Comparison of six dose-response models for use with food-borne pathogens

Two risk assessments: Evaluating the use of indicator HF183 Bacteroides versus pathogen measurements for modelling recreational illness risks in an urban watershed

The purpose of this study was to evaluate the performance of HF183 Bacteroides for estimating pathogen exposures during recreational water activities. We compared the use of Bacteroides-based exposure assessment to exposure assessment that relied on pathogen measurements. We considered two types of recreational water sites: those impacted by combined sewer overflows (CSOs) and those not impacted by CSOs. Samples from CSO-impacted and non-CSO-impacted urban creeks were analysed by quantitative polymerase chain reaction (qPCR) for HF183 Bacteroides and eight human gastrointestinal pathogens. Exposure assessment was conducted two ways for each type of site (CSO-impacted vs. non-CSO impacted): 1) by estimating pathogen concentrations from HF183 Bacteroides concentrations using published ratios of HF183 to pathogens in sewage and 2) by estimating pathogen concentrations from qPCR measurements. QMRA (quantitative microbial risk assessment) was then conducted for swimming, wading, and fishing exposures. Overall, mean risk estimates varied from 0.27 to 53 illnesses per 1,000 recreators depending on exposure assessment, site, activity, and norovirus dose-response model. HF183-based exposure assessment identified CSO-impacted sites as higher risk, and the recommended HF183 risk-based threshold of 525 genomic copies per 100 mL was generally protective of public health at the CSO-impacted sites but was not as protective at the non-CSO-impacted sites. In the context of our urban watershed, HF183-based exposure assessment over- and under-estimated risk relative to exposure assessment based on pathogen measurements, and the etiology of predicted pathogen-specific illnesses differed significantly. Across all sites, the HF183 model overestimated risk for norovirus, adenovirus, and Campylobacter jejuni, and it underestimated risk for E. coli and Cryptosporidium. To our knowledge, this study is the first to directly compare health risk estimates using HF183 and empirical pathogen measurements from the same waterways. Our work highlights the importance of site-specific hazard identification and exposure assessment to decide whether HF183 is applicable for monitoring risk.

Quantitative microbial risk assessment of Salmonella in fresh chicken patties

Quantitative microbial risk assessment (QMRA) has witnessed rapid development within the context of food safety in recent years. As a means of contributing to these advancements, a QMRA for Salmonella spp. in fresh chicken patties for the general European Union (EU) population was developed. A two-dimensional (Second Order) Monte-Carlo simulation method was used for separating variability and uncertainty of model's parameters. The stages of industrial processing, retail storage, domestic storage, and cooking in the domestic environment were considered in the exposure assessment. For hazard characterization, a dose-response model was developed by combining 8 published dose-response models using a Pert distribution for describing uncertainty. The QMRA model predicted a mean probability of illness of 1.19*10-4 (5.28*10-5 - 3.57*10-4 95 % C.I.), and a mean annual number of illnesses per 100,000 people of 2.13 (0.96 - 6.59 95 % C.I.). Moreover, sensitivity analysis was performed, and variability in cooking preferences was found to be the most influential model parameter (r = -0.39), followed by dose-response related variability (r = 0.22), and variability in the concentration of Salmonella spp. at the time of introduction at the processing facility (r = 0.11). Various mitigation strategy scenarios were tested, from which, "increasing the internal temperature of cooking" and "decreasing shelf life" were estimated to be the most effective in reducing the predicted risk of illness. Salmonella-related illnesses exhibit particularly high severity, making them some of the most prominent zoonotic diseases in the EU. Regular monitoring of this hazard in order to further highlight its related parameters and causes is a necessary procedure. This study not only provides an updated assessment of Salmonella spp. risk associated with chicken patties, but also facilitates the identification of crucial targets for scientific investigation and implementation of real-world intervention strategies.

Poultry Food Assess Risk Model for Salmonella and Chicken Gizzards: II. Illness Dose Step

The Illness Dose (ID) step of a Poultry Food Assess Risk Model (PFARM) for Salmonella and chicken gizzards (CGs) was shown in the present study. The illness dose is the minimum dose of Salmonella consumed that causes an illness. It depends on zoonotic potential (ZP) of Salmonella, food consumption behavior (FCB), and consumer health and immunity (CHI) or the disease triangle (DT). Zoonotic potential is the ability of Salmonella to survive, grow, and spread in the production chain or food and then cause illness in humans. Illness dose is predicted in PFARM using a DT, dose-response model (DRM) that was developed with human feeding trial (HFT) data and was validated with human outbreak investigation (HOI) data for Salmonella. The ability of the DT, DRM to predict DR data from HOI and HFT for Salmonella was quantified using the Acceptable Prediction Zones (APZ) method where acceptable performance occurred when the proportion of residuals in the APZ (pAPZ) was ≥ 0.7. United States, Centers for Disease Control and Prevention (CDC) data for human salmonellosis from 2007 to 2016 were used to simulate ZP and only minor changes in ZP of 11 Salmonella serotypes were observed during this time. The performance of the DT, DRM for predicting Salmonella DR data from HFT and HOI was acceptable with pAPZ that ranged from 0.87 to 1 for individual serotypes of Salmonella. Simulation results from the DT, DRM in PFARM indicated that ID decreased (P ≤ 0.05) and ZP increased (P ≤ 0.05) over time in the simulated production chain because the main serotype of Salmonella changed from Kentucky (low ZP) to Infantis (high ZP) while FCB and CHI were held constant. These results indicated that the DT, DRM in PFARM can be used with confidence to predict ID as a function of ZP, FCB, and CHI. In other words, the DT, DRM in PFARM can be used with confidence to predict dose-response for Salmonella and CGs.

[Application of the log-Logistic Model to Dose Response Relation in Microbial Risk Assessment]

Microbial risk assessment in food safety is a valuable tool to reduce the risks of infection by pathogens. The dose-response relation is aimed to establish the relationship between the dose of a pathogen that populations are exposed to and the probability of the adverse health effect by the pathogen. Among many dose-response models ever proposed, the exponential and beta-Poisson models have been internationally applied, but the decision on which model is selected between them solely depends on the goodness of fit to specific data sets. On the other hands, the log-logistic model, one of the alternative models, has been little studied on the dose-response relation. In the present study, thus, the application of the log-logistic model to dose-response relation was studied with hypothetical and experimental data sets of infection (or death), comparing to the above two models. Here the experimental data sets were for pathogenic organisms such as pathogenic Escherichia coli, Listeria monocytogenes, and Cryptosporidium pavrum. Consequently, this model successfully fit to those data sets in comparison to the two models. These results suggested that log-logistic model would have the potential to apply to the dose-response relation, similar to the exponential and beta-Poisson models.

An agent-based simulator for the gastrointestinal pathway of Listeria monocytogenes

We developed an agent-based gastric simulator for a human host to illustrate the within host survival mechanisms of Listeria monocytogenes. The simulator incorporates the gastric physiology and digestion processes that are critical for pathogen survival in the stomach. Mathematical formulations for the pH dynamics, stomach emptying time, and survival probability in the presence of gastric acid are integrated in the simulator to evaluate the portion of ingested bacteria that survives in the stomach and reaches the small intestine. The parameters are estimated using in vitro data relevant to the human stomach and L. monocytogenes. The simulator predicts that 5%-29% of ingested bacteria can survive a human stomach and reach the small intestine. In the absence of extensive scientific experiments, which are not feasible on the grounds of ethical and safety concerns, this simulator may provide a supplementary tool to evaluate pathogen survival and subsequent infection, especially with regards to the ingestion of small doses.

An Assessment of Listeriosis Risk Associated with a Contaminated Production Lot of Frozen Vegetables Consumed under Alternative Consumer Handling Scenarios

Frozen foods do not support the growth of Listeria monocytogenes (LM) and should be handled appropriately for safety. However, consumer trends regarding preparation of some frozen foods may contribute to the risk of foodborne listeriosis, specifically when cooking instructions are not followed and frozen products are instead added directly to smoothies or salads. A quantitative microbial risk assessment model FFLLoRA (Frozen Food Listeria Lot Risk Assessment) was developed to assess the lot-level listeriosis risk due to LM contamination in frozen vegetables consumed as a ready-to-eat food. The model was designed to estimate listeriosis risk per serving and the number of illnesses per production lot of frozen vegetables contaminated with LM, considering individual facility factors such as lot size, prevalence of LM contamination, and consumer handling prior to consumption. A production lot of 1 million packages with 10 servings each was assumed. When at least half of the servings were cooked prior to consumption, the median risk of invasive listeriosis per serving in both the general and susceptible population was <1.0 × 10^-16 with the median (5th, 95th percentiles) predicted number of illnesses per lot as 0 (0, 0) and 0 (0, 1) under the exponential and Weibull-gamma dose-response functions, respectively. In scenarios in which all servings are consumed as ready-to-eat, the median predicted risk per serving was 1.8 × 10^-13 and 7.8 × 10^-12 in the general and susceptible populations, respectively. The median (5th, 95th percentile) number of illnesses was 0 (0, 0) and 0 (0, 6) for the exponential and Weibull-Gamma models, respectively. Classification tree analysis highlighted initial concentration of LM in the lot, temperature at which the product is thawed, and whether a serving is cooked as main predictors for illness from a lot. Overall, the FFLLoRA provides frozen food manufacturers with a tool to assess LM contamination and consumer behavior when managing rare and/or minimal contamination events in frozen foods.

Inactivation kinetics and lethal dose analysis of antimicrobial blue light and photodynamic therapy

BACKGROUND

Antimicrobial Photodynamic therapy (A-PDT) has been used to treat infections. Currently, microbial inactivation data is reported presenting survival fraction averages and standard errors as discrete points instead of a continuous curve of inactivation kinetics. Standardization of this approach would allow clinical protocols to be introduced globally, instead of the piecemeal situation which currently applies.

METHODS

To this end, we used a power-law function to fit inactivation kinetics and directly report values of lethal doses (LD) and a tolerance factor (T) that informs if inactivation rate varies along the irradiation procedure. A deduced formula was also tested to predict LD for any given survival fraction value. We analyzed the photoantimicrobial effect caused by red light activation of methylene blue (MB-APDT) and by blue light (BL) activation of endogenous microbial pigments against 5 clinically relevant pathogens.

RESULTS

Following MB- APDT, Escherichia coli and Staphylococcus aureus cells become increasingly more tolerant to inactivation along the irradiation process (T < 1). Klebsiella pneumoniae presents opposite behavior, i.e., more inactivation is observed towards the end of the process (T > 1). P. aeruginosa and Candida albicans present constant inactivation rate (T˜1). In contrast, all bacterial species presented similar behavior during inactivation caused by BL, i.e., continuously becoming more sensitive to blue light exposure (T > 1).

CONCLUSION

The power-law function successfully fit all experimental data. Our proposed method precisely predicted LD and T values. We expect that these analytical models may contribute to more standardized methods for comparisons of photodynamic inactivation efficiencies.

Microalgae and cyanobacteria modeling in water resource recovery facilities: A critical review

Microalgal and cyanobacterial resource recovery systems could significantly advance nutrient recovery from wastewater by achieving effluent nitrogen (N) and phosphorus (P) levels below the current limit of technology. The successful implementation of phytoplankton, however, requires the formulation of process models that balance fidelity and simplicity to accurately simulate dynamic performance in response to environmental conditions. This work synthesizes the range of model structures that have been leveraged for algae and cyanobacteria modeling and core model features that are required to enable reliable process modeling in the context of water resource recovery facilities. Results from an extensive literature review of over 300 published phytoplankton models are presented, with particular attention to similarities with and differences from existing strategies to model chemotrophic wastewater treatment processes (e.g., via the Activated Sludge Models, ASMs). Building on published process models, the core requirements of a model structure for algal and cyanobacterial processes are presented, including detailed recommendations for the prediction of growth (under phototrophic, heterotrophic, and mixotrophic conditions), nutrient uptake, carbon uptake and storage, and respiration.

Advancing risk assessment: mechanistic dose-response modelling of Listeria monocytogenes infection in human populations

The utility of characterizing the effects of strain variation and individual/subgroup susceptibility on dose-response outcomes has motivated the search for new approaches beyond the popular use of the exponential dose-response model for listeriosis. While descriptive models can account for such variation, they have limited power to extrapolate beyond the details of particular outbreaks. By contrast, this study exhibits dose-response relationships from a mechanistic basis, quantifying key biological factors involved in pathogen-host dynamics. An efficient computational algorithm and geometric interpretation of the infection pathway are developed to connect dose-response relationships with the underlying bistable dynamics of the model. Relying on in vitro experiments as well as outbreak data, we estimate plausible parameters for the human context. Despite the presence of uncertainty in such parameters, sensitivity analysis reveals that the host response is most influenced by the pathogen-immune system interaction. In particular, we show how variation in this interaction across a subgroup of the population dictates the shape of dose-response curves. Finally, in terms of future experimentation, our model results provide guidelines and highlight vital aspects of the interplay between immune cells and particular strains of Listeria monocytogenes that should be examined.

Developing and utilizing controlled human models of infection

The controlled human infection model (CHIM) to assess the efficacy of vaccines against Shigella and enterotoxigenic Escherichia coli (ETEC) has several unique features that could significantly enhance the ability to test candidate vaccines. Despite increasing interest in these models, questions remain as to how to best incorporate them into vaccine development and how to maximize results. We designed a workshop focused on CHIM as part of the Vaccines Against Shigella and ETEC (VASE) Conference. The workshop, using the World Café method, focused on; clinical outcomes, nonclinical outcomes and model standardization. Researchers with a variety of expertise and experience rotated through each focus area and discussed relevant sub-topics. The results of these discussions were presented and questions posed to guide future workshops. Clinical endpoint discussions focused on the need for harmonized definitions; optimized attack rates; difficulties of sample collection and a need for non-stool based endpoints. Nonclinical discussions centered on evolving omics-based opportunities, host predictors of susceptibility and novel characterizations of the immune response. Model standardization focused on the value of shared procedures across institutions for clinical and non-clinical endpoints as well as for strain preparation and administration and subject selection. Participants agreed CHIMs for Shigella and ETEC vaccine development could accelerate vaccine development of a promising candidate; however, it was also appreciated that variability in the model and our limited understand of the host-pathogen interaction may yield results that could negatively impact a suitable candidate. Future workshops on CHIM are needed to ensure the optimal application of these models moving forward.

Dose-response relationships for environmentally mediated infectious disease transmission models

Environmentally mediated infectious disease transmission models provide a mechanistic approach to examining environmental interventions for outbreaks, such as water treatment or surface decontamination. The shift from the classical SIR framework to one incorporating the environment requires codifying the relationship between exposure to environmental pathogens and infection, i.e. the dose-response relationship. Much of the work characterizing the functional forms of dose-response relationships has used statistical fit to experimental data. However, there has been little research examining the consequences of the choice of functional form in the context of transmission dynamics. To this end, we identify four properties of dose-response functions that should be considered when selecting a functional form: low-dose linearity, scalability, concavity, and whether it is a single-hit model. We find that i) middle- and high-dose data do not constrain the low-dose response, and different dose-response forms that are equally plausible given the data can lead to significant differences in simulated outbreak dynamics; ii) the choice of how to aggregate continuous exposure into discrete doses can impact the modeled force of infection; iii) low-dose linear, concave functions allow the basic reproduction number to control global dynamics; and iv) identifiability analysis offers a way to manage multiple sources of uncertainty and leverage environmental monitoring to make inference about infectivity. By applying an environmentally mediated infectious disease model to the 1993 Milwaukee Cryptosporidium outbreak, we demonstrate that environmental monitoring allows for inference regarding the infectivity of the pathogen and thus improves our ability to identify outbreak characteristics such as pathogen strain.

Analysis of Vibrio vulnificus Infection Risk When Consuming Depurated Raw Oysters

A beta Poisson dose-response model for Vibrio vulnificus food poisoning cases leading to septicemia was used to evaluate the effect of depuration at 15 °C on the estimated health risk associated with raw oyster consumption. Statistical variability sources included V. vulnificus level at harvest, time and temperature during harvest and transportation to processing plants, decimal reductions (SV) observed during experimental circulation depuration treatments, refrigerated storage time before consumption, oyster size, and number of oysters per consumption event. Although reaching nondetectable V. vulnificus levels (<30 most probable number per gram) throughout the year and a 3.52 SV were estimated not possible at the 95% confidence level, depuration for 1, 2, 3, and 4 days would reduce the warm season (June through September) risk from 2,669 cases to 558, 93, 38, and 47 cases per 100 million consumption events, respectively. At the 95% confidence level, 47 and 16 h of depuration would reduce the warm and transition season (April through May and October through November) risk, respectively, to 100 cases per 100 million consumption events, which is assumed to be an acceptable risk; 1 case per 100 million events would be the risk when consuming untreated raw oysters in the cold season (December through March).

A dose and time response Markov model for the in-host dynamics of infection with intracellular bacteria following inhalation: with application to Francisella tularensis

In a novel approach, the standard birth–death process is extended to incorporate a fundamental mechanism undergone by intracellular bacteria, phagocytosis. The model accounts for stochastic interaction between bacteria and cells of the immune system and heterogeneity in susceptibility to infection of individual hosts within a population. Model output is the dose–response relation and the dose-dependent distribution of time until response, where response is the onset of symptoms. The model is thereafter parametrized with respect to the highly virulent Schu S4 strain of Francisella tularensis, in the first such study to consider a biologically plausible mathematical model for early human infection with this bacterium. Results indicate a median infectious dose of about 23 organisms, which is higher than previously thought, and an average incubation period of between 3 and 7 days depending on dose. The distribution of incubation periods is right-skewed up to about 100 organisms and symmetric for larger doses. Moreover, there are some interesting parallels to the hypotheses of some of the classical dose–response models, such as independent action (single-hit model) and individual effective dose (probit model). The findings of this study support experimental evidence and postulations from other investigations that response is, in fact, influenced by both in-host and between-host variability.

Comparison of multi-stage dose-response mixture models, with applications

This article concerns the analysis of a stochastic model that we propose for the population that generates a response (response measure) to the dose with the multi-stage model. The parameter uncertainty is dealt with via random dose and random size of the population at risk. The response measure is modeled by a random sum of mixed Bernoulli random variables with arbitrary distribution for the mixing parameters. Some extensions of the model are defined by functionals of the infection probability, fulfilling some convexity properties. We analyze the response by stochastic comparisons under different stochastic relations on the random dosages and the random sizes of the population at risk; or on the random infection rates. We provide stochastic exact bounds of the mixture model for the response, using inequalities and the positive quadrant dependence. Numerical bounds of the response by a dose having a scalar value or having an exponential or uniform distributions are obtained. Some conclusions are derived: the lower estimation of the response measure in the increasing convex order sense by replacing the dosages by their means; effects of the variation of the dose on the magnitude of the probability distribution of the response; effects of parameter correlation on the degree of variability of the response to any random dose; the low-dose region assessment; and also, the classical multi-stage model is compared versus the mixture model featuring independence and versus that with positive quadrant dependence.

New data, strategies, and insights for Listeria monocytogenes dose-response models: summary of an interagency workshop, 2011

Listeria monocytogenes is a leading cause of hospitalization, fetal loss, and death due to foodborne illnesses in the United States. A quantitative assessment of the relative risk of listeriosis associated with the consumption of 23 selected categories of ready-to-eat foods, published by the U.S. Department of Health and Human Services and the U.S. Department of Agriculture in 2003, has been instrumental in identifying the food products and practices that pose the greatest listeriosis risk and has guided the evaluation of potential intervention strategies. Dose-response models, which quantify the relationship between an exposure dose and the probability of adverse health outcomes, were essential components of the risk assessment. However, because of data gaps and limitations in the available data and modeling approaches, considerable uncertainty existed. Since publication of the risk assessment, new data have become available for modeling L. monocytogenes dose-response. At the same time, recent advances in the understanding of L. monocytogenes pathophysiology and strain diversity have warranted a critical reevaluation of the published dose-response models. To discuss strategies for modeling L. monocytogenes dose-response, the Interagency Risk Assessment Consortium (IRAC) and the Joint Institute for Food Safety and Applied Nutrition (JIFSAN) held a scientific workshop in 2011 (details available at http://foodrisk.org/irac/events/). The main findings of the workshop and the most current and relevant data identified during the workshop are summarized and presented in the context of L. monocytogenes dose-response. This article also discusses new insights on dose-response modeling for L. monocytogenes and research opportunities to meet future needs.

Quantitative risk assessment of human salmonellosis in Canadian broiler chicken breast from retail to consumption

The current quantitative risk assessment model followed the framework proposed by the Codex Alimentarius to provide an estimate of the risk of human salmonellosis due to consumption of chicken breasts which were bought from Canadian retail stores and prepared in Canadian domestic kitchens. The model simulated the level of Salmonella contamination on chicken breasts throughout the retail-to-table pathway. The model used Canadian input parameter values, where available, to represent risk of salmonellosis. From retail until consumption, changes in the concentration of Salmonella on each chicken breast were modeled using equations for growth and inactivation. The model predicted an average of 318 cases of salmonellosis per 100,000 consumers per year. Potential reasons for this overestimation were discussed. A sensitivity analysis showed that concentration of Salmonella on chicken breasts at retail and food hygienic practices in private kitchens such as cross-contamination due to not washing cutting boards (or utensils) and hands after handling raw meat along with inadequate cooking contributed most significantly to the risk of human salmonellosis. The outcome from this model emphasizes that responsibility for protection from Salmonella hazard on chicken breasts is a shared responsibility. Data needed for a comprehensive Canadian Salmonella risk assessment were identified for future research.

An expert panel report of a proposed scientific model demonstrating the effectiveness of antibacterial handwash products

In 2005, a US Food and Drug Administration Nonprescription Drug Advisory Committee (NDAC) review of consumer antiseptic handwash product studies concluded that the data regarding existing products failed to demonstrate any association between specific log reductions of bacteria achieved by antiseptic handwashing and reduction of infection. The NDAC recommended that consumer antibacterial handwashing products should demonstrate a reduction in infection compared with non-antibacterial handwash products. In response to the NDAC review, a consumer product industry-sponsored expert panel meeting was held in October 2007 to review new methods for assessing the efficacy of antibacterial handwashes. The expert panel reviewed a newly proposed model for linking the effectiveness of antibacterial handwashing to infection reduction and made recommendations for conducting future studies designed to demonstrate the efficacy of antibacterial handwash formulations. The panel concluded that using the surrogate infection model to demonstrate efficacy has a sound scientific basis, that the use of Shigella flexneri as a test organism coupled with a modified hand contamination procedure is supported by published data, and that the model represents a realistic test for the efficacy of consumer antibacterial handwash products. This article summarizes the expert panel's deliberations, conclusions, and recommendations.

Construction of food and water borne pathogens' dose-response curves using the expanded Fermi Solution

Theoretically, the relationship between the number of pathogens that cause acute infection if settling in the gut, N, and that initially ingested, M, can be constructed from the survival probabilities at the different "stations" along the digestive tract. These probabilities are rarely known exactly, but their ranges can be estimated. If for a given N one generates estimates of M using random probabilities within these ranges, the estimates' distribution will be approximately lognormal and its cumulative (CDF) form will represent the pathogen's dose-response curve. The distribution's logarithmic mean and standard deviation can be calculated from the ranges with a formula and used to plot the curve. The method was used to generate dose-response curves of hypothetical food and waterborne pathogens and calculate their infective dose (ID) at 5%, 50%, and 95% probability. The curves were compatible with the Beta Poisson model and robust against minor perturbations in the underlying probabilities' ranges. The calculation and plotting procedure was automated and posted on the Internet as a freely downloadable interactive Wolfram Demonstration. It allows the user to generate, modify, examine, and compare dose-response curves, and to calculate their characteristics, by moving sliders on the screen.

Dynamic predictive model for the growth of Salmonella spp. in liquid whole egg

A dynamic model for the growth of Salmonella spp. in liquid whole egg (LWE) (approximately pH 7.8) under continuously varying temperature was developed. The model was validated using 2 (5 to 15 °C; 600 h and 10 to 40 °C; 52 h) sinusoidal, continuously varying temperature profiles. LWE adjusted to pH 7.8 was inoculated with approximately 2.5-3.0 log CFU/mL of Salmonella spp., and the growth data at several isothermal conditions (5, 7, 10, 15, 20, 25, 30, 35, 37, 39, 41, 43, 45, and 47 °C) was collected. A primary model (Baranyi model) was fitted for each temperature growth data and corresponding maximum growth rates were estimated. Pseudo-R2 values were greater than 0.97 for primary models. Modified Ratkowsky model was used to fit the secondary model. The pseudo-R2 and root mean square error were 0.99 and 0.06 log CFU/mL, respectively, for the secondary model. A dynamic model for the prediction of Salmonella spp. growth under varying temperature conditions was developed using 4th-order Runge-Kutta method. The developed dynamic model was validated for 2 sinusoidal temperature profiles, 5 to 15 °C (for 600 h) and 10 to 40 °C (for 52 h) with corresponding root mean squared error values of 0.28 and 0.23 log CFU/mL, respectively, between predicted and observed Salmonella spp. populations. The developed dynamic model can be used to predict the growth of Salmonella spp. in LWE under varying temperature conditions.

PRACTICAL APPLICATION

  Liquid egg and egg products are widely used in food processing and in restaurant operations. These products can be contaminated with Salmonella spp. during breaking and other unit operations during processing. The raw, liquid egg products are stored under refrigeration prior to pasteurization. However, process deviations can occur such as refrigeration failure, leading to temperature fluctuations above the required temperatures as specified in the critical limits within hazard analysis and critical control point plans for the operations. The processors are required to evaluate the potential growth of Salmonella spp. in such products before the product can be used, or further processed. Dynamic predictive models are excellent tools for regulators as well as the processing plant personnel to evaluate the microbiological safety of the product under such conditions.

Risk of fetal mortality after exposure to Listeria monocytogenes based on dose-response data from pregnant guinea pigs and primates

One-third of the annual cases of listeriosis in the United States occur during pregnancy and can lead to miscarriage or stillbirth, premature delivery, or infection of the newborn. Previous risk assessments completed by the Food and Drug Administration/the Food Safety Inspection Service of the U.S. Department of Agriculture/the Centers for Disease Control and Prevention (FDA/USDA/CDC) and Food and Agricultural Organization/the World Health Organization (FAO/WHO) were based on dose-response data from mice. Recent animal studies using nonhuman primates and guinea pigs have both estimated LD(50)s of approximately 10(7) Listeria monocytogenes colony forming units (cfu). The FAO/WHO estimated a human LD(50) of 1.9 x 10(6) cfu based on data from a pregnant woman consuming contaminated soft cheese. We reevaluated risk based on dose-response curves from pregnant rhesus monkeys and guinea pigs. Using standard risk assessment methodology including hazard identification, exposure assessment, hazard characterization, and risk characterization, risk was calculated based on the new dose-response information. To compare models, we looked at mortality rate per serving at predicted doses ranging from 10(-4) to 10(12) L. monocytogenes cfu. Based on a serving of 10(6) L. monocytogenes cfu, the primate model predicts a death rate of 5.9 x 10(-1) compared to the FDA/USDA/CDC (fig. IV-12) predicted rate of 1.3 x 10(-7). Based on the guinea pig and primate models, the mortality rate calculated by the FDA/USDA/CDC is underestimated for this susceptible population.

Estimated occupational risk from bioaerosols generated during land application of class B biosolids

Some speculate that bioaerosols from land application of biosolids pose occupational risks, but few studies have assessed aerosolization of microorganisms from biosolids or estimated occupational risks of infection. This study investigated levels of microorganisms in air immediately downwind of land application operations and estimated occupational risks from aerosolized microorganisms. In all, more than 300 air samples were collected downwind of biosolids application sites at various locations within the United States. Coliform bacteria, coliphages, and heterotrophic plate count (HPC) bacteria were enumerated from air and biosolids at each site. Concentrations of coliforms relative to Salmonella and concentrations of coliphage relative to enteroviruses in biosolids were used, in conjunction with levels of coliforms and coliphages measured in air during this study, to estimate exposure to Salmonella and enteroviruses in air. The HPC bacteria were ubiquitous in air near land application sites whether or not biosolids were being applied, and concentrations were positively correlated to windspeed. Coliform bacteria were detected only when biosolids were being applied to land or loaded into land applicators. Coliphages were detected in few air samples, and only when biosolids were being loaded into land applicators. In general, environmental parameters had little impact on concentrations of microorganisms in air immediately downwind of land application. The method of land application was most correlated to aerosolization. From this large body of data, the occupational risk of infection from bioaerosols was estimated to be 0.78 to 2.1%/yr. Extraordinary exposure scenarios carried an estimated annual risk of infection of up to 34%, with viruses posing the greatest threat. Risks from aerosolized microorganisms at biosolids land application sites appear to be lower than those at wastewater treatment plants, based on previously reported literature.

Probability of porcine reproductive and respiratory syndrome (PRRS) virus infection as a function of exposure route and dose

At the most elemental level, the design of effective strategies to control and/or eliminate porcine reproductive and respiratory syndrome (PRRS) virus depend on an accurate and comprehensive understanding of virus transmission. As a general rule, transmission is highly dependent on the route of exposure and the dose of virus. The objective of this study was to derive PRRS virus isolate VR-2332 dose-response curves for oral and intranasal routes of exposure, i.e., determine the probability that a specific virus dose would result in infection. Individually housed pigs approximately 21 days of age were exposed to specific doses of PRRS virus isolate VR-2332 by either oral or intranasal routes. Positive controls were intramuscularly inoculated with 10(2.2) 50% tissue culture infective dose (TCID50) of PRRS virus and negative controls were orally administered 100ml of diluent with no virus. Pigs were monitored for evidence of infection for 21 days following exposure, i.e., serum samples were collected on days 0, 7, 14, 21, and tested for virus and PRRS virus-specific antibodies. Dose-response curves and 95% confidence intervals for oral and intranasal routes of exposure were derived using logistic models (logit and probit). The infectious dose50 (ID50) for oral exposure was estimated to be 10(5.3) TCID50 (95% CI, 10(4.6) and 10(5.9)); the ID50 for intranasal exposure was estimated to be 10(4.0) TCID50 (95% CI, 10(3.0) and 10(5.0)). Given these estimates, it is worth noting that intramuscular exposure of animals to 10(2.2) TCID50 (positive controls) resulted in infection in all animals. Thus pigs were the most susceptible to infection via parenteral exposure.

A quantitative risk assessment of the public health impact of the Finnish Salmonella control program for broilers

In order to study the public health effects of the Finnish Salmonella control program (FSCP), a quantitative risk assessment model of Salmonella from slaughtered broiler flocks to consumers was developed. Based on the model, approximately 0.21% of domestically produced broiler meat mass was contaminated with Salmonella (95% probability interval 0.05-0.48%). This model was combined to the model on primary production of broilers. By this way, the effect of eliminating breeder flocks from production which have tested positive for Salmonella and heat-treating the meat of detected positive broiler flocks on public health could be simulated. Based on the whole model, if detected positive breeder flocks were not removed this would result in 1.0-2.5 more reported human cases compared to the expected number of cases under current FSCP (95% predictive interval). Without heat treatment of meat the increase would be 2.9-5.4-fold and without both interventions 3.8-9.0-fold. In scenarios with one grandparent or five parent flocks infected, the combined effect of these two interventions was 9.3-25.8-fold and 4.9-11.7-fold compared to the baseline level under each scenario, respectively. The scenario analyses suggest that with a higher infection level, inclusion of both interventions will be more effective than either of the interventions alone. Replacement of half of the current retail broiler meat by meat with 20-40% contamination could result in 33-93 times more human cases compared to the expected value under current situations. On the basis of the model, the interventions applied in FSCP clearly protect the public health.

A comparison of microbial dose–response models fitted to human data

A study of eight mathematical dose-response models for microbial risk assessment was conducted using infectivity and illness data on a variety of microbial pathogens from published studies with human volunteers. The purpose was to evaluate variability among the models for human microbial dose-response data in order to determine whether two-parameter models might suffice for most microbial dose-response data or whether three-parameter models should generally be fitted. Model variability was measured in terms of estimated ED01s and ED10s, with the view that these effective dose levels correspond to the lower and upper limits of the 1-10% risk range generally recommended for establishing benchmark doses in risk assessment. An investigation of the ranks of the ED01 and ED10 values among the models led to the conclusion that the two-parameter models captured at least as much uncertainty as the three-parameter models for the data examined. A further evaluation of the two-parameter models did not result in the selection of one "best" model, but it did provide some insights into the models' relative behavior. The model uncertainty analysis proposed by Kang et al. [Regulat. Toxicol. Pharmacol. 32 (2000) 68] using four two-parameter models was reinforced.

A quantitative risk assessment of waterborne cryptosporidiosis in France using second-order Monte Carlo simulation

A pragmatic quantitative risk assessment (QRA) of the risks of waterborne Cryptosporidium parvum infection and cryptosporidiosis in immunocompetent and immunodeficient French populations is proposed. The model takes into account French specificities such as the French technique for oocyst enumeration performance and tap water consumption. The proportion of infective oocysts is based on literature review and expert knowledge. The probability of infection for a given number of ingested viable oocysts is modeled using the exponential dose-response model applied on published data from experimental infections in immunocompetent human volunteers challenged with the IOWA strain. Second-order Monte Carlo simulations are used to characterize the uncertainty and variability of the risk estimates. Daily risk of infection and illness for the immunocompetent and the immunodeficient populations are estimated according to the number of oocysts observed in a single storage reservoir water sample. As an example, the mean daily risk of infection in the immunocompetent population is estimated to be 1.08 x 10(-4) (95% confidence interval: [0.20 x 10(-4); 6.83 x 10(-4)]) when five oocysts are observed in a 100 L storage reservoir water sample. Annual risks of infection and disease are estimated from a set of oocyst enumeration results from distributed water samples, assuming a negative binomial distribution of day-to-day contamination variation. The model and various assumptions used in the model are fully explained and discussed. While caveats of this model are well recognized, this pragmatic QRA could represent a useful tool for the French Food Safety Agency (AFSSA) to define recommendations in case of water resource contamination by C. parvum whose infectivity is comparable to the IOWA strain.

Dose-response model for 13 strains of salmonella

Data from a human feeding trial with healthy men were used to develop a dose-response model for 13 strains of Salmonella and to determine the effects of strain variation on the shape of the dose-response curve. Dose-response data for individual strains were fit to a three-phase linear model to determine minimum, median, and maximum illness doses, which were used to define Pert distributions in a computer simulation model. Pert distributions for illness dose of individual strains were combined in an Excel spreadsheet using a discrete distribution to model strain prevalence. In addition, a discrete distribution was used to model dose groups and thus create a model that simulated human feeding trials. During simulation of the model with @Risk, an illness dose and a dose consumed were randomly assigned to each consumption event in the simulated feeding trial and if the illness dose was greater than the dose consumed then the model predicted no illness, otherwise the model predicted that an illness would occur. To verify the dose-response model predictions, the original feeding trial was simulated. The dose-response model predicted a median of 69 (range of 43-101) illnesses compared to 74 in the original trial. Thus, its predictions were in agreement with the data used to develop it. However, predictions of the model are only valid for eggnog, healthy men, and the strains and doses of Salmonella used to develop it. When multiple strains of Salmonella were simulated together, the predicted dose-response curves were irregular in shape. Thus, the sigmoid shape of dose-response curves in feeding trials with one strain of Salmonella may not accurately reflect dose response in naturally contaminated food where multiple strains may be present.

Modeling microbial growth within food safety risk assessments

Risk estimates for food-borne infection will usually depend heavily on numbers of microorganisms present on the food at the time of consumption. As these data are seldom available directly, attention has turned to predictive microbiology as a means of inferring exposure at consumption. Codex guidelines recommend that microbiological risk assessment should explicitly consider the dynamics of microbiological growth, survival, and death in foods. This article describes predictive models and resources for modeling microbial growth in foods, and their utility and limitations in food safety risk assessment. We also aim to identify tools, data, and knowledge sources, and to provide an understanding of the microbial ecology of foods so that users can recognize model limits, avoid modeling unrealistic scenarios, and thus be able to appreciate the levels of confidence they can have in the outputs of predictive microbiology models. The microbial ecology of foods is complex. Developing reliable risk assessments involving microbial growth in foods will require the skills of both microbial ecologists and mathematical modelers. Simplifying assumptions will need to be made, but because of the potential for apparently small errors in growth rate to translate into very large errors in the estimate of risk, the validity of those assumptions should be carefully assessed. Quantitative estimates of absolute microbial risk within narrow confidence intervals do not yet appear to be possible. Nevertheless, the expression of microbial ecology knowledge in "predictive microbiology" models does allow decision support using the tools of risk assessment.

Modelling the effects of temperature, water activity, pH and lactic acid concentration on the growth rate of Escherichia coli

An extended square root-type model describing Escherichia coli growth rate was developed as a function of temperature (7.63-47.43 degrees C), water activity (0.951-0.999, adjusted with NaCl), pH (4.02-8.28) and lactic acid concentration (0-500 mM). The new model, based on 236 growth rate data, combines and extends previously published square root-type models and incorporates terms for upper and lower limiting temperatures, upper and lower limiting pH, minimum inhibitory concentrations of dissociated and undissociated lactic acid and lower limiting water activity. A term to describe upper limiting water activity was developed but could not be fitted to the E. coli data set because of the difficulty of generating data in the super-optimal water activity range (i.e. >0.998). All data used to generate the model are presented. The model provides an excellent description of the experimental data.

Fermentation and pathogen control: a risk assessment approach

Food fermentation has a long tradition of improving the safety, shelf life and acceptability of foods. Although fermented foods generally enjoy a well-founded reputation for safety, some notable outbreaks of foodborne illness associated with fermented foods have occurred. Microbiological risk assessment (MRA), as it has emerged in recent years, provides the scientific basis for the control and management of risk. Aspects of fermented food processes are discussed under the various stages of risk assessment and data are presented that would inform more detailed risk assessments.

Sensitivity analysis of Salmonella enteritidis levels in contaminated shell eggs using a biphasic growth model

Salmonella enteritidis (SE) is a common foodbome pathogen, the transmission of which is primarily associated with the consumption of contaminated Grade A shell eggs. In order to estimate the level of SE present in raw shell eggs, it is necessary to consider the protective effects of the egg albumin, which effectively inhibits SE growth in a time- and temperature-dependent manner. In this study, a SE growth model was produced by combining two mathematical equations that described both the extended lag phase of SE growth (food component) and a SE growth model (pathogen component). This biphasic growth model was then applied to various egg handling scenarios based on the farm-to-table continuum, including in-line and off-line processing facilities with consideration of key events in production, processing, transportation, and storage. Seasonal effects were also studied. Monte Carlo simulation was used to characterize variability in temperature and time parameter values influencing the level of SE to which individuals are exposed. The total level of SE consumed was estimated under best, most likely, and time-temperature abusive handling scenarios. The model estimated that, in most cases, there was no SE growth in contaminated eggs handled under most likely practices, because 10-70% of the yolk membrane remained intact. Under abusive handling scenarios, complete loss of yolk membrane integrity frequently occurred by the time eggs reach the distribution phase, followed by subsequent SE growth, which was often quite rapid. In general, the effect of season and processing method (in-line vs. off-line) was minimal. Further sensitivity analysis demonstrated that the initial SE contamination level significantly influenced the final exposure levels only under no-abuse or mildly abusive conditions. The results of our study suggest that, for maximum reduction of SE exposure level, cooling strategies should not only focus on the on-farm or processing phases, but should emphasize the importance of cooling strategies at the distribution and consumer phases of the farm-to-fork continuum.