In France, medical doctors who diagnose norovirus gastroenteritis in ≥2 persons who shared a common meal are required to declare a suspected foodborne illness outbreak. All meal participants then receive a standardized questionnaire that addresses the foods consumed, the symptoms, and the timing of illness, allowing the calculation of the relative risk and its 95% CI. Information on outbreaks to laboratories must be transmitted quickly to enable collection of samples that are directly linked to the clinical cases.

Eight outbreaks were considered for this study on the basis of the following criteria: clinical diagnosis of norovirus in sick consumers; epidemiologic confirmation that oysters were implicated; and rapid notification of responsible oyster production areas. The outbreaks occurred during the winter months in private houses except for 1 that occurred in a nursing home (outbreak 8) (Table). The attack rate varied from 43% to 100%, with median incubation times between 0.5 and 2 days. Fecal samples available for 2 outbreaks (1 for outbreak 6 and 3 for outbreak 8) confirmed the presence of norovirus (National Reference Center for Enteric Viruses, Dijon, France, pers. comm.). Eight shellfish samples were collected from batches that were directly implicated, and 1 sample was taken from leftovers in the nursing home’s refrigerator, increasing the likelihood that the samples were representative of consumed oysters. An additional 16 samples were collected from implicated production areas located along different coasts of France, including frozen samples (during the winter and spring months, Ifremer laboratories doing official control monitoring of shellfish for Escherichia coli routinely freeze leftover samples). Viruses were eluted from oyster digestive tissues by using the reference method (6), and then quantified using the QuantStudio 3D Digital PCR system (Thermo Fisher, Villebon, France) (Technical Appendix).

No norovirus was detected in 1 oyster sample; norovirus genogroups GI, GII, or both were detected in 9, 11, and 4 samples, respectively (Table). Overall, norovirus concentrations ranged from 43 to 1,170 RNA copies/oyster; the highest concentrations detected were GI. For outbreak 8, in which a leftover sample from the implicated meal was obtained, norovirus GII was detected at a concentration of 82 RNA copies/oyster, whereas norovirus GI was detected at a concentration of 185 RNA copies/oyster in the same batch collected from the oyster farm.

In a previous dose–response model for norovirus GI and GII based on outbreak investigations, differences were observed between consumers with the secretor phenotype, for which infection and disease probability were high at low doses compared with nonsecretor phenotypes (7). Although method sensitivity may need to be improved, the concentrations reported here are consistent with observed illness in dose–response studies to date (8). Norovirus GI and GII were detected in oyster samples from the production area and in 4 fecal samples (National Reference Center for Enteric Viruses, pers. comm.).

Because oyster contamination occurs through the filtration of seawater contaminated by human sewage, many contamination events involving both norovirus genogroups and different strains have been described worldwide; this study provides additional evidence of the diversity of contamination (1). In contrast to person-to-person transmission in which GII strains dominate, oysters favor the transmission of some specific GI strains, a major consideration for the global epidemiology of norovirus (1,3). Thus, identifying if oysters implicated in outbreaks are contaminated with norovirus GI or GII is important, because genetic susceptibility means that some consumers do not become infected with certain GI or GII strains; this affects the disease and favors the distribution of some norovirus strains. Such a comprehensive approach will provide information for risk analysis and assist in understanding norovirus infections (7,9).

Although we obtained some norovirus sequences from 6 implicated batches, confirming the specificity of the dPCR, we believe that the development of technology such as next-generation sequencing will provide more detailed information on the full range of strains present in samples. Obtaining more accurate information on strain diversity and quantification will be valuable for molecular epidemiology studies and management.

In France, oysters are a popular dish, especially during December–April, when they are in the optimal low-fat condition for consumption. They are opened just before consumption and eaten raw; intravalvular seawater is tipped out, thus eliminating food handler contamination. Because this is the highest period for potential contamination by norovirus, samples are kept frozen by laboratories in France for analysis in case of outbreaks. In the current case, this was useful because it demonstrated the presence of norovirus up to 19 days before the shellfish were marketed. This detection in samples collected 2 weeks before an outbreak suggests that illness could have been prevented. Control shellfish samples from different production areas were analyzed at the same time and were negative for norovirus (data not shown), correlating well with the estimated NoV prevalence of less than 10% in France (10).

This study demonstrates that outbreaks could be prevented by performing shellfish analysis at times of the year at which norovirus risk is elevated, such as the winter season, and following microbial alert events such as sewage overflows and heavy rainfall. Application of dPCR to shellfish implicated in outbreaks will provide accurate quantification, which is useful for further risk analysis studies. This application will help to improve regulations and enhance the safety of products on the market, keeping in mind that the sanitary quality of coastal areas is of primary concern.