This is the third outbreak described in a tourist resort in Southern Italy [12, 18] to date. The outbreak occurred during three weeks in July 2005. The delay in reporting the epidemics to the local Public Health Unit only allowed to investigate cases during the last week. Moreover, data concerning the beginning of the outbreak are limited to those reported from the Local Health Unit (personal communication). No further cases were reported after July 29, when specific public-health measures were implemented (such as drinking only bottled water, super-chlorination of the resort tank, stopping the serving of raw mussels, and disposal of previously stored ice).
Infection appears to have spread rapidly within the resort, hitting persons within just a few days after their arrival, similarly to our previous observations during the 2000 epidemics in Italy .
During the period of the outbreak, no other NoV outbreaks were notified outside the resort in the same area. Moreover, a recent study in a close area (Taranto province, see Figure 1) has shown that during 2005 the number of patients with gastroenteritis seeking hospital care was lower compared to 2006 .
The most likely hypothesis supported by the findings of this epidemiological investigation was that illness was associated with raw mussels and ice made by tap water. However just over half of the cases reported eating raw mussels, eggs and/or ice. There are a number of possible explanations for cases not reporting to have eaten the implicated items. Interviews took place during the outbreak period, but the food buffet served to residents changed weekly during the three weeks of the outbreak and answers given were indeed food preferences rather than food actually eaten. Several cases who did not report the consumption of ice were children aged between 2–8, and parents or guardians who answered on behalf of their children may not have been aware of all food items eaten during a holiday period with several annexed social gatherings. Besides, cases might have been secondary cases, resulting from person-to-person transmission.
Raw mussels and ice are a plausible vehicle for infection as described before in Italy [12, 18]. Nonetheless, it should be noted that no food leftover was available for virological investigations. Also in the case of water and ice samples the presence of virus could not be confirmed by specific laboratory analysis. Although laboratory testing for NoV could not be performed on drinking water, the presence of faecal bacteria suggests that the water system may have been the actual source of NoV. As we were able to investigate only the final week of the outbreak, it cannot be ruled out that the association between water and cases may indeed reflect contamination of water with patients stools from within the hotel late during the epidemic. Such a hypothesis appears however unlikely since the environmental inspection did not identify any failure in the water system of the resort. Rather, the observation that tap water samples from different places in the resort showed contamination with faecal bacteria is suggestive that the water pipeline was polluted ahead of the resort. It cannot be ruled out that a possible food item contaminated with NoV may have been involved in some phase during the outbreak, and may have spread the virus to other foodstuff or environment. In particular, it is possible that mussels may have been rinsed with clear water before consumption raw or lightly cooked, although this could not be decisively ascertained from interviews to cooking personnel. In addition to a point source of infection, a person to person transmission of Nov during the outbreak is also likely to have occurred as supported in particular by both the occurrence of few isolated cases in the first days of the outbreak before the peak four days later, and the occurrence of secondary cases within families. Multiple transmission routes have been reported previously .
A conclusive association of the outbreak with noroviruses is supported by the results of laboratory investigations. We used both NoV-specific ELISA and RT-PCR to test 20 faecal samples collected during the outbreak investigation. Using ELISA test, 2 (5%) of the 20 samples were positive for GII antigens. The RT-PCR test detected NoV RNA in 18 (90%) samples, including the two positive by ELISA, out of the 20 cases. These results fit with previous findings that ELISA assays have a lower sensitivity as RT-PCR methods , although ELISA might represent a simple and rapid diagnostic test for timely investigations of NoV outbreaks in laboratories with low capacities. Sequence analysis of the amplified fragments of the ORF1 and ORF2 from 3 samples showed that a recombinant GIIb(ORF1)-GII.3(ORF2) NoV strain [29, 30] was the causative agent of the outbreak. That is not surprising since most GII.b NoV strains described in the literature appear to be recombinant NoV genotype [31, 32]. Sequences from different patients shared 98–100% nucleotide identity, that supports the involvement of a single strain at least during the third week of the outbreak. The spread of GIIb strain of NoV has been previously reported in Italy in 2002, in association with sporadic cases of gastroenteritis in children requiring hospitalization . GIIb strains have circulated in several European countries since 2000 , and have been associated with outbreaks in schools, nursing homes, rural villages and water-borne outbreaks .