E. Franz

Quantitative Microbial Risk Assessment for Escherichia coli O157:H7, Salmonella, and Listeria monocytogenes in Leafy Green Vegetables Consumed at Salad Bars

Fresh vegetables are increasingly recognized as a source of foodborne outbreaks in many parts of the world. The purpose of this study was to conduct a quantitative microbial risk assessment for Escherichia coli O157:H7, Salmonella, and Listeria monocytogenes infection from consumption of leafy green vegetables in salad from salad bars in The Netherlands. Pathogen growth was modeled in Aladin (Agro Logistics Analysis and Design Instrument) using time-temperature profiles in the chilled supply chain and one particular restaurant with a salad bar. A second-order Monte Carlo risk assessment model was constructed (using @Risk) to estimate the public health effects. The temperature in the studied cold chain was well controlled below 5 degrees C. Growth of E. coli O157:H7 and Salmonella was minimal (17 and 15%, respectively). Growth of L. monocytogenes was considerably greater (194%). Based on first-order Monte Carlo simulations, the average number of cases per year in The Netherlands associated the consumption leafy greens in salads from salad bars was 166, 187, and 0.3 for E. coli O157:H7, Salmonella, and L. monocytogenes, respectively. The ranges of the average number of annual cases as estimated by second-order Monte Carlo simulation (with prevalence and number of visitors as uncertain variables) were 42 to 551 for E. coli O157:H7, 81 to 281 for Salmonella, and 0.1 to 0.9 for L. monocytogenes. This study included an integration of modeling pathogen growth in the supply chain of fresh leafy vegetables destined for restaurant salad bars using software designed to model and design logistics and modeling the public health effects using probabilistic risk assessment software.

A chain modeling approach to estimate the impact of soil cadmium pollution on human dietary exposure.

Cadmium in soil poses a risk for human health, due to its accumulation in food and feed crops. The extent of accumulation depends strongly on soil type and the degree of pollution. The objective of the present study was to develop a predictive model to estimate human dietary cadmium exposure from soil characteristics. This chain model consists of three basic steps: (i) calculation of plant cadmium levels from soil contamination levels and soil characteristics, (ii) calculation of animal transfer from consumption and contamination levels, and (iii) human exposure from both plant and animal products. Six soil scenarios were assessed, reflecting a specific contaminated region and ranging from 0.5 mg/kg of Cd (pH 4.5) to 2.5 mg/kg of Cd (pH 5.5). Cadmium levels in feed crops and vegetables were estimated with regression and mathematical models. Animal exposure and transfer to cattle kidneys, livers, and meat were calculated using a consumption database and a parameterized linear simulation model. Human exposure was estimated by Monte Carlo simulation, using a consumption database. The median human exposure for the different scenarios ranged from 0.24 to 0.98 microg/kg of body weight per day, which is comparable to results obtained from exposure levels based on observed field contamination data. The study shows that a chain model approach from soil contamination to human exposure, including animal exposure and transfer to animal products, can successfully be applied. The model can be used for fast evaluation of dietary cadmium exposure and the identification of risk areas based on soil conditions.

Quantitative microbial risk assessment for Escherichia coli O157:H7, Salmonella enterica, and Listeria monocytogenes in leafy green vegetables consumed at salad bars, based on modeling supply chain logistics.

Quantitative microbial risk assessments do not usually account for the planning and ordering mechanisms (logistics) of a food supply chain. These mechanisms and consumer demand determine the storage and delay times of products. The aim of this study was to quantitatively assess the difference between simulating supply chain logistics (MOD) and assuming fixed storage times (FIX) in microbial risk estimation for the supply chain of fresh-cut leafy green vegetables destined for working-canteen salad bars. The results of the FIX model were previously published (E. Franz, S. O. Tromp, H. Rijgersberg, and H. J. van der Fels-Klerx, J. Food Prot. 73:274-285, 2010). Pathogen growth was modeled using stochastic discrete-event simulation of the applied logistics concept. The public health effects were assessed by conducting an exposure assessment and risk characterization. The relative growths of Escherichia coli O157 (17%) and Salmonella enterica (15%) were identical in the MOD and FIX models. In contrast, the relative growth of Listeria monocytogenes was considerably higher in the MOD model (1,156%) than in the FIX model (194%). The probability of L. monocytogenes infection in The Netherlands was higher in the MOD model (5.18×10(-8)) than in the FIX model (1.23×10(-8)). The risk of listeriosis-induced fetal mortality in the perinatal population increased from 1.24×10(-4) (FIX) to 1.66×10(-4) (MOD). Modeling the probabilistic nature of supply chain logistics is of additional value for microbial risk assessments regarding psychrotrophic pathogens in food products for which time and temperature are the postharvest preventive measures in guaranteeing food safety.

Prediction of deoxynivalenol content in Dutch winter wheat.

Predictive models for the deoxynivalenol (DON) content in wheat can be a useful tool for control authorities and the industry to avoid or limit potential food and/or feed safety problems. The objective of this study was to develop a predictive model for DON in mature Dutch winter wheat. From 2001 to 2007, the concentration of DON was measured in winter wheat samples taken just before harvest from 264 fields throughout The Netherlands. Agronomic and climatic variables were obtained for each field for a 48-day period, centered on the heading date. Multiple regression was used to determine the most important variables and to construct the predictive model. The first model (model 1) was based on 24-day pre- and postheading periods, while the second model (model 2) was based on eight time blocks of 6 days around the heading date. Although both models showed good statistical evaluations and predictive performance, model 1 showed the highest performance (R(2) of 0.59 between observed and predicted values, fraction samples correctly below or above the 1,250 microg/kg threshold of 92%, and sensitivity of 63%). With both models, the predicted DON level increased with a higher average temperature, increased precipitation, and higher relative humidity, but decreased with increased number of hours with the temperature above 25 degrees C. We observed a strong regional effect on the levels of DON, which could not be explained by differences in the recorded agronomic and climatic variables. It is suggested that future model improvement might be realized by indentifying and quantifying the mechanism underlying the region effect.

Farm and slaughterhouse characteristics affecting the occurrence of Salmonella and Campylobacter in the broiler supply chain

Based on a data set on Campylobacter and Salmonella prevalence in the broiler supply chain, collected during the period 2002 through 2005 in the Netherlands, farm- and slaughterhouse-specific characteristics were tested for their effect on Campylobacter and Salmonella prevalence at different stages of the broiler supply chain. Three different sampling points were considered: departure from the farm, arrival at the slaughterhouse, and the end of the slaughterline. Strong associations were found between Salmonella and Campylobacter prevalence at a particular sampling point and their prevalence at the preceding point of the chain. Statistical analyses showed that the country of origin of the broiler farm had a significant effect on the prevalence of Salmonella and Campylobacter at slaughterhouse arrival. The feeding company delivering to the farm also showed a significant effect on the occurrence of both pathogens at departure from the broiler farm. The prevalence of Campylobacter decreased with an increasing number of birds per flock, whereas the prevalence of Salmonella increased with an increasing number of birds per flock. The number of flocks processed within a specific slaughterhouse was not associated with an increased or decreased prevalence of Campylobacter and Salmonella. The results provide more insight into factors related to the occurrence of both pathogens and in understanding their epidemiology. The results can be supportive in decision making on measures to reduce the contamination of broiler products with Salmonella and Campylobacter.

A Model for Setting Performance Objectives for Salmonella in the Broiler Supply Chain

A stochastic model for setting performance objectives for Salmonella in the broiler supply chain was developed. The goal of this study was to develop a model by which performance objectives for Salmonella prevalence at various points in the production chain can be determined, based on a preset final performance objective at the end of the processing line. The transmission of Salmonella through the broiler production chain was modeled. The prevalence at flock level was calculated from the measured prevalence at sample level. The transmission model is based on data on the occurrence of Salmonella collected in the Dutch broiler production chain during several years. The developed model can be used by policymakers and industry to determine economically and politically acceptable performance objectives for various points of the production chain and to draw conclusions about which interventions are most appropriate.

Modeling the impact of the indigenous microbial population on the maximum population density of Salmonella on alfalfa

Within a microbial risk assessment framework, modeling the maximum population density (MPD) of a pathogenic microorganism is important but often not considered. This paper describes a model predicting the MPD of Salmonella on alfalfa as a function of the initial contamination level, the total count of the indigenous microbial population, the maximum pathogen growth rate and the maximum population density of the indigenous microbial population. The model is parameterized by experimental data describing growth of Salmonella on sprouting alfalfa seeds at inoculum size, native microbial load and Pseudomonas fluorescens 2–79. The obtained model fits well to the experimental data, with standard errors less than ten percent of the fitted average values. The results show that the MPD of Salmonella is not only dictated by performance characteristics of Salmonella but depends on the characteristics of the indigenous microbial population like total number of cells and its growth rate. The model can improve the predictions of microbiological growth in quantitative microbial risk assessments. Using this model, the effects of preventive measures to reduce pathogenic load and a concurrent effect on the background population can be better evaluated. If competing microorganisms are more sensitive to a particular decontamination method, a pathogenic microorganism may grow faster and reach a higher level. More knowledge regarding the effect of the indigenous microbial population (size, diversity, composition) of food products on pathogen dynamics is needed in order to make adequate predictions of pathogen dynamics on various food products.