Relation between broiler and human Campylobacter jejuni strains isolated in Belgium from 2011 to 2013.

AIMS: This study inquires the relationship between Campylobacter jejuni isolated from broiler meat carcasses (n = 97) and human clinical samples (n = 72) in Belgium, from 2011 to 2013. METHODS AND RESULTS: The evaluation of the relation was based on the characteristics determined using multilocus sequence typing (MLST) alone and combined with flagellin gene A restriction fragment length polymorphism (flaA-RFLP) typing, antibiotic microbiological resistance profiling (AMRp), lipooligosaccharide class typing or virulence gene profiling (Vp). Clusters containing both human and broiler meat strains were more common when MLST was used alone, followed by MLST/flaA-RFLP and then by MLST/AMRp. More logical chronologically relations broiler-human were obtained for MLST/flaA-RFLP, then for MLST, and finally for MLST/AMRp: i.e. the isolates would first be detected in the broiler meat and at the same time or later in humans. CONCLUSIONS: In several cases, the C. jejuni strains isolated from the consumed broiler meat and from the campylobacteriosis case had the same profile, according to the used typing methods. The circulating Campylobacter strains appear to have remained the same from 2011 till 2013 in Belgium. SIGNIFICANCE AND IMPACT OF THE STUDY: This study corroborates previously published data from Belgium that suggest a strong correlation between C. jejuni strains isolated from broiler meat and from campylobacteriosis patients.

Reduced contamination of pig carcasses using an alternative pluck set removal procedure during slaughter.

This study compared the current pig slaughter procedure where the pluck set is completely removed with a procedure where the pluck set is partially removed, leaving the highly contaminated oral cavity, tonsils and tongue untouched. The effect on carcass contamination was investigated by enumerating hygiene indicator bacteria (total aerobic count, Enterobacteriaceae and E. coli) and cefotaxime-resistant E. coli (CREC) as well as assessing Salmonella and Yersinia enterocolitica presence on the sternum, elbow and throat of pig carcasses. Using the alternative pluck set removal, significantly lower mean numbers of hygiene indicator bacteria on throat samples and E. coli on elbow samples were found. Less pig carcasses were highly contaminated and a lower presence and level of CREC was observed. No difference in Salmonella or Yersinia enterocolitica presence was seen. The data in this study can help to assess the effect of this alternative procedure on the safety of pork and subsequently public health.

Quantification, distribution and diversity of ESBL/AmpC-producing Escherichia coli on freshly slaughtered pig carcasses.

This study quantified cefotaxime-resistant E. coli (CREC) on nine different carcass areas of 104 freshly slaughtered pig carcasses. In 49% [95% confidence interval (95% CI): 29-69%] of the carcasses CREC could be isolated and enumerated (using Tryptone Bile Agar with X-Glucuronide supplemented with 1 mg/L cefotaxime). Proportions of positive samples varied between carcass areas from 1% [95% CI: 0-10%] (loin) to 23% [95% CI: 10-44%] (head). Maximum concentrations on positive samples ranged between -0.6 log10 CFU/cm2 (loin, elbow before evisceration) and 1.7 log10 CFU/cm2 (head). The head was significantly more frequently contaminated than the loin (p = 0.027) and ham (3% [95% CI: 1-15%]). The foreleg was significantly more frequently contaminated (20% [95% CI: 13-30%]) than the ham. Combination disk diffusion assays revealed that 81% of the CREC isolates were extended-spectrum beta-lactamases (ESBL) producers, 13% were AmpC cephalosporinases (AmpC) producers and 2% ESBL and AmpC co-producers. Genotyping denoted blaCTX-M-gr1 (63%) and blaTEM (40%) as most present antibiotic resistance genes. Multiple gene combinations in one isolate and multiple combinations of genotypes and phenotypes among isolates of one sample were observed. These quantitative data can be used for intervention strategies to lower human exposure to CREC.

Correlation between slaughter practices and the distribution of Salmonella and hygiene indicator bacteria on pig carcasses during slaughter.

This study investigated the distribution of hygiene indicator bacteria and Salmonella on pig carcasses. Moreover, the relation between hygiene indicator counts and Salmonella presence as well as associations between specific slaughter practices and carcass contamination were determined for each carcass area. Seven Belgian pig slaughterhouses were visited three times to swab five randomly selected carcasses at nine different areas, after evisceration and trimming. Information about slaughter practices was collected using a questionaire. In all samples, the E. coli and Salmonella presence was analyzed and Enterobacteriaceae and total aerobic bacteria were quantified. Average total aerobic counts ranged from 3.1 (loin, pelvic duct, ham) to 4.4 log10 CFU/cm2 (foreleg). Median Enterobacteriaceae numbers varied between 0.4 (ham) an 1.8 log10 CFU/cm2 (foreleg). E. coli and Salmonella presence ranged from 15% (elbow) to 89% (foreleg) and 5% (elbow) to 38% (foreleg), respectively. Positive relations were found between hygiene indicator counts and Salmonella presence at the head, sternum, loin and throat. Several slaughter practices, such as splitting the head and incising tonsils, were associated with higher levels of hygiene indicator bacteria and Salmonella. These findings can be used to educate slaughterhouse personnel and estimate the public health risk involved in consumption of different pork cuts.

Ensuring food safety in food donations: Case study of the Belgian donation/acceptation chain.

The food donation process in Belgium is mapped and analyzed to identify bottlenecks in compliance with the legal framework and implementation of food safety management, based on literature search and interviews with stakeholders (donors, acceptors, regulators and facilitators) in Belgium and at EU level. The study revealed that the food donation/acceptation chain is far less structured and organized than the conventional food supply chain. The fragmented landscape of many small food banks and charity organizations (acceptors), often directed by and working with volunteers without training in food safety and lack of knowledge of legal food hygiene requirements is a bottleneck to generate trust among food donors and restricts the provision of perishable products in food donations. Lack of refrigerated transport and insufficient cold/freezing capacity in food banks and charity organizations was identified as a barrier to distribute perishable products. Furthermore, in two cities in Flanders (Belgium), at some food donation centers, donated perishable food samples (n=72) were taken and subjected to microbiological analysis to determine their overall food quality, hygiene and food safety status. Twenty-two of 72 analyzed samples showed marginal microbiological quality based on numbers of yeast, lactic acid bacteria or total viable count. In three samples Listeria monocytogenes was detected per 25g among which one ready-to-eat cooked meat product which showed increased numbers of L. monocytogenes (3.5logCFU/g) and Enterobacteriaceae (6.7logCFU/g). Overall, in Belgium, most of the donated foods considers nonperishable foods, with more or less half of the food collected by the food banks being purchased with funds from FEAD (Fund for European Aid to the Most Deprived) and thus not derived from food losses. Efforts are being made by facilitators to provide a platform for better coordination of donors and acceptors to make more efficient use of food losses. Regulators at the national level are taking action to clarify and provide some flexibility in food hygiene regulation and initiatives on EU level to facilitate food donation in the combat of food losses are pending. As from the side of the acceptors, it is recommended to professionalize the acceptation chain in Belgium and seek for a more harmonized approach and concerted action.

Quantitative farm-to-fork human norovirus exposure assessment of individually quick frozen raspberries and raspberry puree.

A quantitative human norovirus (NoV) exposure model describing transmission of NoV during pre-harvest, harvest and further processing of soft red fruits exemplified by raspberries is presented. The outcomes of the model demonstrate the presence of NoV in raspberry puree or individual quick frozen (IQF) raspberry fruits and were generated by Monte Carlo simulations by combining GoldSim® and @Risk® software. Input data were collected from scientific literature, observational studies and assumptions. NoV contamination of soft red fruits is assumed to take place at farms by application of contaminated water for pesticides dilution or by berries' pickers shedding NoV. The model was built simulating that a collection center received berries from ten farms with a total of 245 food handlers picking soft red fruits during a 10-hour day shift. Given 0, 5 and 20 out of 245 berries' pickers were shedding NoV, these conditions were calculated to result in a mean NoV contamination of respectively 0.47, 14.1 and 36.2 NoV particles per kg raspberries in case all raspberries are mixed to one day-batch of 11tons. The NoV contamination of the fruits was mainly driven by the route of NoV shedding food pickers (95.8%) rather than by spraying contaminated pesticide water (4.2%) (baseline scenario with 5 shedding pickers and contaminated pesticide water). Inclusion of appropriate hand washing procedures or hand washing followed by hand disinfection resulted in estimated reductions of the mean NoV levels from 14.1 to 0.16 and 0.17 NoV particles per kg raspberries, respectively, for the baseline scenario with 5 out of 245 food pickers shedding NoV. The use of a mild heat treatment (30s at 75°C) during further processing of berries to purees was noted to reduce mean NoV levels substantially from 14.1 to 0.2 NoV particles per kg raspberry puree. For IQF raspberries, the NoV contamination is heterogeneously distributed and resulted in a mean contamination of 3.1 NoV particles per 250g package containing approximately 115 berries. This farm-to-fork model is a useful tool for evaluating NoV mitigation strategies in the soft red fruit supply chain.

Persistence of Pathogenic and Non-Pathogenic Escherichia coli Strains in Various Tropical Agricultural Soils of India.

The persistence of Shiga-like toxin producing E. coli (STEC) strains in the agricultural soil creates serious threat to human health through fresh vegetables growing on them. However, the survival of STEC strains in Indian tropical soils is not yet understood thoroughly. Additionally how the survival of STEC strain in soil diverges with non-pathogenic and genetically modified E. coli strains is also not yet assessed. Hence in the present study, the survival pattern of STEC strain (O157-TNAU) was compared with non-pathogenic (MTCC433) and genetically modified (DH5α) strains on different tropical agricultural soils and on a vegetable growing medium, cocopeat under controlled condition. The survival pattern clearly discriminated DH5α from MTCC433 and O157-TNAU, which had shorter life (40 days) than those compared (60 days). Similarly, among the soils assessed, the red laterite and tropical latosol supported longer survival of O157-TNAU and MTCC433 as compared to wetland and black cotton soils. In cocopeat, O157 recorded significantly longer survival than other two strains. The survival data were successfully analyzed using Double-Weibull model and the modeling parameters were correlated with soil physico-chemical and biological properties using principal component analysis (PCA). The PCA of all the three strains revealed that pH, microbial biomass carbon, dehydrogenase activity and available N and P contents of the soil decided the survival of E. coli strains in those soils and cocopeat. The present research work suggests that the survival of O157 differs in tropical Indian soils due to varied physico-chemical and biological properties and the survival is much shorter than those reported in temperate soils. As the survival pattern of non-pathogenic strain, MTCC433 is similar to O157-TNAU in tropical soils, the former can be used as safe model organism for open field studies.

Assessing the status of food safety management systems for fresh produce production in East Africa: evidence from certified green bean farms in Kenya and noncertified hot pepper farms in Uganda.

The farms of fresh produce farmers are major sources of food contamination by microbiological organisms and chemical pesticides. In view of their choice for farming practices, producers are influenced by food safety requirements. This study analyzes the role of food safety standard certification toward the maturity of food safety management systems (FSMS) in the primary production of fresh produce. Kenya and Uganda are two East African countries that export green beans and hot peppers, respectively, to the European Union but have contrasting features in terms of agricultural practices and certification status. In the fresh produce chain, a diagnostic instrument for primary production was used to assess context factors, core control and assurance activities, and system output to measure the performance of FSMS for certified green bean farms in Kenya and noncertified hot pepper farms in Uganda. Overall, our findings show that in Uganda, noncertified hot pepper farms revealed only a "basic level of control and assurance" activities in their FSMS, which was not satisfactory, because no insight into potential pesticide microbial contamination was presented by these farmers. On the other hand, certified green bean farms in Kenya had an "average level of control and assurance," providing insight into the delivered food safety and quality by the farmers. Farm size did not impact the maturity level of FSMS. This study confirms the role played by food safety standard certification toward the maturity of FSMS implemented in developing countries and demonstrates the possibility of Ugandan farms to upgrade agricultural practices in the fresh produce sector.

Belgian and Spanish consumption data and consumer handling practices for fresh fruits and vegetables useful for further microbiological and chemical exposure assessment.

A consumer survey was organized in Spain and Belgium to obtain consumption data and to gain insight into consumer handling practices for fresh vegetables consumed raw or minimally processed (i.e., heads of leafy greens, bell peppers, tomatoes, fresh herbs, and precut and packed leafy greens) and fruits to be consumed without peeling (i.e., apples, grapes, strawberries, raspberries, other berries, fresh juices, and precut mixed fruit). This information can be used for microbiological and/or chemical food safety research. After extensive cleanup of rough databases for missing and extreme values and age correction, information from 583 respondents from Spain and 1,605 respondents from Belgium (18 to 65 years of age) was retained. Daily intake (grams per day) was calculated taking into account frequency and seasonality of consumption, and distributions were obtained that can be used in quantitative risk assessment for chemical hazards with chronic effects on human health. Data also were recalculated to obtain discrete distributions of consumption per portion and the corresponding frequency of consumption, which can be used in acute microbiological risk assessment or outbreak investigations. The ranked median daily consumption of fruits and vegetables was similar in Spain and Belgium: apple > strawberry > grapes > strawberries and raspberries; and tomatoes > leafy greens > bell peppers > fresh herbs. However, vegetable consumption was higher (in terms of both portion and frequency of consumption) in Spain than in Belgium, whereas the opposite was found for fruit consumption. Regarding consumer handling practices related to storage time and method, Belgian consumers less frequently stored their fresh produce in a refrigerator and did so for shorter times compared with Spanish consumers. Washing practices for lettuce heads and packed leafy greens also were different. The survey revealed differences between these two countries in consumption and consumer handling practices, which can have an impact on outcomes of future microbiological or chemical risk assessment studies.

Effect of exposure to stress conditions on propidium monoazide (PMA)-qPCR based Campylobacter enumeration in broiler carcass rinses.

Campylobacter quantification by qPCR is unable to distinguish viable vs. dead cells in contrast to the culture-based ISO 10272-2 reference method. Propidium monoazide (PMA) has been used to overcome this disadvantage. A Campylobacter PMA-qPCR enumeration method was evaluated for its consistency and compared to the culture-based enumeration for both artificially and natural contaminated broiler carcass rinses. The PMA effect was further evaluated on stressed cells. Five conditions, commonly encountered during the slaughter process and storage (acid, heat, cold, oxidation and freezing), were inflicted to the broiler carcass rinses artificially contaminated with Campylobacter jejuni or Campylobacter coli. A better correlation between the reference method and the qPCR enumeration was obtained when PMA was used. The two cultured-based methods used showed a significant CFU reduction for heat, cold and acid stresses although the PMA-qPCR enumeration showed that viable bacteria were underestimated. Freezing showed the highest reduction effect, while the reduction extend was also overestimated by the microbiological enumeration procedure. Exposure to a mild oxidative stress was the only stress condition applied at temperatures permitting adaptation of Campylobacter and did not lead to either reduction in CFU nor in the PMA-qPCR signal.

Multiplex real-time PCR and culture methods for detection of Shiga toxin-producing Escherichia coli and Salmonella Thompson in strawberries, a lettuce mix and basil.

An appropriate approach of high throughput multi-screening was verified for Shiga toxin-producing Escherichia coli (STEC) and Salmonella spp. in strawberries, lettuce and basil. Sample replicates were inoculated with STEC O157 or O26 and Salmonella Thompson (ca. 10-70, 100-700 and 1000-7000 cfu/25 g) and analysed after 1 and 5 days of storage (strawberries and lettuce at 7 °C and basil at 10 °C). After 18-24 h of enrichment at 37 °C in buffered peptone water, detection was performed using the GeneDisc multiplex PCR (stx1, stx2, eae and iroB genes) and selective culture media for isolation of STEC (with immunomagnetic separation (IMS)) and Salmonella spp. in parallel. After 1 day, the pathogenic strains were recovered from all samples for all inoculum levels, whereas reduced detection rates of STEC O157 and S. Thompson were observed after 5 days of storage in case of strawberries, in particular for the lowest inoculums level, suggesting superior survival potential for STEC O26. Overall, this study indicates the ability of PCR based screening methods for reproducible multi-detection of low numbers (10-70 cfu/25 g) of STEC and Salmonella in this type of foods. However, for the basil samples, PCR needed twofold dilution of the DNA extract to overcome inhibition. It was noted that on several occasions growth of competitive microbiota obstructed finding presumptive colonies on the selective agar media, whereas the use of an additional agar medium such as CHROMagar STEC (without IMS) improved recovery rate of STEC.

Risk ranking of pathogens in ready-to-eat unprocessed foods of non-animal origin (FoNAO) in the EU: initial evaluation using outbreak data (2007-2011).

Foods of non-animal origin (FoNAO) are consumed in a variety of forms, being a major component of almost all meals. These food types have the potential to be associated with large outbreaks as seen in 2011 associated with VTEC O104. In order to identify and rank specific food/pathogen combinations most often linked to human cases originating from FoNAO in the EU, a semi-quantitative model was developed using seven criteria: strength of associations between food and pathogen based on the foodborne outbreak data from EU Zoonoses Monitoring (2007-2011), incidence of illness, burden of disease, dose-response relationship, consumption, prevalence of contamination and pathogen growth potential during shelf life. The top ranking food/pathogen combination was Salmonella spp. and leafy greens eaten raw followed by (in equal rank) Salmonella spp. and bulb and stem vegetables, Salmonella spp. and tomatoes, Salmonella spp. and melons, and pathogenic Escherichia coli and fresh pods, legumes or grains. Despite the inherent assumptions and limitations, this risk model is considered a tool for risk managers, as it allows ranking of food/pathogen combinations most often linked to foodborne human cases originating from FoNAO in the EU. Efforts to collect additional data even in the absence of reported outbreaks as well as to enhance the quality of the EU-specific data, which was used as input for all the model criteria, will allow the improvement of the model outputs. Furthermore, it is recommended that harmonised terminology be applied to the categorisation of foods collected for different reasons, e.g. monitoring, surveillance, outbreak investigation and consumption. In addition, to assist future microbiological risk assessments, consideration should be given to the collection of additional information on how food has been processed, stored and prepared as part of the above data collection exercises.

An educationally inspired illustration of two-dimensional Quantitative Microbiological Risk Assessment (QMRA) and sensitivity analysis.

Quantitative Microbiological Risk Assessment (QMRA) is a structured methodology used to assess the risk involved by ingestion of a pathogen. It applies mathematical models combined with an accurate exploitation of data sets, represented by distributions and - in the case of two-dimensional Monte Carlo simulations - their hyperparameters. This research aims to highlight background information, assumptions and truncations of a two-dimensional QMRA and advanced sensitivity analysis. We believe that such a detailed listing is not always clearly presented in actual risk assessment studies, while it is essential to ensure reliable and realistic simulations and interpretations. As a case-study, we are considering the occurrence of listeriosis in smoked fish products in Belgium during the period 2008-2009, using two-dimensional Monte Carlo and two sensitivity analysis methods (Spearman correlation and Sobol sensitivity indices) to estimate the most relevant factors of the final risk estimate. A risk estimate of 0.018% per consumption of contaminated smoked fish by an immunocompromised person was obtained. The final estimate of listeriosis cases (23) is within the actual reported result obtained for the same period and for the same population. Variability on the final risk estimate is determined by the variability regarding (i) consumer refrigerator temperatures, (ii) the reference growth rate of L. monocytogenes, (iii) the minimum growth temperature of L. monocytogenes and (iv) consumer portion size. Variability regarding the initial contamination level of L. monocytogenes tends to appear as a determinant of risk variability only when the minimum growth temperature is not included in the sensitivity analysis; when it is included the impact regarding the variability on the initial contamination level of L. monocytogenes is disappearing. Uncertainty determinants of the final risk indicated the need of gathering more information on the reference growth rate and the minimum growth temperature of L. monocytogenes. Uncertainty in the dose-response relationship was not included in the analysis, hence the level of its influence cannot be assessed in the present research. Finally, a baseline global workflow for QMRA and sensitivity analysis is proposed.

Small Bacillus cereus ATCC 14579 subpopulations are responsible for cytotoxin K production.

AIMS: Bacillus cereus diarrhoeal food poisoning can be caused by several potential enterotoxins, including the nonhaemolytic enterotoxin (Nhe), haemolysin BL (Hbl) and cytotoxin K (CytK). To get more insights into the CytK expression, a fluorescent reporter strain was created for CytK expression. METHODS: Bacillus cereus ATCC 14579 was used as the reporter strain that contained the cyan fluorescent protein (CFPopt) gene under control of the cytK promoter. Transcription of enterotoxin genes nheB, hblC and cytK was assessed by messenger RNA analysis (RT-qPCR), and their full expression was assessed by immunological protein detection in the case of Nhe and Hbl and fluorescence microscopy in the case of CytK, using the reporter gene CFPopt. RESULTS: Transcription of enterotoxins Nhe, Hbl and CytK showed similar kinetics with a peak during the late exponential growth phase. Toxin expression of the reporter strain was unaltered in comparison with the wild type. However, fluorescence, and thus CytK expression, only occurred in a small (1-2%) portion of the cell population. CONCLUSIONS: These results suggest that a small subpopulation of B. cereus ATCC 14579 is responsible for CytK production in a homogeneous monoculture. SIGNIFICANCE AND IMPACT OF THE STUDY: Future research is warranted to determine whether genetically homogeneous B. cereus populations utilize differential gene expression for other toxins and virulence genes than CytK and whether this also applies to other B. cereus strains. If so, differential expression of toxin genes could be used by these bacteria to increase the fitness and survival chances of their population by diversification and specialization into different subpopulations.

Assessment of human exposure to 3rd generation cephalosporin resistant E. coli (CREC) through consumption of broiler meat in Belgium.

Acquired resistance of Escherichia coli to 3rd generation cephalosporin antimicrobials is a relevant issue in intensive broiler farming. In Belgium, about 35% of the E. coli strains isolated from live broilers are resistant to 3rd generation cephalosporins while over 60% of the broilers are found to be carrier of these 3rd generation cephalosporin resistant E. coli (CREC) after selective isolation. A model aimed at estimating the exposure of the consumer to CREC by consumption of broiler meat was elaborated. This model consists of different modules that simulate the farm to fork chain starting from primary production, over slaughter, processing and distribution to storage, preparation and consumption of broiler meat. Input data were obtained from the Belgian Food Safety agencies' annual monitoring plan and results from dedicated research programs or surveys. The outcome of the model using the available baseline data estimates that the probability of exposure to 1000 colony forming units (cfu) of CREC or more during consumption of a meal containing chicken meat is ca. 1.5%, the majority of exposure being caused by cross contamination in the kitchen. The proportion of CREC (within the total number of E. coli) at primary production and the overall contamination of broiler carcasses or broiler parts with E. coli are dominant factors in the consumer exposure to CREC. The risk of this exposure for human health cannot be estimated at this stage given a lack of understanding of the factors influencing the transfer of cephalosporin antimicrobial resistance genes from these E. coli to the human intestinal bacteria and data on the further consequences of the presence of CREC on human health.

Microbiological performance of a food safety management system in a food service operation.

The microbiological performance of a food safety management system in a food service operation was measured using a microbiological assessment scheme as a vertical sampling plan throughout the production process, from raw materials to final product. The assessment scheme can give insight into the microbiological contamination and the variability of a production process and pinpoint bottlenecks in the food safety management system. Three production processes were evaluated: a high-risk sandwich production process (involving raw meat preparation), a medium-risk hot meal production process (starting from undercooked raw materials), and a low-risk hot meal production process (reheating in a bag). Microbial quality parameters, hygiene indicators, and relevant pathogens (Listeria monocytogenes, Salmonella, Bacillus cereus, and Escherichia coli O157) were in accordance with legal criteria and/or microbiological guidelines, suggesting that the food safety management system was effective. High levels of total aerobic bacteria (>3.9 log CFU/50 cm(2)) were noted occasionally on gloves of food handlers and on food contact surfaces, especially in high contamination areas (e.g., during handling of raw material, preparation room). Core control activities such as hand hygiene of personnel and cleaning and disinfection (especially in highly contaminated areas) were considered points of attention. The present sampling plan was used to produce an overall microbiological profile (snapshot) to validate the food safety management system in place.

Review: norovirus prevalence in Belgian, Canadian and French fresh produce: a threat to human health?

Foodborne viruses, especially noroviruses (NoV), are increasingly reported as the cause of foodborne outbreaks. NoV outbreaks have been reported linked to fresh soft red fruits and leafy greens. Belgium, Canada and France were the first countries to provide data about the prevalence of NoV on fresh produce. In total, 867 samples of leafy greens, 180 samples of fresh soft red fruits and 57 samples of other types of fresh produce (tomatoes, cucumber and fruit salads) were analyzed. Firstly, the NoV detection methodology, including virus and RNA extraction, real-time RT-PCR and quality controls were compared among the three countries. In addition, confirmation and genotyping of the NoV strains was attempted for a subset of NoV positive samples using conventional RT-PCR targeting an alternative region followed by sequencing. Analysis of the process control showed that 653, 179 and 18 samples of the leafy greens, soft red fruits and other fresh produce types were valid for analysis based on the recovery of the process control. NoV was detected by real-time RT-PCR in 28.2% (N=641), 33.3% (N=6) and 50% (N=6) of leafy greens tested in Canada, Belgium and France, respectively. Soft red fruits were found positive by real-time RT-PCR in 34.5% (N=29) and 6.7% (N=150) of the samples tested in Belgium and France, respectively. 55.5% (N=18) of the other fresh produce types, analyzed in Belgium, were found NoV positive by real-time RT-PCR. Conventional RT-PCR resulted in an amplicon of the expected size in 19.5% (52/266) of the NoV positive samples where this assay was attempted. Subsequent sequencing was only successful in 34.6% (18/52) of the suspected amplicons obtained by conventional RT-PCR. From this study, using the described methodology, NoV genomes were frequently detected in fresh produce however sequence confirmation was not successful for the majority of the samples tested. Infection or outbreaks were rarely or not known to be related to the NoV positive samples. With the increase in sensitivity of the detection methodology, there is an increasing concern about the interpretation of positive NoV results by real-time amplification. Strategies to confirm the results by real-time RT-PCR should be developed in analogy with the detection of microbial pathogens in foods. Detection might indicate contact with NoV in the fresh produce chain. Consequently, a potential risk for infection cannot be excluded but the actual risk from RT-PCR NoV positive produce is still unknown. Studies should be designed determining the probability of infection related to the presence or levels of NoV genomic copies.