated into the hazard characterization due to limitations of available data.

Severity of illness as a function of patient age, Salmonella serovar or pathogen dose was not evaluated. Severity could potentially be influenced by any or all these factors. However, the current database of information was insufficient to derive a quantitative estimate of these factors.

Variations in the virulence of different Salmonella biotypes were not explicitly incorporated into the hazard characterization steps.

The data used to develop the dose-response model was based on outbreak data from a few developed countries. Application of the dose-response model in regions of the world where the population is believed to be distinct may not predict appropriately the probability of illness upon exposure to a dose of Salmonella.

The hazard characterization assumes that the beta-Poisson dose-response function, which has been shown in the literature to provide a good representation of the dose-response relationship for most bacterial pathogens and has low dose linearity properties as recommended in the draft WHO/FAO Guidelines on Hazard Characterization for Pathogens in Food and Water, is the appropriate mathematical form of the dose-response relationship.

There was some uncertainty associated with the outbreak data used to develop the dose-response relationship. It was necessary to introduce assumptions to use these data.

5.5.2 Risk characterization of Salmonella Enteritidis in eggs

This risk characterization for Salmonella Enteritidis in eggs was intentionally developed so as to not be representative of any specific country or region. Yet, within-flock prevalence and other model inputs were based on United States and/or Canadian evidence or assumptions. Therefore, caution is required when extrapolating from this model to countries other than the United States and Canada. At a minimum, data that is representative of a specific country should be used to determine model inputs prior to using the model to predict risk for that country.

A number of assumptions were made concerning the epidemiology of Salmonella Enteritidis. These include the assumptions that; infected hens produce contaminated eggs at a constant frequency regardless of hen strain, bacterial strain or environmental factors; the laying hen population is homogenous with respect to flock size, management factors and environmental factors; and within flock prevalence is random and not affected by factors such as age, the presence of an alternative host, bacterial strain and environment effects. Further investigation of the biology of Salmonella Enteritidis may enable these assumptions to be refined and to understand whether or not they introduce any limitations in the interpretation of findings.

The Salmonella Enteritidis risk characterization was in part based on estimates of the concentration of the organism in contaminated eggs. Evidence regarding enumeration of the organism was based on only 63 Salmonella Enteritidis-contaminated eggs. Of these eggs, only 32 contaminated eggs were from naturally infected hens (the remainder coming from experimentally infected hens). These data were used to describe the initial numbers of Salmonella Enteritidis in contaminated eggs. Also, these data were used to estimate the frequency of contaminated eggs that contain Salmonella Enteritidis organisms in the yolk at the time of lay. Nevertheless, it is difficult to represent uncertainty and variability with such limited data.

Much uncertainty attends the effectiveness of various management interventions for controlling Salmonella Enteritidis. The magnitudes of uncertainty regarding test sensitivity, effectiveness of cleaning and disinfecting, and vaccination efficacy have not been measured. Some data were available to describe these inputs, but the data may not be relevant to all regions or countries where such interventions might be applied.

Statistical or model uncertainty was not fully explored in this risk characterization. For example, alternative distributions to the lognormal for within-flock prevalence were not considered. Also, the predictive microbiology used in this model was dependent on very limited data pertaining to Salmonella Enteritidis growth inside eggs. Alternative functional specifications for Salmonella Enteritidis growth equations were not pursued in this analysis.

5.5.3 Risk characterization of Salmonella spp. in broiler chickens

The risk assessment of Salmonella spp. in broiler chickens did not consider all parts of the farm-to-fork continuum and this limits the range of control options that can be assessed. For example, if the model extended as far back as the pre-harvest stage, then risk management control options at the early stages of the chain could be appropriately evaluated. However, it was acknowledged by the expert consultation that national data exist in some countries that were not made available to the drafting group during the conduct of the work and undoubtedly more will still come to the attention of FAO and WHO upon distribution and discussion of this report. Thus, the opportunity exists to continue to expand upon the work presented in this document.

A ‘hypothetical’ baseline scenario was constructed to evaluate the relative merit of control strategies for Salmonella spp. in broiler chickens. This scenario consists of a composite of input assumptions derived from a range of countries. The findings from the baseline scenario should not be used to draw inferences about a particular nation or region. To make specific inferences it would be necessary to collect inputs that were directly relevant to the particular population of interest.

It was not possible to provide a perfect representation of growth of Salmonella spp. in raw poultry. For example, it was assumed that no growth occurs below 10oC (although in reality a small amount of growth can be expected), and growth was predicted using average temperatures which may be less accurate than predictions of growth based on time-temperature data. Seasonal variations in ambient temperature were not accounted for. The model adopted also assumed ambient temperature had no impact on the rate of change for storage temperatures used for predicting growth and this is intuitively inappropriate in some circumstances.

Similarly, limitations were present in the way the model predicts the death of Salmonella spp. in broiler chicken carcasses during the cooking process. For example, it was not possible to account for the fact that varying locations on the carcass experience a different amount of heat during cooking. Further, assumptions on the duration of cooking of broiler chickens in domestic kitchens were not based on extensive and reliable data, and time-temperature data were not used for these predictions.

At several steps reliance was placed on expert opinion to estimate the value of model inputs. The consultation acknowledged that expert opinion was often easily accessible and sometimes sufficiently accurate for this purpose (for instance information from an Irish retailer was used to estimate retail storage times). However, on some occasions expert opinion might reduce transparency and introduce an unacceptable bias that may not be detected by the risk assessors. This remark is of a general nature and does not only apply to the Salmonella risk assessment.

Surveillance data from some countries often show a marked seasonality in the number of notifications of human salmonellosis with peak incidence occurring in the warmer months. The current model cannot account for or explain this important phenomenon as there is no facility for including time-dependent climatic effects which arise because the biological basis of seasonal effects is presently too uncertain to enable it to be represented in a model.

The consultation agreed that a lack of detailed understanding of all aspects of cross-contamination in the home hampered the ability of the drafting group to address this process. Thus, the drafting group could only consider a limited number of pathways by which cross-contamination could occur, and these required assumptions having a questionable basis. Nevertheless, the consultation agreed that it was necessary for cross-contamination to be represented in the model because it is considered important in the epidemiology of foodborne salmonellosis in humans and because it provides a basis for future improvement of the models.

Although the uncertainty associated with several parameters in the consumption portion of the risk assessment was accounted for, a full analysis of statistical and model uncertainty was not performed. For instance, the influence of uncertainty in the cross contamination pathway was not explored.


5.6.1 Hazard characterization

5.6.2 Exposure assessment of Salmonella Enteriditis in eggs

5.6.3 Exposure assessment of Salmonella spp. in broiler chickens

If an attempt were made to extend the risk assessment to more fully assess pre-slaughter interventions then more data would be required on the prevalence of Salmonella in feed and replacement stock and fasting prior to slaughter. Data on the effect of scalding, de-feathering, evisceration, washing and chilling processes as well as other decontamination treatments are needed to effectively model the benefits of control interventions at the level of processing.


The 33rd session of the CCFH discussed the preliminary report of the Joint FAO/WHO Expert Consultation on Risk Assessment of Microbiological Hazards in Foods (17 - 21 July 2000), which addressed hazard characterization and exposure assessment of Salmonella spp in broiler chickens and eggs and L. monocytogenes in ready-to-eat foods. Through that discussion the Committee identified a number of risk management questions to be addressed by the joint FAO/WHO ad hoc expert consultations. These questions were listed in the report of the session (ALINORM 01/13A) and are outlined below.

Risk management questions for Salmonella Enteritidis in eggs

· Estimate the risk from Salmonella Enteritidis in eggs in the general population and in the various susceptible populations (e.g. elderly, children, immunocompromised patients) at various prevalence and concentration levels of Salmonella Enteritidis in contaminated eggs.

· Estimate the change in risk likely to occur from each of the interventions below including their efficacy.

Risk management questions for Salmonella spp. in chicken (broilers)

· Estimate the risk from pathogenic Salmonella spp. in chicken (broilers) consequential to a range of levels in raw poultry for the general population and for various susceptible population groups (elderly, children, and immunocompromised patients).

· Estimate the change in risk likely to occur for each of the interventions under consideration including their efficacy.

The response to these questions as outlined below is based on the risk assessment documents prepared by the expert drafting group and the discussions that took place during the expert consultation.

5.7.1 Estimation of the risk from pathogenic Salmonella spp. consequential to a range of levels for the general population and for various susceptible population groups. (elderly, children, immunocompromised patients).

The hazard characterization component of the risk assessment was used as the basis for addressing this question. While the risk from different levels of Salmonella was considered using reported outbreak data, there were insufficient data to limit the analysis to that from outbreaks associated with eggs and broiler chickens only.

A relationship between doses (numbers) of Salmonella ingested by people and the probability of developing symptoms of gastrointestinal illness was estimated using data from 34 outbreak investigations. For each outbreak the data described the magnitude of the ingested dose, the number of people consuming the contaminated food and the number who developed illness as a result.. The outbreak data came from Japan, United States, Canada and Europe and included water- and food-related outbreaks.

Comparing the outbreak attack rates for children under five years of age, against those for the rest of the population, there was no overall trend of increased risk of illness for this sub-population. The limited database of outbreak information, however, lacks the power to reveal the existence of any true differences that may exist. There was some indication in two outbreaks for the existence of a different attack rate for the two populations. However, considering the overall trend, and until more information is made available, the dose-response relationship for members of the population across all age groups was assumed to be the same.

There were no outbreak data available that would allow a more comprehensive assessment of other susceptibility factors.

The expert consultation agreed that this was an appropriate response.

5.7.2 The effect of specific management interventions on the risk posed by Salmonella Enteritidis in eggs.

A. Reducing the prevalence of infected flocks and destroying breeding or laying flocks

The model is capable of examining the effect of changing flock prevalence on the probability of illness per serving. For example, if flock prevalence was reduced by an amount, say x%, by the application of some intervention, then risk of human illness would also be reduced by x% (assuming all other model inputs remain constant).

Although the model did not explicitly include breeder flocks, it was expected that reducing infection in breeder flocks would reduce the prevalence of infected commercial flocks. If data were available that explain the contribution of breeder flocks to commercial flock prevalence, the change in risk per serving could be estimated using this model.

B. Reducing the prevalence of Salmonella Enteriditis positive eggs via testing of flocks and diversion of their eggs to pasteurization (including the effect of pasteurization)

Given a baseline scenario with flock prevalence of 25%, the effects of two test and diversion protocols were evaluated using the model. In one protocol, a single test was administered to all commercial flocks at the beginning of their egg production. In the other protocol three tests were administered to commercial flocks throughout the life of each flock. A single test was assumed to detect 44% of flocks. In the model, flocks that test positive for Salmonella Enteriditis were required to divert eggs to pasteurization, and to clean and disinfect the environment following depopulation (50% effective).

The probability of illness per shell egg serving for each year was calculated for each protocol (Figure 5.3). Testing three times per year reduced the risk of human illness from shell eggs by nearly 90% across four years.

While egg diversion from positive flocks reduces the public health risk from shell eggs, it might be expected that there is some increased risk from egg products. Mandatory diversion causes more contaminated eggs to be sent to pasteurization. Nevertheless, in this model the average number of Salmonella Enteriditis in pasteurized liquid whole eggs was reduced by diversion. This is because diverted eggs are stored for shorter periods than those not diverted and this reduces the opportunity for growth of the pathogen within those eggs that are contaminated.

FIGURE 5.3 Predicted probability of illness per serving from shell eggs per year following the implementation of two testing protocols. It was assumed that all flocks in the region were tested each time. Initial flock prevalence was assumed to be 25%. The baseline egg storage time and temperature scenario was used for the four years.

C. Vaccination of flocks against Salmonella Enteritidis and the use of competitive exclusion.

A single test for Salmonella Enteritidis, or two tests four months apart, of 90 faecal samples per test were used to evaluate the effect of vaccination against Salmonella Enteritidis. The vaccine was assumed to be capable of reducing the frequency of contaminated eggs by approximately 75%, based on data discussed in the source document (MRA 01/02).

Assuming 25% flock prevalence and the baseline egg storage time and temperature scenario, the probability of illness per serving for a single faecal test for Salmonella Enteritidis and vaccination protocol is about 70% of a non-vaccination protocol (Figure 5.4). Risk of illness was reduced to 60% of the non-vaccination protocol if two faecal tests for Salmonella Enteritidis were applied. Nevertheless, the actual reduction in risk of illness would depend on the prevalence of Salmonella Enteritidis within flocks and other input values.

Given data concerning competitive exclusion efficacy on infected flocks, an approach similar to that used for vaccination could be used in the model to estimate the effect on the risk of human illness.

D. Refrigeration of eggs after lay and during distribution or requirement of a specific storage time at retail for eggs at ambient temperatures.

The effect of mandatory retail storage times and temperatures was evaluated using baseline assumptions for a country that does not have egg refrigeration requirements.

Truncating retail storage time to <7 or <14 days simulated a shelf-life restriction scenario. Truncating the retail storage temperature to 7oC (<45oF) simulated a refrigeration requirement. The results are summarized in Figure 5.5. Shelf-life restrictions at retail within 14 days reduced risk of illness per serving very little. The refrigeration requirement reduced risk of illness per serving by about 58%. If shelf-life at retail was limited to 7 days, an effect comparable to that of refrigeration was noted. The actual reduction in risk will depend, however, on the prevalence of Salmonella Enteritidis within flocks and other input values.

FIGURE 5.4 Comparison of predicted probability of illness per serving between scenarios where no vaccination was used, where one faecal test for Salmonella Enteritidis was applied at the beginning of production and positive flocks were all vaccinated, and where a second faecal test for Salmonella Enteritidis was applied four months after the first test and additional test-positive flocks were vaccinated. The prevalence was assumed to be 25%, and the baseline egg storage time and temperature scenario was used.

FIGURE 5.5 Probability of illness per serving of shell eggs having a mandatory shelf-life of <7 or <14 days at retail, or a mandatory retail storage temperature of 7oC (<45oF).

5.7.3 The effect of specific management interventions on the risk posed by Salmonella spp. in broiler chickens.

A. On-farm and processing interventions

The approach taken for the risk assessment was to develop a model that uses prevalence and numbers of Salmonella spp. on broiler chickens at the end of processing as the starting point. A baseline model was established in order to provide a means to compare the effects on risk when prevalence and cell numbers were changed. In this way, a reduction in prevalence or cell numbers of Salmonella on raw poultry carcasses could be evaluated by the measure of change in the risk estimate compared to the baseline risk assessment. Farm or process interventions could be evaluated, provided the effect that the interventions have on the prevalence or cell numbers at the end of processing can be estimated.

Specific on-farm interventions, such as those mentioned in the question from CCFH, were not evaluated in the risk assessment model because of a lack of representative data to analyse how much change in either the prevalence and/or level of Salmonella in poultry could be attributable to any specific treatment or action. It is acknowledged that destruction of Salmonella positive flocks will influence public health outcomes. However, without information about the extent to which this practice translates to fewer infected birds at the completion of processing or fewer Salmonella cells per infected bird the magnitude of risk reduction could not be estimated

In addition, while there are many studies reporting the prevalence of Salmonella infection in birds, very few report the numbers of Salmonella that might be found on-farm and in live poultry. The studies providing enumeration data that were reviewed were substantially different with respect to sample sizes, experimental approach, testing methods, test accuracy, and spatial and temporal factors.

If management strategies that impact on the level of contamination, i.e., the numbers of Salmonella spp. on chickens are implemented, the relationship to risk is estimated to be greater than a one-to-one relationship. A shift of the cell number distribution on broiler chickens exiting the chill tank at the end of processing, such that the mean cell number from the baseline scenario is reduced by 40% on the non-log scale, reduces the expected risk of illness per serving by approximately 65%.

A change in the prevalence of contaminated raw product changes the estimated risk of illness to the consumer by altering the frequency of exposure to risk events i.e. exposure to the pathogen. The change in risk of illness as a result of a change in the prevalence of Salmonella-contaminated broiler chicken carcasses was estimated by simulating the model for a range of prevalence that are representative of data reported in various studies and countries (reviewed in Exposure Assessment MRA 00/05). If the prevalence of contaminated chickens leaving processing is altered, through some management practice either at the farm level or at the processing level, the expected risk of illness per serving is altered. The magnitude of the changes in risk of illness per serving and risk of illness per cross-contamination event as a result of changes in prevalence are summarized in Table 5.2.

TABLE 5.2 Change in risk of illnesses a result of decreasing or increasing the prevalence of Salmonella-contaminated broiler carcasses, relative to a baseline prevalence of 20% Salmonella-contamination used in the risk model.

Prevalence of contaminated carcasses exiting processing

Percent change in prevalence relative to assumed baseline

Per serving risk (estimated from model)

Percent change in per serving risk relative to assumed baseline

















20% (assumed baseline)












The change in risk of illness associated with a change in prevalence of Salmonella spp. was estimated to be a one-to-one relationship, assuming everything else remains constant. If the prevalence of Salmonella spp. on contaminated carcasses at the end of slaughter and processing is reduced by a specific percentage, the expected risk is reduced by a similar percentage. For example, reducing prevalence by 50% (e.g. from 20 to 10%) produces a 52% reduction in the expected risk of illness per serving (Table 5.3). Similarly, a larger reduction in prevalence (e.g., from 20% to 0.05%) was estimated to produce a 99.7% reduction in the expected risk of illness.

TABLE 5.3 Reported data from different studies on the effects of chlorine vs. no-chlorine in immersion water chilling on the prevalence of Salmonella on broiler chicken carcasses.

With Chlorine

Study No.

Prevalence before chilling

Prevalence after Chilling

Ratio After/









20-50ppm (tank)









4-9ppm (overflow)









1-5ppm (overflow)









15-50ppm (tank)










Without Chlorine

Prevalence before chilling

Prevalence after Chilling

Ratio (After/


























































There are very few data on the effect of air chilling vs. water chilling on the prevalence or numbers of Salmonella spp. on processed carcasses. Regarding the use of chlorine in water chilling of broilers, it is difficult to objectively evaluate because of differences in the literature on the level of chlorine used, and the location at which chlorine was measured. There is little information on the effect of chlorine addition at levels of 50 ppm or less on the numbers of the pathogen attached to the skin of poultry carcasses. On the other hand, studies reporting changes in prevalence seem to indicate that the effect of chlorine might be to simply prevent an increase in prevalence (i.e. reduce the extent of cross-contamination in the chill tank) as compared with before and after data where chlorine was not used (Table 5.3). However, the reported effectiveness of chlorine ranged from only a small increase in prevalence, to no change, to a decrease, and was not necessarily consistent with the concentration of chlorine.

B. The importance of various routes for introduction of pathogenic Salmonella spp. into flocks

In relation to the importance of various routes for introduction of pathogenic Salmonella spp. into flocks including feed, replacement birds, vectors, and hygiene, the available data were inconclusive. Interpretations of existing studies and results are confounded because of the number of different sampling protocols, specimen types, and laboratory methods, as well as the nature of poultry-rearing operations (e.g., very large vs. very small premises, types of waterers, feeders, etc.). For these reasons, it was not possible to evaluate the importance of on-farm routes of introduction of Salmonella spp., and this stage was not incorporated into the risk assessment.


The consultation expressed its appreciation for the quality and extent of information found in the risk characterization document. It was felt that the work represented a substantial advance in the application of scientific knowledge to improve the objective basis for managing Salmonella spp. in the food chain. The consultation further commended the extent of transparency achieved in the documents and for the pro-active manner in which limitations of the work were highlighted.

The consultation concluded that the new dose-response model derived from outbreak data represents the best available method for estimating probability of illness upon ingestion of a dose of Salmonella. However, with respect to the hazard characterization it was recommended that future analyses of data attempt to quantify the extent to which other factors effect the form of the dose-response relationship.

There is a need to increase the understanding of cross-contamination processes (on-farm, during transport, during processing, during storage and food preparation in the home and food service establishments) so that these can be modelled. The consultation recommended that additional data to improve these aspects of the model should be collected.

Special emphasis should also be placed on improvement of the survival and growth modules, for example modelling survival and growth of Salmonella spp. below 10oC. In relation to Salmonella spp. in broiler chickens the consultation recommended that when technically possible the model be extended to include the whole production chain, from farm-to-table. The consultation was informed that some data are available in this respect; occurrence of Salmonella in the environment and feed; flock-prevalence of Salmonella in replacement birds, layers and broilers; prevalence in broiler chickens after slaughter; and data showing a recent decrease of cases of Salmonella spp. in humans due to preharvest risk management actions (actions taken in broiler chicken production). In relation to data generation the expert consultation also noted that current enumeration methods vary in sensitivity. The most sensitive method (MPN) is labour intensive and expensive. Improvements are needed to overcome these difficulties and develop cost effective methods to enumerate small populations of Salmonella.

It was also recommended that a sensitivity analysis of the model be performed aimed at identifying the parameters that have the most impact on the predictions of probability of human illness.

The consultation recommended that the model, and in particular the data inputs to the model, are evaluated prior to their use by member countries. If possible, user-friendly versions of the models should be made available to member countries provided the new versions could be made to accurately reflect the behaviour of the models reviewed in this expert consultation.

[1] Monte Carlo: In Monte Carlo methods, the computer uses random number simulation techniques to mimic a statistical population. For each Monte Carlo replication, the computer: simulates a random sample from the population, analyzes the sample, and stores the result. After many replications, the stored results will mimic the sampling distribution of the statistic.
[2] A flock is a group of hens of similar age that are managed and housed together.

Previous Page Top of Page Next Page