Norovirus gastroenteritis general outbreak associated with raw shellfish consumption in South Italy

Descriptive and analytical studies

The outbreak occurred during the Easter holidays, between March 31^st and April 7^th 2002, in the province of Bari (Puglia region, South Italy). In Italy, at Easter many people eat in restaurants and the day after Easter most people go picnicking. In the province of Bari seafood (especially mussels) are often consumed on these occasions. The Regional Epidemiological Office, alerted by the Local Health Unit, performed a field study, collecting data from all General Practitioners and all Emergency Units in the province. Active surveillance was conducted until April 15^th 2002 (one week after the last case onset). The aim of the investigation was to describe the outbreak and to identify the etiologic agent, the source of infection and the means of transmission.

A case was defined as illness in a resident in the area during the period March 31^st – April 7^th, who had diarrhoea (three or more loose stools in any 24 hour period) or vomiting (at least one episode). Fever greater than or equal to 38°C, abdominal pain or nausea were considered additional symptoms. A "probable" secondary case was defined as illness in a household with onset of symptoms more than 24 hours after the primary case.

A retrospective cohort study was performed in order to assess risk factors associated with illness.

All households where a case occurred were included in the study.

Information, collected by means of a standardised questionnaire, included: 1) demographic individual characteristics such as age, gender, occupation; 2) type of food consumed during the last 72 hours before onset of symptoms; 3) if ill, type, date and time of onset of symptoms.

To assess the association between food consumption and disease, relative risks (RR) and 95% confidence intervals (95% CI) were calculated. Age and gender were compared between ill and unaffected individuals by the chi squared and Mann-Whitney tests. A p-value less than 0.05 was considered significant. Variables associated to the illness in the univariate analysis were included in the stratified analysis; summary RR and 95% CI were calculated from the formulas of Greenland and Robins [2]. Collected data were analysed by using Epi Info 6.04 (Centres for Disease Control and Prevention, Atlanta, GA) and Statview 4.0 (Sas Institute Inc., Cary, NC) software.

Laboratory investigations

Faecal specimens were collected from ill individuals. Part of the specimens was refrigerated and processed within 12 hours to detect ova and parasites by direct microscopy and common bacteria by standard methods. The rest was stored at -20°C until examination by NV-specific reverse transcription-polymerase chain reaction (RT-PCR).

Viral nucleic acid extraction and purification from stool specimens was performed as previously reported [3]. RT-PCR was carried out with the primers JV12-SM31 specific for the polymerase gene of NV [4].

Eleven samples of mussels were collected from two fish-markets from which the cases had bought the shellfish they consumed. Mussel samples were also processed within 12 hours to detect common bacteria by standard methods.

Then a nested PCR was used. The procedure for mussel processing as well as viral RNA extraction and purification has been previously described in full [5].

The primers used for first round PCR were JV12 and SM31. Nested PCR was carried out with the use of the primers SR33 for negative strand DNA synthesis and SR48 and SR46 for positive strand synthesis of genogroup I (GI) and genogroup II (GII) sequences, respectively [6].

The 333-bp or 123-bp amplification products from cases and from mussels were subjected to sequencing with PCR primers. When required, cloning was carried out on PCR products. Sequences obtained were aligned with those available in the GenBank.

Epidemiological issues

The clinical pattern of the disease was characterised by the presence of vomiting (84.5%), nausea (58.3%), diarrhoea (53.4%), abdominal pain (47.6%), fever (16.5%). No difference was observed in the clinical pattern by gender. Age was significantly higher in ill individuals with severe diarrhoea (lasting more than 24 hours or undergoing specific treatment) (46.3 vs 38.3 years; p = 0.036). No difference in age was found for the other symptoms.

The outbreak started on March 31^st at 5,00 pm and lasted eight days. The epidemic curve shows a single peak on April 2^nd (35 cases between 1:00 and 12:00 AM) and a right tail probably due to secondary cases (figure 1). Incubation time was not calculated because it was not possible to state a single exposure time. In fact this outbreak could represent the result of multiple small household outbreaks, the source of which could have been either a restaurant meal (on Easter Sunday) or a picnic the day after Easter.

All cases belonged to 30 households that included 139 individuals in all. All members of each household ate together at least once (at the same restaurant or during the same picnic) in the days before the onset of symptoms of relatives. All 36 healthy subjects were interviewed. Nineteen (52.8%) of them were female, the mean age was 42.5 years. Gender and age were not significantly different between healthy and ill individuals (table 1).

An attack rate of 74.1% (103/139) was observed. In the univariate analysis, the association between raw mussel consumption and illness was significant (RR: 1.50; 95% CI: 1.18 – 1.89; p < 0.001). The attack rate according to raw mussel consumption was 86.3% (69/80).

Subjects defined "probable secondary cases" were considered ill according to the risk assessment. In fact, we could not be certain that "probable secondary cases" were actually co-primary cases; on the other hand, the association between raw mussels and illness was stronger excluding from analysis such cases (RR: 1.63; 95% CI: 1.22–2.18; p < 0.001).

Laboratory issues

Both stool and mussel samples were negative for parasites and bacterial enteropathogens. All stool samples from 24 cases were positive for NV by RT-PCR and 5 of 11 mussel samples by nested PCR. The sequences of strains revealed great heterogeneity. In fact, the simultaneous occurrence of GI and GII viruses within the same outbreak was observed. One sample of mussels showed the presence of a mixed genogroup. Sequence analysis showed that strains from 19 cases and 3 mussel samples had identical sequences and belonged to GII, clustering narrowly to the strain NLV/Tarrag/238/2001/Sp, thus strengthening the epidemiologic link of the mussels to the cases. The nucleotide sequences from three cases and 2 mussel samples formed two other distinct clusters showing the best fit with the strains Saitama U25 and Khs1-1997-JP. Sequences from two further cases and one mussel sample were assigned to genogroup I and showed a high degree of identity with the strain NLV/Steinbach/EG/2001/CA.

Three familial groups consisting of 7 cases and three members of a further family group from the outbreak showed identical sequences and were classified as NV GII (the main cluster of 19 cases). A fourth member of the latter family group was positive for one of the two GI strains. The other strains investigated by sequence analysis were unrelated to those from these familial groups.