Gastrointestinal tract infections are common in both developing and developed countries [1] and are produced by a wide variety of enteropathogens, including bacteria, viruses and parasites. The main causes of viral gastroenteritis include noroviruses and rotaviruses [2–4]. Since the first described outbreak of Norwalk virus gastroenteritis in an elementary school in Norwalk (OH, USA) in the fall of 1968 [5], studies have shown that viral infections, especially those due to Noroviruses (NoVs), are the most frequent causes of acute gastroenteritis in the community [6–8]. NoVs infect all age groups, with particularly severe disease occurring in young children, the elderly and persons with chronic illnesses [9]. Because of the low infectious dose (10–100 viral particles can induce symptoms), short-lived immunity, and high stability in the environment, these viruses are especially contagious and outbreaks are characterised by high secondary attack rates [9]. NoV infections have historically been described as mild and self-limiting [6,8]. Studies in the 1970s were the basis for the criteria used by Kaplan et al. [10] to discern outbreaks with a viral aetiology, which include stool culture negative for bacterial pathogens, mean (or median) duration of illness 12–60 h, vomiting in 50% of cases, and, if known, mean (or median) incubation period of 24–48 h [10]. Kaplan et al. criteria are still considered valid [11]. NoVs cause outbreaks in a wide range of settings, especially in hotels [12], hospitals [13], retirement centres [14], schools [15,16], and cruise ships [17]. The mode of transmission may be foodborne, waterborne or person-to-person contact. Point source outbreaks are usually related to contaminated food or water [17,18], while secondary transmission often results from person-to-person contact [19]. NoVs have also been identified in sporadic cases of gastroenteritis [20,21]. Although outbreaks and sporadic disease may occur year round, countries of the Northern hemisphere mainly show a seasonal pattern of increased occurrence during the winter months [22].

NoVs form a genus within the family of Caliciviridae and are genetically and antigenically highly variable. Five genogroups (G) of NoVs have been tentatively assigned from the molecular characterization of complete capsid gene sequences. Strains of three genogroups, GI, GII, and GIV, are found in humans (GII/11 are porcine), and GIII and GV strains are found in cows and mice, respectively [23].

Several large foodborne and waterborne outbreaks due to NoVs have been described [17,19,24]. However, there are only a few reports in which both epidemiological and environmental data have been confirmed by molecular data on the waterborne NoVs [25,26]. Up to 93% of the outbreaks and sporadic cases of nonbacterial, acute gastroenteritis in humans and about 60–85% of all gastroenteritis outbreaks, specifically within the United States, Europe, and Japan, are associated with NoVs [24].

In Greece, which has no surveillance system for non-bacterial gastroenteritis, the impact of NoVs infection is unknown, and very few epidemiological studies about NoVs have been reported so far. In fact, only two epidemiological studies have been reported previously in Greece [27,28] and only one refers to a possible waterborne outbreak [28]. Furthermore, there is no previous molecular epidemiological study in Greece about NoVs. The epidemiological as well as the molecular investigation of non-bacterial gastroenteritis is rather absent in Greece.

This report describes a large outbreak of non-bacterial gastroenteritis caused by NoVs in North Eastern Greece. It is the first study, to our knowledge in Greece, trying to combine environmental, epidemiological and molecular investigations in order to specify the cause and the conditions of the outbreak and to examine its characteristics. Also it is the first study which includes a molecular investigation of the NoVs identified during the studied outbreak.

Xanthi is the capital of the prefecture of the same name located in North Eastern Greece. The population of Xanthi prefecture was 100;882 inhabitants, according to the National Census of 2001 [29]. Water supply of the capital and the main communities of the prefecture mainly depend on surface water sources, as well as on drilled wells. In the first days of June of 2006, reports of gastroenteritis cases to the General Hospital of Xanthi increased markedly.

The epidemic curve presented a rapid increase of gastroenteritis cases during June, as well as high occurrence of cases is observed during July. The number of cases declined slowly during the next three months and the epidemic finally stopped during the first days of November. This epidemic curve suggests an occurrence of an outbreak and its beginning related to a possible common source of contamination. The initial source could have been contaminated drinking water, caused by the poor hygienic conditions of the closeby local communities and the intense rainfall that preceded the emergence of the outbreak. The environmental and hygienic conditions in the areas where water supply sources are installed, suggest an increased risk of drinking water contamination, although no analytical epidemiological study was performed to confirm the waterborne source of the outbreak. Although the water samples tested did not confirm this suggestion and no NoVs were detected, we have to clarify that these samples were taken after a significant time delay. The delay of the water sampling was due to delayed announcement of the outbreak by the water company as well as the lack of a surveillance system for non-bacterial gastroenteritis. Therefore, the negative results of the environmental investigation do not exclude the possibility of a waterborne emergence of the outbreak. The waterborne hypothesis is supported by the fact that the outbreak studied is the second gastroenteritis epidemic reported in almost the same area [28]. A possible waterborne gastroenteritis outbreak due to Norovirus infection in Xanthi during the period January 28 to February 10, 2005 had been previously described. Papadopoulos et al. stated that for the most suspicious event that could have been responsible for the contamination of a main well and probably responsible for the NoVs outbreak was a water flood which resulted from heavy rain during the night of January 27, 2005 [28]. During this outbreak, a total of 709 patients visited the local hospital over a period of two weeks with symptoms of acute gastroenteritis. This is the only Norovirus outbreak reported in the literature in Greece [28]. The occurrence of two large outbreaks of gastroenteritis in the same area within two years after a similar event (heavy rainfall) suggests a low level of sanitary conditions and a possible source of repeated contamination of drinking water.

As observed in other outbreaks, this outbreak affected all age groups [33–36]. Cases represented all ages (range 0.25–89 y.o). However, young people presented higher attack rates, this could have been related to a lower level of natural immunity, to frequent contacts among children or to the fact that people tend to go to the hospital more quickly for their children. In Greece, there is no reporting surveillance system for gastroenteritis outbreaks, and it is estimated that consequently a large number of gastroenteritis cases are not recorded and as a result, many cases of gastroenteritis that occurred during the outbreak may have not been recorded. This could be a fact mainly for mild cases or cases occurring in rural areas. A number of patients probably contacted private physicians, or rural Health Centers existing in the studied area. Such patients were not recorded as part of the cases of the studied outbreak. These methodological issues concerning case identification and case ascertainment put some limitations to the interpretation of our data. However, there are strong indications for the outbreak such as that the epidemic curve showed a fast increase of cases at the emergence and the duration of the outbreak was many times longer than the incubation period of noroviruses (12–48 hours).

Sequence BLAST search and the Norovirus genotyping tool of Greek NoVs strains’ sequences along with reference GII strains confirmed that the outbreak isolates belonged to GII but the identification of GII genotypes was not possible due to the short amplified sequence of the RNA polymerase gene. While the RNA polymerase region is mostly used for diagnosis of NoVs, a complete identification requires the use of the capsid region [30] which was not possible in our study (due to the high cost). The use of short sequences has been successful for establishing diagnoses of NoV infection, but it becomes problematic for the classification or phylogenetic analysis [23]. Our results show that the outbreak was caused by GII strains. This is in accordance with studies from different parts of the world reporting the predominance of GII [6,16,21,36,37].

Outbreaks in hospitals and institutions are more easily identified, cause more disruption compared to outbreaks in the community and are more likely to be reported [38]. Also the community outbreaks are more difficult to yield clinical specimens for identification of Norovirus. The emerging behavior of Norovirus has stimulated a considerable interest in the molecular epidemiology of these viruses. In turn, this has been complicated by the underestimated rate of recombination between strains [35,36] and the lack of a simple in vitro culture system with high yield.

The outbreak is one of the very few extended outbreaks due to Noroviruses in Greece. Due to the lack of a surveillance system for acute gastroenteritis in Greece, the accomplishment of the study proved very difficult. The emergence of the outbreak was possibly due to contaminated drinking water, although this was not proved by water analyses data. Its long duration is likely due to a large number of cases related to person-to-person contamination. The fact that the described outbreak was the second gastroenteritis outbreak in the same area, within the time interval of two years, suggests prolonged negligence in drinking water sanitary conditions (e.g., lack of stable chlorination). Also, the present epidemiological study confirmed the necessity of the development of a surveillance system for acute non-bacterial gastroenteritis outbreaks in Greece.