In 1998, an outbreak of gastroenteritis affected at least 448 persons including 122 staff at a resort hotel in Bermuda. A survey among staff indicated that gastroenteritis was associated with eating or drinking at the hotel (OR = 6·0, 95% CI = 2·4–15·1). Multiple specimens of drinking water had elevated faecal coliform levels and Escherichia coli present, suggestive of faecal contamination. Stools from 18 of the 19 persons with gastroenteritis that were tested were positive for genogroup-II Norwalk-like viruses (NLVs). RT–PCR analysis of a 3 l specimen of water produced a genogroup-II NLV genome with a sequence identical to that of NLVs in the stools of three ill persons. This outbreak shows the value of new molecular diagnostics to link illness with a contaminated source through the use of sequence analysis. The risk of outbreaks such as these could be reduced in tourism dependent regions like Bermuda and the Caribbean by regular evaluation of data from the inspection and monitoring of drinking water supplies and waste water systems, by ensuring the chlorination of supplemental drinking water supplies and by establishing food-safety initiatives.

Although food handlers are often implicated as the source of infection in outbreaks of food-borne viral gastroenteritis, little is known about the timing of infectivity in relation to illness. We investigated a gastroenteritis outbreak among employees of a manufacturing company and found an association (RR=14·1, 95% CI=2·0–97·3) between disease and eating sandwiches prepared by 6 food handlers, 1 of whom reported gastroenteritis which had subsided 4 days earlier. Norwalk-like viruses were detected by electron microscopy or reverse transcriptase-polymerase chain reaction (RT-PCR) in stool specimens from several company employees, the sick food handler whose specimen was obtained 10 days after resolution of illness, and an asymptomatic food handler. All RT-PCR product sequences were identical, suggesting a common source of infection. These data support observations from recent volunteer studies that current recommendations to exclude food handlers from work for 48–72 h after recovery from illness may not always prevent transmission of Norwalk-like viruses because virus can be shed up to 10 days after illness or while exhibiting no symptoms.