Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A One-Year Survey of Norovirus in UK Oysters Collected at the Point of Sale

A One-Year Survey of Norovirus in UK Oysters Collected at the Point of Sale Contamination of bivalve shellfish, particularly oysters, with norovirus is recognised as a food safety risk and a potential contributor to the overall burden of gastroenteritis in the community. The United Kingdom (UK) has comprehensive national baseline data on the prevalence, levels, and seasonality of norovirus in oysters in production areas resulting from a previ- ous two-year study (2009–2011). However, previously, data on final product as sold to the consumer have been lacking. As part of a wider project to establish the overall burden of foodborne norovirus in the UK, this study aimed to address this data gap. A one-year survey of oysters collected from the point-of-sale to the consumer was carried out from March 2015 to March 2016. A total of 630 samples, originating in five different European Union Member States, were collected from 21 regions across the UK using a randomised sampling plan, and tested for norovirus using a method compliant with ISO 15216-1, in addition to Escherichia coli as the statutory indicator of hygiene status. As in the previous production area study, norovirus RNA was detected in a high proportion of samples (68.7%), with a strong winter seasonality noted. Some statistically significant differences in prevalences and levels in oysters from different countries were noted, with samples originating in the Netherlands showing lower prevalences and levels than those from either the UK or Ireland. Overall, levels detected in positive samples were considerably lower than seen previously. Investigation of potential contributing factors to this pattern of results was carried out. Application of normalisation factors to the data from the two studies based on both the numbers of norovirus illness reports received by national surveillance systems, and the national average environmental temperatures during the two study periods resulted in a much closer agreement between the two data sets, with the notably different numbers of illness reports making the major contribution to the differences observed in norovirus levels in oysters. The large majority of samples (76.5%) contained no detectable E. coli; however, in a small number of samples (2.4%) levels above the statutory end product standard (230 MPN/100 g) were detected. This study both revealed the high prevalence of norovirus RNA in oysters directly available to the UK consumer, despite the high level of compliance with the existing E. coli-based health standards, while also highlighting the difficulty in comparing the results of surveys carried out in different time periods, due to variability in risk factors. Keywords Norovirus · Oysters · qRT-PCR · Survey Introduction Contamination of bivalve shellfish, particularly oysters, with norovirus is recognised as a food safety risk, with a Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s1256 0-018-9338-4) contains considerable number of reports of outbreaks in the litera- supplementary material, which is available to authorized users. ture (reviewed in Bellou et al. 2013). In both the European Union and the United States, viral contamination in shellfish * J. A. Lowther is regulated indirectly using enteric bacteria as an indicator james.lowther@cefas.co.uk of faecal pollution (Anonymous 2004, 2015b). However, this Centre for Environment, Fisheries and Aquaculture Science, approach has been demonstrated to inadequately address Weymouth DT4 8UB, UK the risk from human enteric viruses, with a number of viral Institute of Psychology, Health & Society, University outbreaks caused by batches compliant with the regulations of Liverpool, Liverpool, England, UK Vol:.(1234567890) 1 3 Food and Environmental Virology (2018) 10:278–287 279 (Chalmers and McMillan 1995; Le Guyader et al. 2008; wholesalers. A total of 373 vendors were identified across Dore et al. 2010). Considerable progress has been made the 21 areas. towards the development of detection methods for norovirus A randomised sampling plan was drawn up aiming to in molluscan shellfish and an ISO/CEN technical specifica- obtain a total of 630 oyster samples over a 1-year period tion including such a method (ISO/TS 15216) was published (16th March 2015–15th March 2016), with monthly tar- in 2013 (Anonymous 2013), with a subsequent update to gets of 26 samples in the truncated months of March 2015 a fully validated standard in 2017 (Anonymous 2017). EU and March 2016, and 53 or 52 samples alternatively for the legislative texts foreshadow the adoption of virus controls months April 2015 to February 2016. For each month, the in bivalve shellfish when the methods are sufficiently devel- vendors targeted were selected randomly from a subset of oped (Anonymous 2005) and the options for improvement the list of all 373 vendors with no weighting by region. Over of EU legislation to better address the virus risk have been the course of the survey, any shortfall in sample numbers actively discussed in recent years (EFSA Panel on Biologi- collected in a given month was compensated by the addition cal Hazards 2012). It has therefore become important to of extra samples (selected at random from the same region gain information about the application of the new methods, for logistical regions) in the sampling schedule for the fol- and the potential impact of possible legislative standards on lowing month. bivalve shellfish production. Amongst current European Union Member States, the Sample Collection United Kingdom (UK) has some of the most comprehen- sive national baseline data on the prevalence, levels and sea- Samples (except online sales) were pre-ordered through sonality of norovirus in oysters resulting from a two-year direct contact with the vendor, then collected by sam- study carried out on samples taken directly from production pling officers at the point-of-sale to the consumer. Samples areas in 2009–2011. (Lowther et al. 2012b). However, until sourced from online sales were ordered for delivery to the recently, data on final product as sold to the consumer in the sampling officer. Within each vendor, samples were lim- UK have been lacking, with to the best of our knowledge ited to native, Pacific, or other oyster species, sold as either the only published study having tested oysters from a single ambient, chilled, or frozen. To avoid possible contamination dispatch centre only (Lowther et al. 2010). As part of a wider by food handlers live animals in restaurants were obtained project to establish the overall burden of foodborne norovi- before shucking by restaurant staff. Cooked, pasteurised, rus in the UK (“NoVAS: Assessing the contribution made smoked, or otherwise processed oysters were not sampled. by the food chain to the burden of UK-acquired norovirus Where multiple products or batches of the same product infection”—UK Food Standards Agency Project reference: were available, one was picked at random by the sampler. A FS101040), this study aimed to address this data gap, as well sample consisted of individual animals from the same batch as to compare results for final product with those obtained (same origin and production date). in the previous production area survey. Given sufficient availability samples consisted of 25 oys- ters (with a minimum number of 12 oysters required for a valid sample). At the point of sampling, full sample details Materials and Methods including date, time, vendor name and address, sample type, sample temperature at the point of sale (ambient, fresh, fro- Sampling Plan zen), sample origin/identification mark (if available) were recorded by the sampling officer. A high-resolution digital The survey design was informed by a comprehensive prac- photograph of the sample packaging and identification mark tical evaluation of the purchase routes for oysters available (if available) was taken. This information with accompany- to the UK consumer. This evaluation was undertaken by a ing photographs was then e-mailed to the Stericycle project specialist product retrieval company contracted to collect the co-ordinator for collation in a sample database. survey samples (Stericycle ExpertSOLUTIONS, Reading, Samples were packaged in temperature controlled Cole- UK), through phone interviews with and physical visits to man food boxes with cool packs and despatched to the labo- identified vendors. This market research was conducted in ratory via overnight courier service accompanied by a sam- 21 selected cities/regions across the UK (selected to give a ple submission form including a unique sample identifier, good geographical spread and including regions in each of the date and time of collection, the storage temperature of the four constituent countries of England, Wales, Scotland the sample at the collection point and the date and time of and Northern Ireland). Vendors directly available to con- despatch. Details of the vendor and the origin of the oysters sumers of oysters were subdivided into the following types: were not included such that the sample testing was carried supermarkets, fishmongers, restaurants, online sales and out blind. 1 3 280 Food and Environmental Virology (2018) 10:278–287 Upon receipt at the laboratory, each sample was pro- COG2R primers, and QNIFS probe were used (Kageyama cessed according to standard procedures. If the sample tem- et al. 2003, Loisy et al. 2005). For mengo virus, the mengo perature on receipt was > 18 °C, fewer than 10 live animals 110 and mengo 209 primers, and the mengo 147 probe were were available, or the condition of the sample was otherwise used (Costafreda et al. 2006). For both norovirus genogroup- unsatisfactory, samples were not tested, and replacement specific assays, three aliquots of 5 μl sample or extraction samples were collected. In addition, if the sample tempera- control RNA was tested in 25 µl total volume with one-step ture on receipt was > 10 °C, fewer than 20 live animals reaction mix prepared using the RNA Ultrasense one-step were available, or a period of > 48 h had elapsed between qRT-PCR system (Invitrogen) (final concentrations of 1x sample collection and receipt at the laboratory, samples were Reaction Mix, 500 nM forward and 900 nM reverse primers, analysed for norovirus only (not E. coli); under these circum- and 250 nM probe, plus 0.5 µl Rox and 1.25 µl Enzyme Mix stances replacement samples were not sought. per reaction). For mengo virus, two aliquots of 5 μl cDNA were used. Amplification was performed using the following Detection and Quantification of Norovirus cycling parameters: 55 °C for 60 min, 95 °C for 5 min, and then 45 cycles of 95 °C for 15 s, 60 °C for 1 min and 65 °C Oyster samples were tested for norovirus according to the for 1 min on an Mx3005P real-time PCR machine (Strata- draft international standard ISO 15216-1 (now published in gene). Wells containing nuclease free water and the above Anonymous 2017). qRT-PCR reaction mixes were included on each plate as a negative control. Quantification used a log dilution series 5 1 Virus Extraction (range 1 × 10 to 1 × 10  copies/µl) of linear dsDNA mol- ecules carrying the GI and GII target sequences and fol- For each sample, ten oysters were selected. The digestive lowed the principles outlined in ISO 15216-1 (Anonymous tissues (stomach and digestive diverticula) of these oysters 2017). All samples were assessed for extraction efficiency by were excised, pooled, and then finely chopped using a razor the comparison of sample Ct values for mengo virus with a blade. A 2-g subsample of chopped digestive tissues was standard curve generated from the process control material transferred to a clean tube. 10 µl of mengo virus vMC0 tis- and for qRT-PCR inhibition using RNA external controls as sue culture supernatant was added to the 2-g subsample described in ISO 15216-1 (Anonymous 2017). Samples were as a within-sample virus/RNA extraction process control. retested if extraction or qRT-PCR inhibition levels fell below Homogenates were prepared by adding 2 ml of a 100 μg/ml 1% or above 75% respectively, where positive qRT-PCR con- Proteinase K solution to the digestive tissues. This was then trols indicated reagent failure, or for any positive sample incubated at 37 °C with shaking at 320 rpm for a duration where the negative extraction or PCR controls showed con- of 1 h, and subsequently incubated at 60 °C for a duration of tamination. Quantitative results were not adjusted for losses 15 min. Finally, the sample was centrifuged at 3000×g for during processing or RT-PCR inhibition. 5 min.; the volume of the soluble portion (homogenate) was measured and then retained for downstream testing and the Detection and quantification of E. coli pellet discarded. Homogenates were stored at 4 °C for up to one month prior to testing. Oyster samples were tested for E. coli according to ISO 16649-3 (Anonymous 2015a). Whole animal homogenates RNA Extraction were prepared from the flesh and intravalvular fluid of 10 oysters and assayed using a most-probable-number (MPN) Total RNA was extracted from 500 µl of shellfish homogen- method. Results are expressed per 100 g of shellfish flesh ate using a NucliSENS miniMAG extraction machine and and intravalvular fluid. NucliSENS magnetic extraction reagents (BioMerieux) following the manufacturer’s instructions (eluting in 100 µl Statistical Analysis elution buffer). A negative (water only) extraction control sample was also prepared and tested in parallel with each Relevant statistical analyses (Fisher’s exact test, set of samples extracted. Eluted RNA was stored at − 20 °C Kruskal–Wallis test) were carried out using the Minitab until required. software package. For statistical analysis and calculation of geometric means, positive results of < 100 copies/g (the One‑Step qRT‑PCR limit of quantification of the assay) were scored at 50, and not detected samples were scored at 20 (half the limit of For norovirus GI, QNIF4 and NV1LCR primers, and TM9 detection). Scores for GI and GII were combined prior to probe were used (da Silva et al. 2007; Hoehne and Schreier analysis. In this way, samples that were not detected for both 2006; Svraka et al. 2007). For norovirus GII, QNIF2 and genogroups scored 40 copies/g, and this figure should be 1 3 Food and Environmental Virology (2018) 10:278–287 281 considered a baseline for levels. Confidence intervals (95%) both surveys, an average level for each calendar month was for datasets were calculated as the geometric mean ± 2x the calculated. geometric standard deviation; at the lower end, these are censored at 40  copies/g where the calculated value was less than this. Due to the large number of censored values Results and Discussion in the dataset, non-parametric statistical tests were used throughout. Norovirus Results All 630 samples subjected to testing passed quality control Normalisation Factors criteria for extraction efficiency and RT-PCR inhibition on initial or retesting. The average extraction efficiency obtained In order to compare the contribution of different risk fac- was 28.7% (range 1.1–99.6%), while the average RT-PCR tors to the results obtained in the current and previous stud- inhibition was 14.0% (range 0–74.3%). Of the 630 samples, ies, month-by-month normalisation factors were generated 433 (68.7%) were positive for norovirus RNA. Of these, 99 for norovirus illness (using data on illness reports in Eng- samples (15.7%) were positive for GI only, 88 (14.0%) were land and Wales provided by Public Health England) and positive for GII only and 246 (39.0%) were positive for both environmental temperatures (using data on UK average air GI and GII. A clear seasonality was observed with 79.7% temperatures obtained from the UK Meteorological Office of samples collected in the months October–March positive website—http://www.met office.gov.uk) as follows. compared with 57.0% in the months April–September. This For illness reports, the normalisation factor N was deter- difference was found to be statistically significant (Fisher’s mined as exact test; p < 0.0001). The highest and lowest monthly prevalences were recorded in February 2016 (96.3%) and N = , September 2015 (34.6%), respectively (Fig. 1a). where I is the average illness reports per day for the relevant calendar month in the period of the production area survey (May 2009–Apr 2011) and I average illness reports per day for the month in question, such that where illness reports for a given month were lower than the average for that calen- dar month in 2009–2011, the normalisation factor was > 1. For example, in April 2015, the average number of illness reports per day was 30.27, compared with the average for April during the production area survey of 34.05 reports per day. The normalisation factor for April 2015 was therefore 34.05 ÷ 30.27 = 1.12. For temperatures, the normalisation factor N was deter- mined as 20 − T N = , 20 − T where T is the long-term time series average temperature for the relevant calendar month (1981–2010) and T  is the recorded monthly UK average temperature for the month in question, such that where the UK average air temperature for a given month was higher than the long-term average for that calendar month in 1981–2010, the normalisation Fig. 1 Monthly proportion of samples giving total norovirus results factor was > 1. For example, in April 2015, the UK average in different quantity brackets (copies/g) in the current (retail) sur - temperature was 7.9 °C, compared with a long-term average vey and a previous production area survey. ND not detected. Results for April of 7.4 °C. The normalisation factor for April 2015 are for GI and GII combined; samples that were positive at levels of  <  100 copies/g for both genogroups are included in the  <  100 was therefore (20 − 7.4) ÷ (20 − 7.9) = 1.04. quantity bracket. a Results for the retail survey. b Results for the pro- Normalisation factors calculated in this way were applied duction area survey (Lowther et  al. 2012b)—proportions calculated to the geometric mean norovirus levels recorded for each for each calendar month across the survey duration, March shown month of both the retail and production area surveys. For twice to allow comparison with the retail survey 1 3 282 Food and Environmental Virology (2018) 10:278–287 Norovirus levels were also higher during the winter interval 40–129  copies/g) than in samples from the UK period with a geometric mean level of 87 copies/g (95% (71.7% positive, geometric mean of 78 copies/g, 95% con- confidence interval 40–309 copies/g) in the months Octo- fidence interval 40–277 copies/g). These differences were ber–March compared with 65 copies/g (95% confidence found to be statistically significant (Fisher’s exact test, interval 40–202 copies/g) in samples collected from April to p = 0.0144; Kruskal–Wallis test, p < 0.001). Further sub- September. This difference was found to be statistically sig- division of non-UK samples to enable country-by-country nificant using the Kruskal–Wallis test (p  < 0.001). The high- analysis showed that for oysters from the Netherlands both est levels recorded in individual samples were 586 copies/g prevalence and levels were significantly lower than for for GI and 1802 copies/g for GII; however, in the majority the UK (Fisher’s exact test, p < 0.0001; Kruskal–Wallis of samples testing positive (85.9%), the levels recorded were test, p < 0.0001). Prevalence and levels for oysters from below the limit of quantification of the assay (100 copies/g) the Republic of Ireland were not significantly different for both norovirus GI and GII. In total, 61 samples pro- from those for the UK, but were significantly higher than duced results of > 100 copies/g for one or both genogroups, those for the Netherlands (Fisher’s exact test, p = 0.0001; representing 14.1% of positive samples, and 9.7% of total Kruskal–Wallis test, p = 0.0081). No apparent seasonal bias samples. Of these 61 samples, 7 produced results of > 100 in collection dates for samples from the three countries were copies/g for both genogroups, 2 for GI only and 52 for GII found to explain these differences (no significant difference only. The highest monthly incidence of samples giving was found between the proportions of samples collected dur- results > 100 copies/g was March 2015 (37.5%). Over the ing the winter months October–March using Fisher’s exact course of the survey, 5 samples (0.8% of total samples) pro- test). Statistical analysis of norovirus results for samples duced results for GI and GII combined of > 1000 copies/g; from France and Spain was not carried out due to the small three of these samples were collected in September 2015, number of samples. and 2 in February 2016. Comparison with the Production Area Study Comparison of Oysters Originating in Different Countries The prevalence of norovirus RNA in oyster samples recorded in this survey (the “retail survey”; 68.7%) was similar but For 492 samples (78.1% of the total), the dispatch centre slightly lower than that found in a previous two-year sur- from which the oysters originated could be identified as a vey (2009–2011) of oysters from UK production areas (the result of information collected by the sampling officer (this “production area survey” 76.2%) (Lowther et al. 2012b). In identification was supported by a photograph of the identi- addition, a similar seasonality with increased prevalences fication mark or other identifying labels/packaging in 378 and levels in the winter months was noted in both surveys. cases). Oysters originated from 33 die ff rent dispatch centres However, the overall levels of norovirus recorded in the in 5 different EU Member States. Of the 492 samples with retail survey were considerably lower than in the production identified dispatch centres, 434 samples (88.2%) originated area survey. In the latter, 36.5% of total samples contained in the UK, 29 (5.9%) from the Netherlands, 25 (5.1%) from levels  >  100  copies/g (the limit of quantification of the Ireland, 3 (0.6%) from France and 1 (0.2%) from Spain. assay) for one or both norovirus genogroups, combined lev- Prevalences of norovirus detection and geometric mean els of > 1000 copies/g were found in 14.6% of samples, and levels of norovirus for samples originating in different EU combined levels of > 10,000 copies/g were found in 1.1% of Member States are shown in Table 1. samples. Geometric means for all results were 76 copies/g Overall prevalence and levels of norovirus were lower (95% confidence interval 40–261 copies/g) and 159 copies/g in samples originating outside the UK (55.2% of samples (95% confidence interval 40–2964 copies/g) for the retail positive, geometric mean of 58 copies/g, 95% confidence and production area surveys, respectively. This difference Table 1 Norovirus results by Country of origin Number of Norovirus results country of origin samples Prevalence (% posi- Geometric mean (copies/g; 95% tive) (%) confidence interval in parentheses) UK 434 71.7 78 (40–277) Netherlands 29 31.0 49 (40–91) Ireland 25 84.0 69 (40–120) France 3 33.3 48 (40–92) Spain 1 100.0 275 (n/a) 1 3 Food and Environmental Virology (2018) 10:278–287 283 was found to be statistically significant (Kruskal–Wallis test; normalisation factors were determined using the PHE data p < 0.001). on illness reports in England and Wales (treating these fig- Possible underlying causes for this pattern of results ures as a proxy for community levels as a whole) and Met include:- Office data on UK national average monthly air temperatures (treating these figures as a proxy for environmental tem- (1) Risk reduction measures by Food Business Operators peratures as a whole—equivalent national average seawater including e.g. use of enhanced depuration conditions or temperatures are not available) as described in materials and use of norovirus testing to inform decisions on choice methods. of supply for processing and marketing Application of the normalisation factors based on illness (2) Representativeness of samples: It is possible that the reports resulted in a notable improvement in correspond- production area survey was not representative of the ence in results by calendar month between the two surveys volumes of oysters placed on the UK market as the (see Fig. 2). Geometric mean levels for each month in the selection of sites for the production area survey was two surveys are plotted against each other in Fig. 2b, d, f, h meant to provide a representative selection of pro- alongside lines of best fit and equality; for data normalised duction areas with different risk profiles and a good according to illness reports (Fig. 2c, d), the slope of the line geographical spread, but not to represent production of best fit (0.4723) is considerably closer to equality and volumes or market share. the correlation is considerably closer to total (r  = 0.9506) (3) Variation in norovirus shedding rates in the community: than for non-normalised data (Fig. 2a, b; slope = 0.0887 and Sewage treatment is known to only reduce norovirus r  = 0.5384). by a limited extent (Campos and Lees 2014). Conse- Application of the normalisation factors for temperature quently, a key factor influencing norovirus contami- in isolation yielded only a modest improvement in agree- nation in filter-feeding shellfish impacted by sewage ment between the results of the two studies (Fig.  2e, f). discharges will be the degree of virus infection, and However, application of both the illness and temperature- hence the degree of virus shedding in faeces, in the based normalisation factors in combination produced the population contributing to the sewage inputs. Dur- best line of best fit overall (Fig.  2g, h; slope = 0.5626 and ing this study, unusually low levels of norovirus were r  = 0.9576). observed in the community in England and Wales dur- This analysis indicates that much of the difference in the ing the winter of 2015–2016, particularly during the norovirus levels between the retail and production area sur- months November to January, compared with unusu- veys can be attributed to the different levels of norovirus in ally high levels during the winters of 2009–2010 and the community between the two study periods, with some 2010–2011 (Supplementary Figure S1, data provided portion of the remaining difference explained by the differ - by Public Health England, equivalent data for other ing temperatures, particularly during the early part of winter. parts of the UK are not available). Nevertheless, even normalising using these factors together (4) Variation in environmental temperatures: Shellfish are results in levels in the retail study on average ~ 56% as high poikilothermic (Gosling 2008), and their metabolic as during the production area survey, suggesting other fac- rate, and hence the degree of contaminant uptake and tors as discussed above also contributed to the different pat- removal, is significantly influenced by the temperature tern of results. of their environment. In this study, environmental tem- peratures in the UK were unusually high during the E. coli Results winter of 2015–2016, particularly during the months November to January, compared with unusually low Out of 630 samples received, E. coli analysis was carried temperatures during the production area study winters out in 452 cases (71.7%). For the remaining samples, E. of 2009–2010 and 2010–2011 (Supplementary Figure coli testing was not carried out primarily due to insufficient S2; data obtained from the UK Meteorological Office live animals in the sample to conduct this test in addition website—http://www.met office.gov.uk). to norovirus analysis (< 20), or elevated temperatures on arrival (> 10 °C). Of the above factors potentially influencing the variation Of the samples tested for E. coli, the bacterium was not seen between contamination levels in the production area detected (< 18 MPN/100 g) in 346 cases (76.5%). In 11 study and in this study, it was only possible to perform fur- samples (2.4%), levels in excess of the EU legal end product ther analysis on the impact of general population shedding standard (230 MPN/100 g; Anonymous 2005) were detected. rates and environmental temperatures due to the unavail- In these cases, the UK Food Standards Agency as the Com- ability of data relevant to the other factors. To further inves- petent Authority was informed on the same working day tigate these possible contributing elements, month-by-month that the result became available. All 11 of these samples 1 3 284 Food and Environmental Virology (2018) 10:278–287 Fig. 2 Application of normalisation factors to monthly geometric and best fit (dotted and labelled with associated equation and r val- mean norovirus levels obtained during the retail and production area ues) are shown. a, b No normalisation applied. c, d Normalisation surveys. a, c, e, f comparison of monthly geomean levels for the retail factors derived from illness reports applied. e, f Normalisation factors (dashed lines) and production area (dotted lines) surveys. b, d, f, h; derived from average temperatures applied. g, h Normalisation fac- correlation between geometric mean norovirus levels for each calen- tors derived from illness reports and average temperatures applied dar month obtained during the two surveys. Lines of equality (solid) 1 3 Food and Environmental Virology (2018) 10:278–287 285 were collected between March and September 2015, with the human health risks is complex (EFSA Panel on Biological highest monthly incidence of five samples > 230 MPN/100 g Hazards 2012); however in an analysis of outbreak-related in July 2015, representing 15.2% of the samples collected in oyster samples carried out in this laboratory (Lowther et al. that month. In one sample, a level in excess of the upper limit 2012a), an association between increased norovirus lev- of quantification of the E. coli assay (> 18,000 MPN/100 g) els and increased likelihood of norovirus-type illness was was recorded from a sample collected on 15 July 2015. observed, with no outbreak-related sample recording levels In comparison with the production area survey, levels of below 152 copies/g. The human health consequences of the E. coli recorded in this study were very low. No E. coli was large proportion of positive samples in this survey are there- detected in the majority of the samples, while results over fore not certain. the A classification and end product standard were rare. The majority of oyster samples tested originated from In the production area survey by contrast, E. coli propor- dispatch centres in the UK (88.2% of samples where the tions were 14.3% undetected and 40.0% > 230 MPN/100 g. dispatch centre could be identified), with the remainder Although other factors may have contributed, this difference originating in other countries in Western Europe. System- is likely to be largely the result of the well-established high atic comparison of prevalences and levels of norovirus in efficacy of standard depuration conditions for the removal of oysters from different countries was complicated by the low E. coli bacteria (Doré and Lees 1995). Since the removal of numbers of samples from each exporter country; however E. coli is a good proxy for other bacterial pathogens derived oysters from the Netherlands showed significantly lower from sewage contamination (Lees 2000), this demonstrates levels and prevalences than oysters from both the UK and the contribution to public health of the classification and Ireland. There is some evidence that oyster growing waters depuration regulations for protection from bacterial illness. in the Netherlands are impacted by lower levels of faecal This finding is supported by the low numbers of bacterial pollution; six out of seven (86%) oyster production areas in infections associated with consumption of oysters in the UK the country are at the time of writing classified A (Nether - (Lees 2000). lands National Reference Laboratory for monitoring bacte- The small number of results of > 230 MPN/100 g, includ- riological and viral contamination of bivalve molluscs; per- ing one result of > 18,000 MPN/100 g, indicates that despite sonal communication), the cleanest status based on E. coli the high level of adherence to the legal standards, problems monitoring results according to EU legislation (Anonymous can nevertheless occur. The root cause of the high E. coli 2004). By contrast, in the UK, 37% of oyster production levels detected in some samples could not be investigated, areas are wholly or partially classified A, either permanently but could conceivably be linked to problems post-harvest, or for part of the year (UK National Reference Laboratory during transportation, or at the point-of-sale. for monitoring bacteriological and viral contamination of bivalve molluscs; personal communication). During the year of the survey, some significant potential Conclusion risk factors were low compared with the previous study on oysters from UK production areas (Lowther et al. 2012b). The survey described here is the first systematic study of The number of norovirus cases in the general population norovirus in oysters collected at the point-of-sale in the UK. and hence the likely extent of virus shedding into shellfish Norovirus RNA was detected in 68.7% of samples tested, production areas was considerably lower than previously, comparable with the prevalence found in a previous survey and environmental temperatures during the winter were carried out using the same methods on oysters from UK higher. The datasets used to quantify these risk factors had production areas (76.2%; Lowther et al. 2012b). The preva- some limitations; air temperatures were used as an indicator lence described here is considerably higher than recorded of overall environmental temperatures, rather than directly in surveys of norovirus in bivalve shellfish collected at using seawater temperatures (no national average seawater the point-of-sale in some other countries, for example the temperature data is available). In addition, for both factors, United States (3.9%; Woods and Burkhardt 2010), France data collected in the UK were extrapolated to normalise (9%; Schaeffer et al. 2013) and Thailand (12.3%: Kittigul results based on all samples collected during the retail sur- et al. 2016); however, comparatively frequent detection of vey, including those originating outside the UK. However, norovirus has been reported in shellfish from production 430 out of 432 samples (99.6%) where origin data existed areas in Ireland (37.1%; Flannery et al. 2009), Italy (51.5%, either originated in the UK, or in bodies of water abutting Suffredini et al. 2014) and Spain (52.4%; Polo et al. 2015). UK territorial waters (the Irish Sea, the English Channel and Although the majority of samples were found to be positive, the North Sea), while the illness data used broadly reflects levels exceeding 100 norovirus copies/g were found in only global trends in norovirus infections. The two winter peri- a relatively small percentage of samples (9.7%). The rela- ods in which illness levels in the dataset used were highest tionship between levels of norovirus as detected by PCR and (2009–2010 and 2012–2013) both followed directly on from 1 3 286 Food and Environmental Virology (2018) 10:278–287 infection (NoVAS)”. The authors thank Laura Boyd, Michelle Paice the emergence of a global pandemic strain; New Orleans and the sampling team from Stericycle ExpertSOLUTIONS for collec- 2009 (Vega et al. 2011) and Sydney 2012 (van Beek et al. tion of oyster samples. We also thank the NoVAS Consortium for help- 2013), respectively. For these reasons, we therefore consid- ful comments on the manuscript. The NoVAS Consortium in addition ered that use of these suboptimal datasets for determination to the authors comprises the University of Liverpool (Miren Iturriza- Gomara), the University of East Anglia (Paul Hunter, Jim Maas), Pub- of normalisation factors was unlikely to confound the analy- lic Health England (David James Allen, Nicola Elviss, Andrew Fox), sis we carried out. Leatherhead Food Research (Angus Knight) and Fera Science Ltd. This analysis offers some insights into the contribution (Nigel Cook, Martin D’Agostino). of these two factors to the pattern of results observed in the different surveys, and highlights the difficulty of comparing Open Access This article is distributed under the terms of the Crea- tive Commons Attribution 4.0 International License (http://creat iveco results from surveys carried out in different time periods, or mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- of treating the results of a short survey as completely indica- tion, and reproduction in any medium, provided you give appropriate tive of the long-term characteristics of the surveyed area. It credit to the original author(s) and the source, provide a link to the is however possible that some of the differences observed Creative Commons license, and indicate if changes were made. were down to inherent differences between production area and retail-ready oysters. For example, Food Business Opera- tor risk management interventions (such as virus testing, or selection of product from cleaner areas) may have contrib- References uted to the low virus levels seen in this study. Direct com- Anonymous. (2004). Regulation (EC) No 854/2004 of the European parison, within the same time period, of levels in production Parliament and of the Council of 29 April 2004 laying down spe- areas with those seen in retail-ready oysters would assist cific rules for the organisation of official controls on products of assessment of the contribution made by producer practices. animal origin intended for human consumption. Official Journal An ongoing EU-wide survey of norovirus in oysters from of the European Communities. 25.06.2004, L226, 83–127. Anonymous. (2005). European Commission Regulation (EC) No both production areas and dispatch centres, organised by the 2073/2005 of 15 November 2005 on microbiological criteria for European Food Safety Authority (EFSA 2012) may help to foodstuffs. Official Journal of the European Union. 22.12.2005, illuminate this issue. L338, 1–26. The very high proportion of samples compliant with the Anonymous. (2013). Microbiology of food and animal feed—horizon- tal method for determination of hepatitis A virus and norovirus E. coli end product standard (97.6%) in this study indicates in food using real-time RT-PCR—Part 1: Method for quantifica- the good compliance with current regulatory requirements tion. ISO/TS 15216-1:2013 (withdrawn). Geneva: International in the UK oyster supply chain and the consequential prob- Organization for Standardization. able low risk from bacterial pathogens such as Salmonella. Anonymous. (2015a). Microbiology of the food chain—Horizon- tal method for the enumeration of beta-glucuronidase-positive However, the contrast between the results for norovirus and Escherichia coli—Part 3: Detection and most probable num- E. coli again illustrates the limitations of this organism as ber technique using 5-bromo-4-chloro-3-indolyl-ß-D-glucu- an indicator of viral risk in shellfish, particularly in depu - ronide. ISO 16649-3. Geneva: International Organization for rated animals. Shellfish-related outbreaks of norovirus have Standardization. Anonymous. (2015b). National Shellfish Sanitation Program. Guide often been linked to batches compliant with the end product for the Control of Molluscan Shellfish, 2015 Revision. https :// standard for E. coli (Chalmers and McMillan 1995; Le Guy- www.fda.gov/food/guida nceregulat ion/f ederals tatef oodpr ogr ams/ ader et al. 2008; Dore et al. 2010). On the other hand, we ucm20 06754 .htm. have previously observed a site-by-site correlation between Anonymous. (2017). Microbiology of the food chain—Horizontal method for determination of hepatitis A virus and norovirus using E. coli and norovirus levels in oysters from UK production real-time RT-PCR—Part 1: Method for quantification. ISO 15216- areas (Lowther et al. 2012b). 1:2017. Geneva: International Organization for Standardization. In summary, although illuminating the difficulties with Bellou, M., Kokkinos, P., & Vantarakis, A. (2013). Shellfish-borne comparing data collected in different periods, this study viral outbreaks: A systematic review. Food Environ. Virol., 5, 13–23. provides a comprehensive and systematic assessment of the Campos, C. J., & Lees, D. N. (2014). Environmental transmission of extent of both norovirus and E. coli contamination in oysters human noroviruses in shellfish waters. Applied and Environment available to the UK consumer. These data can assist risk Microbiology, 80, 3552–3561. managers both in the UK and beyond to manage the health Chalmers, J. W. T., & McMillan, J. H. (1995). An outbreak of viral gastroenteritis associated with adequately prepared oysters. Epi- risks posed by norovirus in oysters, and as part of the wider demiology and Infection, 115, 163–167. NoVAS project will contribute to an assessment of the con- Costafreda, M. I., Bosch, A., & Pintó, R. M. (2006). Development, tribution of the food chain to norovirus infections in the UK. evaluation, and standardization of a real-time TaqMan reverse transcription-PCR assay for quantification of hepatitis A virus in Acknowledgements This work was supported by the United Kingdom clinical and shellfish samples. Applied and Environment Micro- Food Standards Agency project FS101040: “Assessing the contribu- biology, 72, 3846–3855. tion made by the food chain to the burden of UK-acquired norovirus 1 3 Food and Environmental Virology (2018) 10:278–287 287 da Silva, A. K., Le Saux, J. C., Parnaudeau, S., Pommepuy, M., Elime- Loisy, F., Atmar, R. L., Guillon, P., Le Cann, P., Pommepuy, M., & Le lech, M., & Le Guyader, F. S. (2007). Evaluation of removal of Guyader, F. S. (2005). Real-time RT-PCR for norovirus screening noroviruses during wastewater treatment, using real-time reverse in shellfish. Journal of Virological Methods, 123, 1–7. transcription-PCR: Different behaviors of genogroups I and II. Lowther, J. A., Avant, J. M., Gizynski, K., Rangdale, R. E., & Lees, Applied and Environment Microbiology, 73, 7891–7897. D. N. (2010). Comparison between quantitative real-time reverse Dore, B., Keaveney, S., Flannery, J., & Rajko-Nenow, P. (2010). Man- transcription PCR results for norovirus in oysters and self-reported agement of health risks associated with oysters harvested from gastroenteric illness in restaurant customers. Journal of Food Pro- a norovirus contaminated area, Ireland, February–March 2010. tection, 73, 305–311. Eurosurveillance Weekly, 15, 19567. Lowther, J. A., Gustar, N. E., Hartnell, R. E., & Lees, D. N. (2012a). Doré, W. J., & Lees, D. N. (1995). Behavior of Escherichia coli and Comparison of norovirus RNA levels in outbreak-related oysters male-specific bacteriophage in environmentally contaminated with background environmental levels. Journal of Food Protec- bivalve molluscs before and after depuration. Applied and Envi- tion, 75, 389–393. ronment Microbiology, 61, 2830–2834. Lowther, J. A., Gustar, N. E., Powell, A. L., Hartnell, R. E., & Lees, D. EFSA. (2012). Technical specifications for a European baseline survey N. (2012b). Two-year systematic study to assess norovirus con- of norovirus in oysters. EFSA Journal, 14, 4414. tamination in oysters from commercial harvesting areas in the EFSA Panel on Biological Hazards. (2012). Norovirus (NoV) in oys- United Kingdom. Applied and Environment Microbiology, 78, ters: methods, limits and control options. EFSA Journal, 10, 2500. 5812–5817. Flannery, J., Keaveney, S., & Doré, W. (2009). Use of FRNA bacte- Polo, D., Varela, M. F., & Romalde, J. L. (2015). Detection and quan- riophages to indicate the risk of norovirus contamination in Irish tification of hepatitis A virus and norovirus in Spanish authorized oysters. Journal of Food Protection, 72, 2358–2362. shellfish harvesting areas. International Journal of Food Micro- Gosling, E. (2008). Bivalve molluscs: Biology, ecology and culture. biology, 193, 43–50. Hoboken, NJ: Wiley-Blackwell. Schaeffer, J., Le Saux, J.-C., Lora, M., Atmar, R. L., & Le Guyader, F. Hoehne, M., & Schreier, E. (2006). Detection of Norovirus genogroup S. (2013). Norovirus contamination on French marketed oysters. I and II by multiplex real-time RT-PCR using a 3′-minor groove International Journal of Food Microbiology, 166, 244–248. binder-DNA probe. BMC Infectious Diseases, 6, 69. Suffredini, E., Lanni, L., Arcangeli, G., Pepe, T., Mazzette, R., Cic- Kageyama, T., Kojima, S., Shinohara, M., Uchida, K., Fukushi, S., caglioni, G., et al. (2014). Qualitative and quantitative assess- Hoshino, F., et al. (2003). Broadly reactive and highly sensitive ment of viral contamination in bivalve molluscs harvested in Italy. assay for Norwalk-like viruses based on real-time quantitative International Journal of Food Microbiology, 184, 21–26. reverse transcription-PCR. Journal of Clinical Microbiology, 41, Svraka, S., Duizer, E., Vennema, H., de Bruin, E., van der Veer, B., 1548–1557. Dorresteijn, B., et al. (2007). Etiological role of viruses in out- Kittigul, L., Thamjaroen, A., Chiawchan, S., Chavalitshewinkoon- breaks of acute gastroenteritis in The Netherlands from 1994 Petmitr, P., Pombubpa, K., & Diraphat, P. (2016). Prevalence and through 2005. Journal of Clinical Microbiology, 45, 1389–1394. molecular genotyping of noroviruses in market oysters, mussels, van Beek, J., Ambert-Balay, K., Botteldoorn, N., Eden, J. S., Fonager, and cockles in Bangkok, Thailand. Food and Environmental Virol- J., Hewitt, J., et al. (2013). Indications for worldwide increased ogy, 8, 133–140. norovirus activity associated with emergence of a new variant of Le Guyader, F. S., Le Saux, J. C., Ambert-Balay, K., Krol, J., Serais, genotype II.4, late 2012. Eurosurveillance Weekly, 18, 8–9. O., Parnaudeau, S., et al. (2008). Aichi virus, norovirus, astrovi- Vega, E., Barclay, L., Gregoricus, N., Williams, K., Lee, D., & Vinjé, rus, enterovirus, and rotavirus involved in clinical cases from a J. (2011). Novel surveillance network for norovirus gastroenteri- French oyster-related gastroenteritis outbreak. Journal of Clinical tis outbreaks, United States. Emerging Infectious Diseases, 17, Microbiology, 46, 4011–4017. 1389–1395. Lees, D. (2000). Viruses and bivalve shellfish. International Journal of Food Microbiology, 59, 81–116. 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Food and Environmental Virology Springer Journals

A One-Year Survey of Norovirus in UK Oysters Collected at the Point of Sale

Loading next page...
 
/lp/springer-journals/a-one-year-survey-of-norovirus-in-uk-oysters-collected-at-the-point-of-QQtPrHsGwD
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Biomedicine; Virology; Food Science; Chemistry/Food Science, general
ISSN
1867-0334
eISSN
1867-0342
DOI
10.1007/s12560-018-9338-4
Publisher site
See Article on Publisher Site

Abstract

Contamination of bivalve shellfish, particularly oysters, with norovirus is recognised as a food safety risk and a potential contributor to the overall burden of gastroenteritis in the community. The United Kingdom (UK) has comprehensive national baseline data on the prevalence, levels, and seasonality of norovirus in oysters in production areas resulting from a previ- ous two-year study (2009–2011). However, previously, data on final product as sold to the consumer have been lacking. As part of a wider project to establish the overall burden of foodborne norovirus in the UK, this study aimed to address this data gap. A one-year survey of oysters collected from the point-of-sale to the consumer was carried out from March 2015 to March 2016. A total of 630 samples, originating in five different European Union Member States, were collected from 21 regions across the UK using a randomised sampling plan, and tested for norovirus using a method compliant with ISO 15216-1, in addition to Escherichia coli as the statutory indicator of hygiene status. As in the previous production area study, norovirus RNA was detected in a high proportion of samples (68.7%), with a strong winter seasonality noted. Some statistically significant differences in prevalences and levels in oysters from different countries were noted, with samples originating in the Netherlands showing lower prevalences and levels than those from either the UK or Ireland. Overall, levels detected in positive samples were considerably lower than seen previously. Investigation of potential contributing factors to this pattern of results was carried out. Application of normalisation factors to the data from the two studies based on both the numbers of norovirus illness reports received by national surveillance systems, and the national average environmental temperatures during the two study periods resulted in a much closer agreement between the two data sets, with the notably different numbers of illness reports making the major contribution to the differences observed in norovirus levels in oysters. The large majority of samples (76.5%) contained no detectable E. coli; however, in a small number of samples (2.4%) levels above the statutory end product standard (230 MPN/100 g) were detected. This study both revealed the high prevalence of norovirus RNA in oysters directly available to the UK consumer, despite the high level of compliance with the existing E. coli-based health standards, while also highlighting the difficulty in comparing the results of surveys carried out in different time periods, due to variability in risk factors. Keywords Norovirus · Oysters · qRT-PCR · Survey Introduction Contamination of bivalve shellfish, particularly oysters, with norovirus is recognised as a food safety risk, with a Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s1256 0-018-9338-4) contains considerable number of reports of outbreaks in the litera- supplementary material, which is available to authorized users. ture (reviewed in Bellou et al. 2013). In both the European Union and the United States, viral contamination in shellfish * J. A. Lowther is regulated indirectly using enteric bacteria as an indicator james.lowther@cefas.co.uk of faecal pollution (Anonymous 2004, 2015b). However, this Centre for Environment, Fisheries and Aquaculture Science, approach has been demonstrated to inadequately address Weymouth DT4 8UB, UK the risk from human enteric viruses, with a number of viral Institute of Psychology, Health & Society, University outbreaks caused by batches compliant with the regulations of Liverpool, Liverpool, England, UK Vol:.(1234567890) 1 3 Food and Environmental Virology (2018) 10:278–287 279 (Chalmers and McMillan 1995; Le Guyader et al. 2008; wholesalers. A total of 373 vendors were identified across Dore et al. 2010). Considerable progress has been made the 21 areas. towards the development of detection methods for norovirus A randomised sampling plan was drawn up aiming to in molluscan shellfish and an ISO/CEN technical specifica- obtain a total of 630 oyster samples over a 1-year period tion including such a method (ISO/TS 15216) was published (16th March 2015–15th March 2016), with monthly tar- in 2013 (Anonymous 2013), with a subsequent update to gets of 26 samples in the truncated months of March 2015 a fully validated standard in 2017 (Anonymous 2017). EU and March 2016, and 53 or 52 samples alternatively for the legislative texts foreshadow the adoption of virus controls months April 2015 to February 2016. For each month, the in bivalve shellfish when the methods are sufficiently devel- vendors targeted were selected randomly from a subset of oped (Anonymous 2005) and the options for improvement the list of all 373 vendors with no weighting by region. Over of EU legislation to better address the virus risk have been the course of the survey, any shortfall in sample numbers actively discussed in recent years (EFSA Panel on Biologi- collected in a given month was compensated by the addition cal Hazards 2012). It has therefore become important to of extra samples (selected at random from the same region gain information about the application of the new methods, for logistical regions) in the sampling schedule for the fol- and the potential impact of possible legislative standards on lowing month. bivalve shellfish production. Amongst current European Union Member States, the Sample Collection United Kingdom (UK) has some of the most comprehen- sive national baseline data on the prevalence, levels and sea- Samples (except online sales) were pre-ordered through sonality of norovirus in oysters resulting from a two-year direct contact with the vendor, then collected by sam- study carried out on samples taken directly from production pling officers at the point-of-sale to the consumer. Samples areas in 2009–2011. (Lowther et al. 2012b). However, until sourced from online sales were ordered for delivery to the recently, data on final product as sold to the consumer in the sampling officer. Within each vendor, samples were lim- UK have been lacking, with to the best of our knowledge ited to native, Pacific, or other oyster species, sold as either the only published study having tested oysters from a single ambient, chilled, or frozen. To avoid possible contamination dispatch centre only (Lowther et al. 2010). As part of a wider by food handlers live animals in restaurants were obtained project to establish the overall burden of foodborne norovi- before shucking by restaurant staff. Cooked, pasteurised, rus in the UK (“NoVAS: Assessing the contribution made smoked, or otherwise processed oysters were not sampled. by the food chain to the burden of UK-acquired norovirus Where multiple products or batches of the same product infection”—UK Food Standards Agency Project reference: were available, one was picked at random by the sampler. A FS101040), this study aimed to address this data gap, as well sample consisted of individual animals from the same batch as to compare results for final product with those obtained (same origin and production date). in the previous production area survey. Given sufficient availability samples consisted of 25 oys- ters (with a minimum number of 12 oysters required for a valid sample). At the point of sampling, full sample details Materials and Methods including date, time, vendor name and address, sample type, sample temperature at the point of sale (ambient, fresh, fro- Sampling Plan zen), sample origin/identification mark (if available) were recorded by the sampling officer. A high-resolution digital The survey design was informed by a comprehensive prac- photograph of the sample packaging and identification mark tical evaluation of the purchase routes for oysters available (if available) was taken. This information with accompany- to the UK consumer. This evaluation was undertaken by a ing photographs was then e-mailed to the Stericycle project specialist product retrieval company contracted to collect the co-ordinator for collation in a sample database. survey samples (Stericycle ExpertSOLUTIONS, Reading, Samples were packaged in temperature controlled Cole- UK), through phone interviews with and physical visits to man food boxes with cool packs and despatched to the labo- identified vendors. This market research was conducted in ratory via overnight courier service accompanied by a sam- 21 selected cities/regions across the UK (selected to give a ple submission form including a unique sample identifier, good geographical spread and including regions in each of the date and time of collection, the storage temperature of the four constituent countries of England, Wales, Scotland the sample at the collection point and the date and time of and Northern Ireland). Vendors directly available to con- despatch. Details of the vendor and the origin of the oysters sumers of oysters were subdivided into the following types: were not included such that the sample testing was carried supermarkets, fishmongers, restaurants, online sales and out blind. 1 3 280 Food and Environmental Virology (2018) 10:278–287 Upon receipt at the laboratory, each sample was pro- COG2R primers, and QNIFS probe were used (Kageyama cessed according to standard procedures. If the sample tem- et al. 2003, Loisy et al. 2005). For mengo virus, the mengo perature on receipt was > 18 °C, fewer than 10 live animals 110 and mengo 209 primers, and the mengo 147 probe were were available, or the condition of the sample was otherwise used (Costafreda et al. 2006). For both norovirus genogroup- unsatisfactory, samples were not tested, and replacement specific assays, three aliquots of 5 μl sample or extraction samples were collected. In addition, if the sample tempera- control RNA was tested in 25 µl total volume with one-step ture on receipt was > 10 °C, fewer than 20 live animals reaction mix prepared using the RNA Ultrasense one-step were available, or a period of > 48 h had elapsed between qRT-PCR system (Invitrogen) (final concentrations of 1x sample collection and receipt at the laboratory, samples were Reaction Mix, 500 nM forward and 900 nM reverse primers, analysed for norovirus only (not E. coli); under these circum- and 250 nM probe, plus 0.5 µl Rox and 1.25 µl Enzyme Mix stances replacement samples were not sought. per reaction). For mengo virus, two aliquots of 5 μl cDNA were used. Amplification was performed using the following Detection and Quantification of Norovirus cycling parameters: 55 °C for 60 min, 95 °C for 5 min, and then 45 cycles of 95 °C for 15 s, 60 °C for 1 min and 65 °C Oyster samples were tested for norovirus according to the for 1 min on an Mx3005P real-time PCR machine (Strata- draft international standard ISO 15216-1 (now published in gene). Wells containing nuclease free water and the above Anonymous 2017). qRT-PCR reaction mixes were included on each plate as a negative control. Quantification used a log dilution series 5 1 Virus Extraction (range 1 × 10 to 1 × 10  copies/µl) of linear dsDNA mol- ecules carrying the GI and GII target sequences and fol- For each sample, ten oysters were selected. The digestive lowed the principles outlined in ISO 15216-1 (Anonymous tissues (stomach and digestive diverticula) of these oysters 2017). All samples were assessed for extraction efficiency by were excised, pooled, and then finely chopped using a razor the comparison of sample Ct values for mengo virus with a blade. A 2-g subsample of chopped digestive tissues was standard curve generated from the process control material transferred to a clean tube. 10 µl of mengo virus vMC0 tis- and for qRT-PCR inhibition using RNA external controls as sue culture supernatant was added to the 2-g subsample described in ISO 15216-1 (Anonymous 2017). Samples were as a within-sample virus/RNA extraction process control. retested if extraction or qRT-PCR inhibition levels fell below Homogenates were prepared by adding 2 ml of a 100 μg/ml 1% or above 75% respectively, where positive qRT-PCR con- Proteinase K solution to the digestive tissues. This was then trols indicated reagent failure, or for any positive sample incubated at 37 °C with shaking at 320 rpm for a duration where the negative extraction or PCR controls showed con- of 1 h, and subsequently incubated at 60 °C for a duration of tamination. Quantitative results were not adjusted for losses 15 min. Finally, the sample was centrifuged at 3000×g for during processing or RT-PCR inhibition. 5 min.; the volume of the soluble portion (homogenate) was measured and then retained for downstream testing and the Detection and quantification of E. coli pellet discarded. Homogenates were stored at 4 °C for up to one month prior to testing. Oyster samples were tested for E. coli according to ISO 16649-3 (Anonymous 2015a). Whole animal homogenates RNA Extraction were prepared from the flesh and intravalvular fluid of 10 oysters and assayed using a most-probable-number (MPN) Total RNA was extracted from 500 µl of shellfish homogen- method. Results are expressed per 100 g of shellfish flesh ate using a NucliSENS miniMAG extraction machine and and intravalvular fluid. NucliSENS magnetic extraction reagents (BioMerieux) following the manufacturer’s instructions (eluting in 100 µl Statistical Analysis elution buffer). A negative (water only) extraction control sample was also prepared and tested in parallel with each Relevant statistical analyses (Fisher’s exact test, set of samples extracted. Eluted RNA was stored at − 20 °C Kruskal–Wallis test) were carried out using the Minitab until required. software package. For statistical analysis and calculation of geometric means, positive results of < 100 copies/g (the One‑Step qRT‑PCR limit of quantification of the assay) were scored at 50, and not detected samples were scored at 20 (half the limit of For norovirus GI, QNIF4 and NV1LCR primers, and TM9 detection). Scores for GI and GII were combined prior to probe were used (da Silva et al. 2007; Hoehne and Schreier analysis. In this way, samples that were not detected for both 2006; Svraka et al. 2007). For norovirus GII, QNIF2 and genogroups scored 40 copies/g, and this figure should be 1 3 Food and Environmental Virology (2018) 10:278–287 281 considered a baseline for levels. Confidence intervals (95%) both surveys, an average level for each calendar month was for datasets were calculated as the geometric mean ± 2x the calculated. geometric standard deviation; at the lower end, these are censored at 40  copies/g where the calculated value was less than this. Due to the large number of censored values Results and Discussion in the dataset, non-parametric statistical tests were used throughout. Norovirus Results All 630 samples subjected to testing passed quality control Normalisation Factors criteria for extraction efficiency and RT-PCR inhibition on initial or retesting. The average extraction efficiency obtained In order to compare the contribution of different risk fac- was 28.7% (range 1.1–99.6%), while the average RT-PCR tors to the results obtained in the current and previous stud- inhibition was 14.0% (range 0–74.3%). Of the 630 samples, ies, month-by-month normalisation factors were generated 433 (68.7%) were positive for norovirus RNA. Of these, 99 for norovirus illness (using data on illness reports in Eng- samples (15.7%) were positive for GI only, 88 (14.0%) were land and Wales provided by Public Health England) and positive for GII only and 246 (39.0%) were positive for both environmental temperatures (using data on UK average air GI and GII. A clear seasonality was observed with 79.7% temperatures obtained from the UK Meteorological Office of samples collected in the months October–March positive website—http://www.met office.gov.uk) as follows. compared with 57.0% in the months April–September. This For illness reports, the normalisation factor N was deter- difference was found to be statistically significant (Fisher’s mined as exact test; p < 0.0001). The highest and lowest monthly prevalences were recorded in February 2016 (96.3%) and N = , September 2015 (34.6%), respectively (Fig. 1a). where I is the average illness reports per day for the relevant calendar month in the period of the production area survey (May 2009–Apr 2011) and I average illness reports per day for the month in question, such that where illness reports for a given month were lower than the average for that calen- dar month in 2009–2011, the normalisation factor was > 1. For example, in April 2015, the average number of illness reports per day was 30.27, compared with the average for April during the production area survey of 34.05 reports per day. The normalisation factor for April 2015 was therefore 34.05 ÷ 30.27 = 1.12. For temperatures, the normalisation factor N was deter- mined as 20 − T N = , 20 − T where T is the long-term time series average temperature for the relevant calendar month (1981–2010) and T  is the recorded monthly UK average temperature for the month in question, such that where the UK average air temperature for a given month was higher than the long-term average for that calendar month in 1981–2010, the normalisation Fig. 1 Monthly proportion of samples giving total norovirus results factor was > 1. For example, in April 2015, the UK average in different quantity brackets (copies/g) in the current (retail) sur - temperature was 7.9 °C, compared with a long-term average vey and a previous production area survey. ND not detected. Results for April of 7.4 °C. The normalisation factor for April 2015 are for GI and GII combined; samples that were positive at levels of  <  100 copies/g for both genogroups are included in the  <  100 was therefore (20 − 7.4) ÷ (20 − 7.9) = 1.04. quantity bracket. a Results for the retail survey. b Results for the pro- Normalisation factors calculated in this way were applied duction area survey (Lowther et  al. 2012b)—proportions calculated to the geometric mean norovirus levels recorded for each for each calendar month across the survey duration, March shown month of both the retail and production area surveys. For twice to allow comparison with the retail survey 1 3 282 Food and Environmental Virology (2018) 10:278–287 Norovirus levels were also higher during the winter interval 40–129  copies/g) than in samples from the UK period with a geometric mean level of 87 copies/g (95% (71.7% positive, geometric mean of 78 copies/g, 95% con- confidence interval 40–309 copies/g) in the months Octo- fidence interval 40–277 copies/g). These differences were ber–March compared with 65 copies/g (95% confidence found to be statistically significant (Fisher’s exact test, interval 40–202 copies/g) in samples collected from April to p = 0.0144; Kruskal–Wallis test, p < 0.001). Further sub- September. This difference was found to be statistically sig- division of non-UK samples to enable country-by-country nificant using the Kruskal–Wallis test (p  < 0.001). The high- analysis showed that for oysters from the Netherlands both est levels recorded in individual samples were 586 copies/g prevalence and levels were significantly lower than for for GI and 1802 copies/g for GII; however, in the majority the UK (Fisher’s exact test, p < 0.0001; Kruskal–Wallis of samples testing positive (85.9%), the levels recorded were test, p < 0.0001). Prevalence and levels for oysters from below the limit of quantification of the assay (100 copies/g) the Republic of Ireland were not significantly different for both norovirus GI and GII. In total, 61 samples pro- from those for the UK, but were significantly higher than duced results of > 100 copies/g for one or both genogroups, those for the Netherlands (Fisher’s exact test, p = 0.0001; representing 14.1% of positive samples, and 9.7% of total Kruskal–Wallis test, p = 0.0081). No apparent seasonal bias samples. Of these 61 samples, 7 produced results of > 100 in collection dates for samples from the three countries were copies/g for both genogroups, 2 for GI only and 52 for GII found to explain these differences (no significant difference only. The highest monthly incidence of samples giving was found between the proportions of samples collected dur- results > 100 copies/g was March 2015 (37.5%). Over the ing the winter months October–March using Fisher’s exact course of the survey, 5 samples (0.8% of total samples) pro- test). Statistical analysis of norovirus results for samples duced results for GI and GII combined of > 1000 copies/g; from France and Spain was not carried out due to the small three of these samples were collected in September 2015, number of samples. and 2 in February 2016. Comparison with the Production Area Study Comparison of Oysters Originating in Different Countries The prevalence of norovirus RNA in oyster samples recorded in this survey (the “retail survey”; 68.7%) was similar but For 492 samples (78.1% of the total), the dispatch centre slightly lower than that found in a previous two-year sur- from which the oysters originated could be identified as a vey (2009–2011) of oysters from UK production areas (the result of information collected by the sampling officer (this “production area survey” 76.2%) (Lowther et al. 2012b). In identification was supported by a photograph of the identi- addition, a similar seasonality with increased prevalences fication mark or other identifying labels/packaging in 378 and levels in the winter months was noted in both surveys. cases). Oysters originated from 33 die ff rent dispatch centres However, the overall levels of norovirus recorded in the in 5 different EU Member States. Of the 492 samples with retail survey were considerably lower than in the production identified dispatch centres, 434 samples (88.2%) originated area survey. In the latter, 36.5% of total samples contained in the UK, 29 (5.9%) from the Netherlands, 25 (5.1%) from levels  >  100  copies/g (the limit of quantification of the Ireland, 3 (0.6%) from France and 1 (0.2%) from Spain. assay) for one or both norovirus genogroups, combined lev- Prevalences of norovirus detection and geometric mean els of > 1000 copies/g were found in 14.6% of samples, and levels of norovirus for samples originating in different EU combined levels of > 10,000 copies/g were found in 1.1% of Member States are shown in Table 1. samples. Geometric means for all results were 76 copies/g Overall prevalence and levels of norovirus were lower (95% confidence interval 40–261 copies/g) and 159 copies/g in samples originating outside the UK (55.2% of samples (95% confidence interval 40–2964 copies/g) for the retail positive, geometric mean of 58 copies/g, 95% confidence and production area surveys, respectively. This difference Table 1 Norovirus results by Country of origin Number of Norovirus results country of origin samples Prevalence (% posi- Geometric mean (copies/g; 95% tive) (%) confidence interval in parentheses) UK 434 71.7 78 (40–277) Netherlands 29 31.0 49 (40–91) Ireland 25 84.0 69 (40–120) France 3 33.3 48 (40–92) Spain 1 100.0 275 (n/a) 1 3 Food and Environmental Virology (2018) 10:278–287 283 was found to be statistically significant (Kruskal–Wallis test; normalisation factors were determined using the PHE data p < 0.001). on illness reports in England and Wales (treating these fig- Possible underlying causes for this pattern of results ures as a proxy for community levels as a whole) and Met include:- Office data on UK national average monthly air temperatures (treating these figures as a proxy for environmental tem- (1) Risk reduction measures by Food Business Operators peratures as a whole—equivalent national average seawater including e.g. use of enhanced depuration conditions or temperatures are not available) as described in materials and use of norovirus testing to inform decisions on choice methods. of supply for processing and marketing Application of the normalisation factors based on illness (2) Representativeness of samples: It is possible that the reports resulted in a notable improvement in correspond- production area survey was not representative of the ence in results by calendar month between the two surveys volumes of oysters placed on the UK market as the (see Fig. 2). Geometric mean levels for each month in the selection of sites for the production area survey was two surveys are plotted against each other in Fig. 2b, d, f, h meant to provide a representative selection of pro- alongside lines of best fit and equality; for data normalised duction areas with different risk profiles and a good according to illness reports (Fig. 2c, d), the slope of the line geographical spread, but not to represent production of best fit (0.4723) is considerably closer to equality and volumes or market share. the correlation is considerably closer to total (r  = 0.9506) (3) Variation in norovirus shedding rates in the community: than for non-normalised data (Fig. 2a, b; slope = 0.0887 and Sewage treatment is known to only reduce norovirus r  = 0.5384). by a limited extent (Campos and Lees 2014). Conse- Application of the normalisation factors for temperature quently, a key factor influencing norovirus contami- in isolation yielded only a modest improvement in agree- nation in filter-feeding shellfish impacted by sewage ment between the results of the two studies (Fig.  2e, f). discharges will be the degree of virus infection, and However, application of both the illness and temperature- hence the degree of virus shedding in faeces, in the based normalisation factors in combination produced the population contributing to the sewage inputs. Dur- best line of best fit overall (Fig.  2g, h; slope = 0.5626 and ing this study, unusually low levels of norovirus were r  = 0.9576). observed in the community in England and Wales dur- This analysis indicates that much of the difference in the ing the winter of 2015–2016, particularly during the norovirus levels between the retail and production area sur- months November to January, compared with unusu- veys can be attributed to the different levels of norovirus in ally high levels during the winters of 2009–2010 and the community between the two study periods, with some 2010–2011 (Supplementary Figure S1, data provided portion of the remaining difference explained by the differ - by Public Health England, equivalent data for other ing temperatures, particularly during the early part of winter. parts of the UK are not available). Nevertheless, even normalising using these factors together (4) Variation in environmental temperatures: Shellfish are results in levels in the retail study on average ~ 56% as high poikilothermic (Gosling 2008), and their metabolic as during the production area survey, suggesting other fac- rate, and hence the degree of contaminant uptake and tors as discussed above also contributed to the different pat- removal, is significantly influenced by the temperature tern of results. of their environment. In this study, environmental tem- peratures in the UK were unusually high during the E. coli Results winter of 2015–2016, particularly during the months November to January, compared with unusually low Out of 630 samples received, E. coli analysis was carried temperatures during the production area study winters out in 452 cases (71.7%). For the remaining samples, E. of 2009–2010 and 2010–2011 (Supplementary Figure coli testing was not carried out primarily due to insufficient S2; data obtained from the UK Meteorological Office live animals in the sample to conduct this test in addition website—http://www.met office.gov.uk). to norovirus analysis (< 20), or elevated temperatures on arrival (> 10 °C). Of the above factors potentially influencing the variation Of the samples tested for E. coli, the bacterium was not seen between contamination levels in the production area detected (< 18 MPN/100 g) in 346 cases (76.5%). In 11 study and in this study, it was only possible to perform fur- samples (2.4%), levels in excess of the EU legal end product ther analysis on the impact of general population shedding standard (230 MPN/100 g; Anonymous 2005) were detected. rates and environmental temperatures due to the unavail- In these cases, the UK Food Standards Agency as the Com- ability of data relevant to the other factors. To further inves- petent Authority was informed on the same working day tigate these possible contributing elements, month-by-month that the result became available. All 11 of these samples 1 3 284 Food and Environmental Virology (2018) 10:278–287 Fig. 2 Application of normalisation factors to monthly geometric and best fit (dotted and labelled with associated equation and r val- mean norovirus levels obtained during the retail and production area ues) are shown. a, b No normalisation applied. c, d Normalisation surveys. a, c, e, f comparison of monthly geomean levels for the retail factors derived from illness reports applied. e, f Normalisation factors (dashed lines) and production area (dotted lines) surveys. b, d, f, h; derived from average temperatures applied. g, h Normalisation fac- correlation between geometric mean norovirus levels for each calen- tors derived from illness reports and average temperatures applied dar month obtained during the two surveys. Lines of equality (solid) 1 3 Food and Environmental Virology (2018) 10:278–287 285 were collected between March and September 2015, with the human health risks is complex (EFSA Panel on Biological highest monthly incidence of five samples > 230 MPN/100 g Hazards 2012); however in an analysis of outbreak-related in July 2015, representing 15.2% of the samples collected in oyster samples carried out in this laboratory (Lowther et al. that month. In one sample, a level in excess of the upper limit 2012a), an association between increased norovirus lev- of quantification of the E. coli assay (> 18,000 MPN/100 g) els and increased likelihood of norovirus-type illness was was recorded from a sample collected on 15 July 2015. observed, with no outbreak-related sample recording levels In comparison with the production area survey, levels of below 152 copies/g. The human health consequences of the E. coli recorded in this study were very low. No E. coli was large proportion of positive samples in this survey are there- detected in the majority of the samples, while results over fore not certain. the A classification and end product standard were rare. The majority of oyster samples tested originated from In the production area survey by contrast, E. coli propor- dispatch centres in the UK (88.2% of samples where the tions were 14.3% undetected and 40.0% > 230 MPN/100 g. dispatch centre could be identified), with the remainder Although other factors may have contributed, this difference originating in other countries in Western Europe. System- is likely to be largely the result of the well-established high atic comparison of prevalences and levels of norovirus in efficacy of standard depuration conditions for the removal of oysters from different countries was complicated by the low E. coli bacteria (Doré and Lees 1995). Since the removal of numbers of samples from each exporter country; however E. coli is a good proxy for other bacterial pathogens derived oysters from the Netherlands showed significantly lower from sewage contamination (Lees 2000), this demonstrates levels and prevalences than oysters from both the UK and the contribution to public health of the classification and Ireland. There is some evidence that oyster growing waters depuration regulations for protection from bacterial illness. in the Netherlands are impacted by lower levels of faecal This finding is supported by the low numbers of bacterial pollution; six out of seven (86%) oyster production areas in infections associated with consumption of oysters in the UK the country are at the time of writing classified A (Nether - (Lees 2000). lands National Reference Laboratory for monitoring bacte- The small number of results of > 230 MPN/100 g, includ- riological and viral contamination of bivalve molluscs; per- ing one result of > 18,000 MPN/100 g, indicates that despite sonal communication), the cleanest status based on E. coli the high level of adherence to the legal standards, problems monitoring results according to EU legislation (Anonymous can nevertheless occur. The root cause of the high E. coli 2004). By contrast, in the UK, 37% of oyster production levels detected in some samples could not be investigated, areas are wholly or partially classified A, either permanently but could conceivably be linked to problems post-harvest, or for part of the year (UK National Reference Laboratory during transportation, or at the point-of-sale. for monitoring bacteriological and viral contamination of bivalve molluscs; personal communication). During the year of the survey, some significant potential Conclusion risk factors were low compared with the previous study on oysters from UK production areas (Lowther et al. 2012b). The survey described here is the first systematic study of The number of norovirus cases in the general population norovirus in oysters collected at the point-of-sale in the UK. and hence the likely extent of virus shedding into shellfish Norovirus RNA was detected in 68.7% of samples tested, production areas was considerably lower than previously, comparable with the prevalence found in a previous survey and environmental temperatures during the winter were carried out using the same methods on oysters from UK higher. The datasets used to quantify these risk factors had production areas (76.2%; Lowther et al. 2012b). The preva- some limitations; air temperatures were used as an indicator lence described here is considerably higher than recorded of overall environmental temperatures, rather than directly in surveys of norovirus in bivalve shellfish collected at using seawater temperatures (no national average seawater the point-of-sale in some other countries, for example the temperature data is available). In addition, for both factors, United States (3.9%; Woods and Burkhardt 2010), France data collected in the UK were extrapolated to normalise (9%; Schaeffer et al. 2013) and Thailand (12.3%: Kittigul results based on all samples collected during the retail sur- et al. 2016); however, comparatively frequent detection of vey, including those originating outside the UK. However, norovirus has been reported in shellfish from production 430 out of 432 samples (99.6%) where origin data existed areas in Ireland (37.1%; Flannery et al. 2009), Italy (51.5%, either originated in the UK, or in bodies of water abutting Suffredini et al. 2014) and Spain (52.4%; Polo et al. 2015). UK territorial waters (the Irish Sea, the English Channel and Although the majority of samples were found to be positive, the North Sea), while the illness data used broadly reflects levels exceeding 100 norovirus copies/g were found in only global trends in norovirus infections. The two winter peri- a relatively small percentage of samples (9.7%). The rela- ods in which illness levels in the dataset used were highest tionship between levels of norovirus as detected by PCR and (2009–2010 and 2012–2013) both followed directly on from 1 3 286 Food and Environmental Virology (2018) 10:278–287 infection (NoVAS)”. The authors thank Laura Boyd, Michelle Paice the emergence of a global pandemic strain; New Orleans and the sampling team from Stericycle ExpertSOLUTIONS for collec- 2009 (Vega et al. 2011) and Sydney 2012 (van Beek et al. tion of oyster samples. We also thank the NoVAS Consortium for help- 2013), respectively. For these reasons, we therefore consid- ful comments on the manuscript. The NoVAS Consortium in addition ered that use of these suboptimal datasets for determination to the authors comprises the University of Liverpool (Miren Iturriza- Gomara), the University of East Anglia (Paul Hunter, Jim Maas), Pub- of normalisation factors was unlikely to confound the analy- lic Health England (David James Allen, Nicola Elviss, Andrew Fox), sis we carried out. Leatherhead Food Research (Angus Knight) and Fera Science Ltd. This analysis offers some insights into the contribution (Nigel Cook, Martin D’Agostino). of these two factors to the pattern of results observed in the different surveys, and highlights the difficulty of comparing Open Access This article is distributed under the terms of the Crea- tive Commons Attribution 4.0 International License (http://creat iveco results from surveys carried out in different time periods, or mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- of treating the results of a short survey as completely indica- tion, and reproduction in any medium, provided you give appropriate tive of the long-term characteristics of the surveyed area. It credit to the original author(s) and the source, provide a link to the is however possible that some of the differences observed Creative Commons license, and indicate if changes were made. were down to inherent differences between production area and retail-ready oysters. For example, Food Business Opera- tor risk management interventions (such as virus testing, or selection of product from cleaner areas) may have contrib- References uted to the low virus levels seen in this study. Direct com- Anonymous. (2004). Regulation (EC) No 854/2004 of the European parison, within the same time period, of levels in production Parliament and of the Council of 29 April 2004 laying down spe- areas with those seen in retail-ready oysters would assist cific rules for the organisation of official controls on products of assessment of the contribution made by producer practices. animal origin intended for human consumption. Official Journal An ongoing EU-wide survey of norovirus in oysters from of the European Communities. 25.06.2004, L226, 83–127. Anonymous. (2005). European Commission Regulation (EC) No both production areas and dispatch centres, organised by the 2073/2005 of 15 November 2005 on microbiological criteria for European Food Safety Authority (EFSA 2012) may help to foodstuffs. Official Journal of the European Union. 22.12.2005, illuminate this issue. L338, 1–26. The very high proportion of samples compliant with the Anonymous. (2013). Microbiology of food and animal feed—horizon- tal method for determination of hepatitis A virus and norovirus E. coli end product standard (97.6%) in this study indicates in food using real-time RT-PCR—Part 1: Method for quantifica- the good compliance with current regulatory requirements tion. ISO/TS 15216-1:2013 (withdrawn). Geneva: International in the UK oyster supply chain and the consequential prob- Organization for Standardization. able low risk from bacterial pathogens such as Salmonella. Anonymous. (2015a). Microbiology of the food chain—Horizon- tal method for the enumeration of beta-glucuronidase-positive However, the contrast between the results for norovirus and Escherichia coli—Part 3: Detection and most probable num- E. coli again illustrates the limitations of this organism as ber technique using 5-bromo-4-chloro-3-indolyl-ß-D-glucu- an indicator of viral risk in shellfish, particularly in depu - ronide. ISO 16649-3. Geneva: International Organization for rated animals. Shellfish-related outbreaks of norovirus have Standardization. Anonymous. (2015b). National Shellfish Sanitation Program. Guide often been linked to batches compliant with the end product for the Control of Molluscan Shellfish, 2015 Revision. https :// standard for E. coli (Chalmers and McMillan 1995; Le Guy- www.fda.gov/food/guida nceregulat ion/f ederals tatef oodpr ogr ams/ ader et al. 2008; Dore et al. 2010). On the other hand, we ucm20 06754 .htm. have previously observed a site-by-site correlation between Anonymous. (2017). Microbiology of the food chain—Horizontal method for determination of hepatitis A virus and norovirus using E. coli and norovirus levels in oysters from UK production real-time RT-PCR—Part 1: Method for quantification. ISO 15216- areas (Lowther et al. 2012b). 1:2017. Geneva: International Organization for Standardization. In summary, although illuminating the difficulties with Bellou, M., Kokkinos, P., & Vantarakis, A. (2013). Shellfish-borne comparing data collected in different periods, this study viral outbreaks: A systematic review. Food Environ. Virol., 5, 13–23. provides a comprehensive and systematic assessment of the Campos, C. J., & Lees, D. N. (2014). Environmental transmission of extent of both norovirus and E. coli contamination in oysters human noroviruses in shellfish waters. Applied and Environment available to the UK consumer. These data can assist risk Microbiology, 80, 3552–3561. managers both in the UK and beyond to manage the health Chalmers, J. W. T., & McMillan, J. H. (1995). An outbreak of viral gastroenteritis associated with adequately prepared oysters. Epi- risks posed by norovirus in oysters, and as part of the wider demiology and Infection, 115, 163–167. NoVAS project will contribute to an assessment of the con- Costafreda, M. I., Bosch, A., & Pintó, R. M. (2006). Development, tribution of the food chain to norovirus infections in the UK. evaluation, and standardization of a real-time TaqMan reverse transcription-PCR assay for quantification of hepatitis A virus in Acknowledgements This work was supported by the United Kingdom clinical and shellfish samples. Applied and Environment Micro- Food Standards Agency project FS101040: “Assessing the contribu- biology, 72, 3846–3855. tion made by the food chain to the burden of UK-acquired norovirus 1 3 Food and Environmental Virology (2018) 10:278–287 287 da Silva, A. K., Le Saux, J. C., Parnaudeau, S., Pommepuy, M., Elime- Loisy, F., Atmar, R. L., Guillon, P., Le Cann, P., Pommepuy, M., & Le lech, M., & Le Guyader, F. S. (2007). Evaluation of removal of Guyader, F. S. (2005). Real-time RT-PCR for norovirus screening noroviruses during wastewater treatment, using real-time reverse in shellfish. Journal of Virological Methods, 123, 1–7. transcription-PCR: Different behaviors of genogroups I and II. Lowther, J. A., Avant, J. M., Gizynski, K., Rangdale, R. E., & Lees, Applied and Environment Microbiology, 73, 7891–7897. D. N. (2010). Comparison between quantitative real-time reverse Dore, B., Keaveney, S., Flannery, J., & Rajko-Nenow, P. (2010). Man- transcription PCR results for norovirus in oysters and self-reported agement of health risks associated with oysters harvested from gastroenteric illness in restaurant customers. Journal of Food Pro- a norovirus contaminated area, Ireland, February–March 2010. tection, 73, 305–311. Eurosurveillance Weekly, 15, 19567. Lowther, J. A., Gustar, N. E., Hartnell, R. E., & Lees, D. N. (2012a). Doré, W. J., & Lees, D. N. (1995). Behavior of Escherichia coli and Comparison of norovirus RNA levels in outbreak-related oysters male-specific bacteriophage in environmentally contaminated with background environmental levels. Journal of Food Protec- bivalve molluscs before and after depuration. Applied and Envi- tion, 75, 389–393. ronment Microbiology, 61, 2830–2834. Lowther, J. A., Gustar, N. E., Powell, A. L., Hartnell, R. E., & Lees, D. EFSA. (2012). Technical specifications for a European baseline survey N. (2012b). Two-year systematic study to assess norovirus con- of norovirus in oysters. EFSA Journal, 14, 4414. tamination in oysters from commercial harvesting areas in the EFSA Panel on Biological Hazards. (2012). Norovirus (NoV) in oys- United Kingdom. Applied and Environment Microbiology, 78, ters: methods, limits and control options. EFSA Journal, 10, 2500. 5812–5817. Flannery, J., Keaveney, S., & Doré, W. (2009). Use of FRNA bacte- Polo, D., Varela, M. F., & Romalde, J. L. (2015). Detection and quan- riophages to indicate the risk of norovirus contamination in Irish tification of hepatitis A virus and norovirus in Spanish authorized oysters. Journal of Food Protection, 72, 2358–2362. shellfish harvesting areas. International Journal of Food Micro- Gosling, E. (2008). Bivalve molluscs: Biology, ecology and culture. biology, 193, 43–50. Hoboken, NJ: Wiley-Blackwell. Schaeffer, J., Le Saux, J.-C., Lora, M., Atmar, R. L., & Le Guyader, F. Hoehne, M., & Schreier, E. (2006). Detection of Norovirus genogroup S. (2013). Norovirus contamination on French marketed oysters. I and II by multiplex real-time RT-PCR using a 3′-minor groove International Journal of Food Microbiology, 166, 244–248. binder-DNA probe. BMC Infectious Diseases, 6, 69. Suffredini, E., Lanni, L., Arcangeli, G., Pepe, T., Mazzette, R., Cic- Kageyama, T., Kojima, S., Shinohara, M., Uchida, K., Fukushi, S., caglioni, G., et al. (2014). Qualitative and quantitative assess- Hoshino, F., et al. (2003). Broadly reactive and highly sensitive ment of viral contamination in bivalve molluscs harvested in Italy. assay for Norwalk-like viruses based on real-time quantitative International Journal of Food Microbiology, 184, 21–26. reverse transcription-PCR. Journal of Clinical Microbiology, 41, Svraka, S., Duizer, E., Vennema, H., de Bruin, E., van der Veer, B., 1548–1557. Dorresteijn, B., et al. (2007). Etiological role of viruses in out- Kittigul, L., Thamjaroen, A., Chiawchan, S., Chavalitshewinkoon- breaks of acute gastroenteritis in The Netherlands from 1994 Petmitr, P., Pombubpa, K., & Diraphat, P. (2016). Prevalence and through 2005. Journal of Clinical Microbiology, 45, 1389–1394. molecular genotyping of noroviruses in market oysters, mussels, van Beek, J., Ambert-Balay, K., Botteldoorn, N., Eden, J. S., Fonager, and cockles in Bangkok, Thailand. Food and Environmental Virol- J., Hewitt, J., et al. (2013). Indications for worldwide increased ogy, 8, 133–140. norovirus activity associated with emergence of a new variant of Le Guyader, F. S., Le Saux, J. C., Ambert-Balay, K., Krol, J., Serais, genotype II.4, late 2012. Eurosurveillance Weekly, 18, 8–9. O., Parnaudeau, S., et al. (2008). Aichi virus, norovirus, astrovi- Vega, E., Barclay, L., Gregoricus, N., Williams, K., Lee, D., & Vinjé, rus, enterovirus, and rotavirus involved in clinical cases from a J. (2011). Novel surveillance network for norovirus gastroenteri- French oyster-related gastroenteritis outbreak. Journal of Clinical tis outbreaks, United States. Emerging Infectious Diseases, 17, Microbiology, 46, 4011–4017. 1389–1395. Lees, D. (2000). Viruses and bivalve shellfish. International Journal of Food Microbiology, 59, 81–116. 1 3

Journal

Food and Environmental VirologySpringer Journals

Published: May 2, 2018

References