Control of human pathogenic Yersinia enterocolitica in minced meat: Comparative analysis of different interventions using a risk assessment approach.

This study aimed to evaluate the effect of different processing scenarios along the farm-to-fork chain on the contamination of minced pork with human pathogenic Y. enterocolitica. A modular process risk model (MPRM) was used to perform the assessment of the concentrations of pathogenic Y. enterocolitica in minced meat produced in industrial meat processing plants. The model described the production of minced pork starting from the contamination of pig carcasses with pathogenic Y. enterocolitica just before chilling. The endpoints of the assessment were (i) the proportion of 0.5 kg minced meat packages that contained pathogenic Y. enterocolitica and (ii) the proportion of 0.5 kg minced meat packages that contained more than 10³ pathogenic Y. enterocolitica at the end of storage, just before consumption of raw pork or preparation. Comparing alternative scenarios to the baseline model showed that the initial contamination and different decontamination procedures of carcasses have an important effect on the proportion of highly contaminated minced meat packages at the end of storage. The addition of pork cheeks and minimal quantities of tonsillar tissue into minced meat also had a large effect on the endpoint estimate. Finally, storage time and temperature at consumer level strongly influenced the number of highly contaminated packages.

Evaluation of a cross contamination model describing transfer of Salmonella spp. and Listeria monocytogenes during grinding of pork and beef.

In a previous study, a model was developed to describe the transfer and survival of Salmonella during grinding of pork (Møller, C.O.A., Nauta, M.J., Christensen, B.B., Dalgaard, P., Hansen, T.B., 2012. Modelling transfer of Salmonella typhimurium DT104 during simulation of grinding of pork. Journal of Applied Microbiology 112 (1), 90-98). The robustness of this model is now evaluated by studying its performance for predicting the transfer and survival of Salmonella spp. and Listeria monocytogenes during grinding of different types of meat (pork and beef), using two different grinders, different sizes and different numbers of pieces of meats to be ground. A total of 19 grinding trials were collected. Acceptable Simulation Zone (ASZ), visual inspection of the data, Quantitative Microbiological Risk Assessment (QMRA), as well as the Total Transfer Potential (TTP) were used as approaches to evaluate model performance and to access the quality of the cross contamination model predictions. Using the ASZ approach and considering that 70% of the observed counts have to be inside a defined acceptable zone of ±0.5 log10CFU per portion, it was found that the cross contamination parameters suggested by Møller et al. (2012) were not able to describe all 19 trials. However, for each of the collected grinding trials, the transfer event was well described when fitted to the model structure proposed by Møller et al. (2012). Parameter estimates obtained by fitting observed trials performed at different conditions, such as size and number of pieces of meat to be ground, may not be applied to describe cross contamination of unlike processing. Nevertheless, the risk estimates, as well as the TTP, revealed that the risk of disease may be reduced when the grinding of meat is performed in a grinder made of stainless steel (for all surfaces in contact with the meat), using a well-sharpened knife and holding at room temperatures lower than 4°C.

A QMRA Model for Salmonella in Pork Products During Preparation and Consumption.

As part of a quantitative microbiological risk assessment (QMRA) food chain model, this article describes a model for the consumer phase for Salmonella-contaminated pork products. Three pork products were chosen as a proxy for the entire pork product spectrum: pork cuts, minced meat patties, and fermented sausages. For pork cuts cross-contamination is considered the most important process and therefore it is modeled in detail. For minced meat, both cross-contamination and undercooking are the relevant processes. For those commodities bacterial growth during transport and storage is also modeled. Fermented sausages are eaten raw and the production may be defective. Variability between consumers' behavior and the impact of variability between production processes at the farm and abattoir are taken into account. Results indicate that Salmonella levels on products may increase significantly during transport and storage. Heating is very efficient at lowering concentrations, yet cross-contamination plays an important role in products that remain contaminated. For fermented sausage it is found that drying is important for Salmonella reduction. Sensitivity analysis revealed that cross- contamination factors "knife cleaning" and "preparation of a salad" are important parameters for pork cuts. For minced meat cleaning of the board, salad consumption, refrigerator temperature, and storage time were significant.

Impact of microbial count distributions on human health risk estimates.

Quantitative microbiological risk assessment (QMRA) is influenced by the choice of the probability distribution used to describe pathogen concentrations, as this may eventually have a large effect on the distribution of doses at exposure. When fitting a probability distribution to microbial enumeration data, several factors may have an impact on the accuracy of that fit. Analysis of the best statistical fits of different distributions alone does not provide a clear indication of the impact in terms of risk estimates. Thus, in this study we focus on the impact of fitting microbial distributions on risk estimates, at two different concentration scenarios and at a range of prevalence levels. By using five different parametric distributions, we investigate whether different characteristics of a good fit are crucial for an accurate risk estimate. Among the factors studied are the importance of accounting for the Poisson randomness in counts, the difference between treating "true" zeroes as such or as censored below a limit of quantification (LOQ) and the importance of making the correct assumption about the underlying distribution of concentrations. By running a simulation experiment with zero-inflated Poisson-lognormal distributed data and an existing QMRA model from retail to consumer level, it was possible to assess the difference between expected risk and the risk estimated with using a lognormal, a zero-inflated lognormal, a Poisson-gamma, a zero-inflated Poisson-gamma and a zero-inflated Poisson-lognormal distribution. We show that the impact of the choice of different probability distributions to describe concentrations at retail on risk estimates is dependent both on concentration and prevalence levels. We also show that the use of an LOQ should be done consciously, especially when zero-inflation is not used. In general, zero-inflation does not necessarily improve the absolute risk estimation, but performance of zero-inflated distributions in QMRA tends to be more robust to changes in prevalence and concentration levels, and to the use of an LOQ to interpret zero values, compared to that of their non-zero-inflated counterparts.

Fitting a distribution to microbial counts: making sense of zeroes.

The accurate estimation of true prevalence and concentration of microorganisms in foods is an important element of quantitative microbiological risk assessment (QMRA). This estimation is often based on microbial detection and enumeration data. Among such data are artificial zero counts, that originated by chance from contaminated food products. When these products are not differentiated from uncontaminated products that originate true zero counts, the estimates of true prevalence and concentration may be inaccurate. This inaccuracy is especially relevant in situations where highly pathogenic bacteria are involved and where growth can occur along the food pathway. Our aim was to develop a method that provides accurate estimates of concentration parameters and differentiates between artificial and true zeroes, thus also accurately estimating true prevalence. We first show the disadvantages of using a limit of quantification (LOQ) threshold for the analysis of microbial enumeration data. We show that, depending on the original distribution of concentrations and the LOQ value, it may be incorrect to treat artificial zeroes as censored below a quantification threshold. Next, a method is developed that estimates the true prevalence of contamination within a food lot and the parameters characterizing the within-lot distribution of concentrations, without assuming a LOQ, and using raw plate count data as an input. Counts resulting both from contaminated and uncontaminated sample units are analysed together. This procedure allows the estimation of the proportion of artificial zeroes among the total of zero counts, and therefore the estimation of true prevalence from enumeration results. We observe that this method yields best estimates of mean, standard deviation and prevalence at low true prevalence levels and low expected standard deviation. Furthermore, we conclude that the estimation of prevalence and the estimation of the distribution of concentrations are interrelated and therefore should be estimated simultaneously. We also conclude that one of the keys to an accurate characterization of the overall microbial contamination is the correct identification and separation of true and artificial zeroes. Our method for the analysis of quantitative microbial data shows a good performance in the estimation of true prevalence and the parameters of the distribution of concentrations, which indicates that it is a useful data analysis tool in the field of QMRA.

Estimating the true incidence of campylobacteriosis and salmonellosis in the European Union, 2009.

We estimated the true incidence of campylobacteriosis and salmonellosis in the European Union (EU) in 2009. The estimate was based on disease risks of returning Swedish travellers, averaged over the years 2005-2009, and anchored to a Dutch population-based study on incidence and aetiology of gastroenteritis. For the 27 EU member states the incidence of campylobacteriosis was about 9·2 (95% CI 2·8-23) million cases, while the incidence of salmonellosis was 6·2 (95% CI 1·0-19) million cases. Only 1/47 (95% CI 14-117) cases of campylobacteriosis and one 1/58 (95% CI 9-172) cases of salmonellosis were reported in the EU. The incidence rate of campylobacteriosis in EU member states varied between 30 and 13 500/100 000 population and was significantly correlated with the prevalence of Campylobacter spp. in broiler chickens. The incidence rate of salmonellosis in EU member states varied between 16 and 11 800/100 000 population and was significantly correlated with the prevalence of Salmonella Enteritidis in laying hens.

Risk analysis of analytical validations by probabilistic modification of FMEA.

Risk analysis is a valuable addition to validation of an analytical chemistry process, enabling not only detecting technical risks, but also risks related to human failures. Failure Mode and Effect Analysis (FMEA) can be applied, using a categorical risk scoring of the occurrence, detection and severity of failure modes, and calculating the Risk Priority Number (RPN) to select failure modes for correction. We propose a probabilistic modification of FMEA, replacing the categorical scoring of occurrence and detection by their estimated relative frequency and maintaining the categorical scoring of severity. In an example, the results of traditional FMEA of a Near Infrared (NIR) analytical procedure used for the screening of suspected counterfeited tablets are re-interpretated by this probabilistic modification of FMEA. Using this probabilistic modification of FMEA, the frequency of occurrence of undetected failure mode(s) can be estimated quantitatively, for each individual failure mode, for a set of failure modes, and the full analytical procedure.

Modelling transfer of Salmonella Typhimurium DT104 during simulation of grinding of pork.

AIMS: The aim of this study was to develop a model to predict cross-contamination of Salmonella during grinding of pork. METHODS AND RESULTS: Transfer rates of Salmonella were measured in three experiments, where between 10 and 20 kg meat was ground into 200-g portions. In each experiment, five pork slices of about 200 g per slice were inoculated with 8-9 log-units of Salmonella Typhimurium DT104 and used for building up the contamination in the grinder. Subsequently, Salmonella-free slices were ground and collected as samples of c. 200 g minced pork. Throughout the process, representative samples were quantitatively analysed for Salmonella. A model suggested by Nauta et al. (2005) predicting cross-contamination of Campylobacter in poultry processing and two modified versions of this model were tested. CONCLUSIONS: The present study observed a tailing phenomenon of transfer of Salmonella during a small-scale grinding process. It was, therefore, hypothesized that transfer occurred from two environmental matrices inside the grinder and a model was developed. The developed model satisfactorily predicted the observed concentrations of Salmonella during its cross-contamination in the grinding of up to 110 pork slices. SIGNIFICANCE AND IMPACT OF THE STUDY: The proposed model provides an important tool to examine the effect of cross-contamination in quantitative microbial risk assessments and might also be applied to various other food processes where cross-contamination is involved.

Consistency of FMEA used in the validation of analytical procedures.

In order to explore the consistency of the outcome of a Failure Mode and Effects Analysis (FMEA) in the validation of analytical procedures, an FMEA was carried out by two different teams. The two teams applied two separate FMEAs to a High Performance Liquid Chromatography-Diode Array Detection-Mass Spectrometry (HPLC-DAD-MS) analytical procedure used in the quality control of medicines. Each team was free to define their own ranking scales for the probability of severity (S), occurrence (O), and detection (D) of failure modes. We calculated Risk Priority Numbers (RPNs) and we identified the failure modes above the 90th percentile of RPN values as failure modes needing urgent corrective action; failure modes falling between the 75th and 90th percentile of RPN values were identified as failure modes needing necessary corrective action, respectively. Team 1 and Team 2 identified five and six failure modes needing urgent corrective action respectively, with two being commonly identified. Of the failure modes needing necessary corrective actions, about a third were commonly identified by both teams. These results show inconsistency in the outcome of the FMEA. To improve consistency, we recommend that FMEA is always carried out under the supervision of an experienced FMEA-facilitator and that the FMEA team has at least two members with competence in the analytical method to be validated. However, the FMEAs of both teams contained valuable information that was not identified by the other team, indicating that this inconsistency is not always a drawback.

Evaluation of the "testing and scheduling" strategy for control of Campylobacter in broiler meat in The Netherlands.

"Testing and scheduling" has been proposed as a strategy for control of Campylobacter in broiler meat. By this strategy, flocks with high numbers of Campylobacter in fecal samples would be diverted away from fresh meat production at the entrance of the broiler meat processing plant. Risk assessment studies suggest that this would effectively decrease human health risks, if these flocks are responsible for the meat products with the highest Campylobacter numbers. To investigate the effect of this control strategy, the numbers of Campylobacter were determined in fecal samples from transport containers, and in cecal and breast meat samples from birds in 62 broiler chicken flocks. Results from direct plating and enrichment were combined by a statistical method that allows the inclusion of censored data. As the implementation of "testing and scheduling" requires a rapid on-site test to detect high numbers of Campylobacter, a lateral flow immuno-assay (LFA) was developed and applied to the fecal samples collected from containers. The Campylobacter prevalence in broiler flocks in the autumn of 2007 was found to be 85.4% by traditional microbiological methods. Campylobacter could be isolated from breast meat samples from 42% of the flocks. There was limited agreement between Campylobacter results for the three types of samples and weak correlation between the quantitative results for fecal or cecal samples and meat samples. Agreement between the results of LFA and traditional methods was poor. These findings do not support the implementation of "testing and scheduling" as a practical control strategy, because of both measurement uncertainties and shortcomings in understanding the dynamics of transmission and survival of Campylobacter in the broiler meat processing plant. The limited correlation between Campylobacter contamination of cecal samples and breast meat samples, as observed in this study, suggests that cecal samples are no good indicator for human exposure to Campylobacter.

Risk analysis by FMEA as an element of analytical validation.

We subjected a Near-Infrared (NIR) analytical procedure used for screening drugs on authenticity to a Failure Mode and Effects Analysis (FMEA), including technical risks as well as risks related to human failure. An FMEA team broke down the NIR analytical method into process steps and identified possible failure modes for each step. Each failure mode was ranked on estimated frequency of occurrence (O), probability that the failure would remain undetected later in the process (D) and severity (S), each on a scale of 1-10. Human errors turned out to be the most common cause of failure modes. Failure risks were calculated by Risk Priority Numbers (RPNs)=O x D x S. Failure modes with the highest RPN scores were subjected to corrective actions and the FMEA was repeated, showing reductions in RPN scores and resulting in improvement indices up to 5.0. We recommend risk analysis as an addition to the usual analytical validation, as the FMEA enabled us to detect previously unidentified risks.

Cross-contamination in the kitchen: estimation of transfer rates for cutting boards, hands and knives.

AIMS: To quantify cross-contamination in the home from chicken to ready-to-eat salad. METHODS AND RESULTS: Based on laboratory scenarios performed by de Jong et al. (2008), transfer rates were estimated for Campylobacter jejuni and Lactobacillus casei as a tracer organism. This study showed that transfer characteristics for both micro-organisms were comparable when washing regimes and transfer via items (cutting board, hands and knives) were compared. Furthermore, the study showed that the use of separate transfer rates for transfer from chicken to items and from items to salad will lead to an overestimation of campylobacteriosis risk. Applying good hygienic practices resulted in final levels of bacteria in the salad below the detection limit. Our study showed that it is important to include these data points in model fitting. CONCLUSIONS: Results obtained in observational studies with Lact. casei can be translated to Camp. jejuni using the transfer rates obtained in this study. Cross-contamination by hands, cutting boards and knives was equally important. SIGNIFICANCE AND IMPACT OF THE STUDY: Cross-contamination should be incorporated in microbiological risk assessments. The present study contributes to this by quantifying transfer of Camp. jejuni and Lact. casei from raw chicken via various contact surfaces into the ready-to-eat product.

Cross-contamination in the kitchen: effect of hygiene measures.

AIMS: To determine the effect of hygiene measures on cross-contamination of Campylobacter jejuni at home and to select a safe tracer organism for C. jejuni. METHODS AND RESULTS: Comparative tests were conducted with nonpathogenic Escherichia coli and Lactobacillus casei and L. casei was chosen as the safe tracer organism. Salads containing chicken breast fillet contaminated with a known number of C. jejuni and L. casei were prepared according to different cross-contamination scenarios and contamination levels of salads were determined. Cross-contamination could be strongly reduced when cleaning cutting board and cutlery with hot water (68 degrees C), but generally was not prevented using consumer-style cleaning methods for hands and cutting board. CONCLUSIONS: Dish-washing does not sufficiently prevent cross-contamination, thus different cutting boards for raw meat and other ingredients should be used and meat-hand contact should be avoided or hands should be thoroughly cleaned with soap. Lactobacillus casei can be used as a safe tracer organism for C. jejuni in consumer observational studies. SIGNIFICANCE AND IMPACT OF THE STUDY: Cross-contamination plays an important role in the transmission of food-borne illness, especially for C. jejuni. This study delivers suitable data to quantitatively assess the risk of campylobacteriosis caused by cross-contamination and it shows the effect of different preventive hygiene measures.

Modelling of Campylobacter survival in frozen chicken meat.

AIMS: To model the survival kinetics of Campylobacter jejuni on frozen chicken meat. METHODS AND RESULTS: Three different types of chicken meat surface (skin, skinned muscle and cut muscle) were inoculated with stationary phase cells of C. jejuni (8 log(10) CFU cm(-2)) and frozen for 5 weeks at -20 degrees C. Bacterial numbers were determined weekly using two different methods of enumeration to quantify uninjured and injured cells. Analysis of variance of the results showed that the type of chicken surface and the method used to enumerate surviving cells were the most significant sources of variations in the numbers recovered (P < 0.0001), much more than the freezing time. To identify an appropriate model for the description of effects of freezing on survival over time, several models were fitted to the count data. Decay was found to be nonlinear. In general, survival was least on skin, better on skinned muscle and best on cut muscle. After 2 weeks, additional inactivation by freezing appeared to be negligible. CONCLUSION: Because of the variability of survival it was not possible to fit and select a general model useful for all the different surfaces types. SIGNIFICANCE AND IMPACT OF THE STUDY: The injured state of the cells leads to variability and the underestimation of bacterial survival. This is an essential factor for the assessment of Campylobacter-associated risk.

Disease burden in The Netherlands due to infections with Shiga toxin-producing Escherichia coli O157.

Surveys carried out between 1990 and 2000 indicated that the incidence of STEC O157-associated gastroenteritis in The Netherlands was 1250 cases/year (median), of which 180 visited a general practitioner, 40 are reported and 0.6 are fatal, mainly in the elderly. There are approximately 20 cases of STEC O157-associated haemolytic-uraemic syndrome (HUS) per year, mainly in children. There are 2.5 HUS patients per year who develop end-stage renal disease (ESRD). There are an estimated 2 HUS-related and 0.5 ESRD-related fatalities per year. The mean disease burden associated with STEC O157 in the Dutch population is 116 (90% confidence interval 85-160) Disability Adjusted Life Years (DALYs) per year. Mortality due to HUS (58 DALYs), and ESRD (21 DALYs) and dialysis due to ESRD (21 DALYs) constitute the main determinants of disease burden. Sensitivity analysis indicates that uncertainty associated with model assumptions did not have a major effect on these estimates.

Estimation of animal-level prevalence from pooled samples in animal production.

Often, the prevalence of an infection in the animal-production sector is determined at the group level. The prevalence at animal level (p) gives more-precise information on the infection status of the sector. This paper shows that pooled-sample data together with mathematical models allow for estimation of p. For this, model assumptions have to be made on the variation of p between groups separated in space and/or time. Formulas were derived for four models that were based on different assumptions. Model 1 assumed that p has the same value for all groups. Models 2-4 assumed that some of the groups were not infected. In addition, model 2 assumed that p has the same value for all infected groups; model 3 assumed that for an infected group, p is equal to either p(1) or p(2); and model 4 assumed that p was Beta distributed among infected groups. The models were applied to data sets on Salmonella infection in broiler flocks, including serotype data dominated by S. Hadar and S. Paratyphi B, var. Java. Based on likelihood-ratio tests, models 3 and 4 consistently fitted significantly better to the data. The applicability of model 4 is numerically bounded, related to the shape of the Beta distribution of p. Model calculations show that flock-level prevalence of Salmonella is much higher after than before slaughter. This difference (which possibly is related to different types of samples) is much smaller at the animal level. An important result of the estimation of p is that it in turn allows for an estimation of the proportion of false-negative groups--which is important in estimating the effect of veterinary or public-health measures.

A model for evaluating intervention strategies to control salmonella in the poultry meat production chain.

A model of the transmission of salmonella through the poultry meat production chain is developed, to predict the effects of intervention strategies for salmonella control. The model first describes the situation before intervention in terms of salmonella prevalences at flock level and some transmission parameters. After single control measures are translated into effects on these transmission parameters, the effects of sets of control measures (intervention strategies), can be calculated with the model. As research data are lacking, the model input parameters were derived from expert opinion. As an example, the effects of two intervention strategies proposed for the Dutch poultry industry are predicted. A sensitivity analysis is performed to indicate where the most effective control measures may be expected. Additionally, the reliability of the model predictions is studied by an uncertainty analysis. The use of the model as a tool for policy makers deciding about salmonella control strategies is discussed.

Human exposure to Mycobacterium paratuberculosis via pasteurised milk: a modelling approach.

Paratuberculosis is a disease of cattle caused by infection with Mycobacterium paratuberculosis, and it has been suggested that this bacterium may also play a role in the aetiology of Crohn's disease in humans. M paratuberculosis is shed in the milk and may be able to survive pasteurisation. Therefore, people may be exposed to it by the consumption of pasteurised milk. The risk of such exposure has been analysed using a modelling approach and the model has been used to evaluate the effects of intervention measures at different points in the potential route of transmission. On the basis of data from the literature and expert opinion, an initial point estimate of the exposure level of about 0-5 cfu/litre pasteurised milk was derived, mainly due to milk from clinically affected animals. The model indicates the need for quantitative data on variations in the shedding rates of M paratuberculosis in faeces and milk, and the levels of faecal contamination of milk. Such data are essential for a proper analysis of potential exposure, and may result in a 100-fold increase in the estimated median level of exposure.

Constraints on allele size at microsatellite loci: implications for genetic differentiation.

Microsatellites are promising genetic markers for studying the demographic structure and phylogenetic history of populations. We present theoretical arguments indicating that the usefulness of microsatellite data for these purposes may be limited to a short time perspective and to relatively small populations. The evolution of selectively neutral markers is governed by the interaction of mutation and random genetic drift. Mutation pressure has the inherent tendency to shift different populations to the same distribution of alleles. Hence, mutation pressure is a homogenizing force, and population divergence is caused by random genetic drift. In case of allozymes or sequence data, the diversifying effect of drift is typically orders of magnitude larger than the homogenizing effect of mutation pressure. By a simple model, we demonstrate that the situation may be different for microsatellites where mutation rates are high and the range of alleles is limited. With the help of computer simulations, we investigate to what extent genetic distance measures applied to microsatellite data can nevertheless yield useful estimators for phylogenetic relationships or demographic parameters. We show that predictions based on microsatellite data are quite reliable in small populations, but that already in moderately sized populations the danger of misinterpretation is substantial.

Evolutionary dynamics of spore killers.

Spore killing in ascomycetes is a special form of segregation distortion. When a strain with the Killer genotype is crossed to a Sensitive type, spore killing is expressed by asci with only half the number of ascospores as usual, all surviving ascospores being of the Killer type. Using population genetic modeling, this paper explores conditions for invasion of Spore killers and for polymorphism of Killers, Sensitives and Resistants (which neither kill, nor get killed), as found in natural populations. The models show that a population with only Killers and Sensitives can never be stable. The invasion of Killers and stable polymorphism only occur if Killers have some additional advantage during the process of spore killing. This may be due to the effects of local sib competition or some kind of "heterozygous" advantage in the stage of ascospore formation or in the short diploid stage of the life cycle. This form of segregation distortion appears to be essentially different from other, well-investigated forms, and more field data are needed for a better understanding of spore killing.