The epidemiological investigation showed that eggs used in the preparation of egg-fried rice, which in turn was used in the preparation of some of the other rice items was the vehicle of infection. Isolation of the unusual phage type 34a strengthened the conclusion that eating from the restaurant was linked to a point source outbreak. In general, by the time investigations are initiated often no food material is available for laboratory analysis and the investigator has to rely on epidemiological evidence. The first step in identifying the source of an outbreak is the calculation of attack rates and the responsible food should have a significantly higher attack rate . In this study, three types of food were found to have higher attack rates and were considered responsible for the outbreak (Table 1). All these items contained egg or egg rice mixture, which was left at room temperature for a long time. Cohort analysis also showed elevated RRs which were significant (Table 2). Seven sick patrons did not eat egg-fried rice. Descriptive analysis showed that all except one gave a history of eating chicken and or special fried rice.
Statistical methods have better power while there is an "a priori" hypothesis as shown in the study of summer excess of leukaemia . We had an "a priori" hypothesis that food items containing eggs increased the risk of illness.
To our knowledge, this is the third report of an outbreak due to phage type 34a and the first of its kind in England published in the literature. In the UK, this phage type has been associated with travel abroad especially to southern Spain  and indigenous infections are rare. The restaurant received eggs from two sources and one of which was a packaging firm. Hence, it was not possible to determine the origin of the eggs. An outbreak due to closely related phage type 34, associated with an egg-containing dish in a Mexican restaurant in the United States has also been described . There have been earlier reports of S. Enteritidis outbreaks associated with Chinese food businesses in England , Scotland  and the United States  although it is not clear whether any shortfalls in specific food handling techniques are responsible. In one instance , egg roll batter was made from pooled shelled eggs which were left at room temperature throughout the day.
The proportion of eggs infected with S. Enteritidis has been reported to be low  and hence the risk of acquiring infection from consuming a single raw egg is much lower. However, the practice of pooling shelled eggs together with storage at room temperature as happened in our outbreak promotes bacterial multiplication and a single contaminated egg can contaminate different types of food. The role of S. Enteritidis in causing food borne outbreaks is well known as it has the ability to contaminate eggs without causing discernible illness in the birds affected . Eggs have been implicated as the source of Salmonella infection in many previous outbreaks [18–22]. Hayes et al  in their case control study in Wales found that undercooked hens eggs are an important risk factor for sporadic Salmonella infections.
We could not find any veterinary data on phage type 34 in British flocks. We searched the literature to determine whether there is any molecular relationship between phage type 34a and phage type 4. Hudson et al  based on the results of pulsed-field gel electrophoresis concluded that different S. Enteritidis phage types appear to be genetically related or clonal. Discussion on stability of phage types of S. Enteritidis can also be found in the literature. Conversion of phage type 4 to 24, phage type 23 to 8 and 4 to 7 have been reported. We could not find reports linking phage type 34a and phage type 4. In a recent public health investigation  of S. Enteritidis in raw eggshells, various serotypes of Salmonella were isolated from 23 out of 449 (5.1%) pooled samples labelled as originating from Spain. These sero/phage types included S. Enteritidis PT6a, PT5c, 13a, 14b, 58, PT6d, PT1, PT1c and PT12.
A few limitations of this study are to be noted. The origin of the suspected contaminated eggs could not be traced. Although trace back exercises are key in epidemic investigations, often they are not successful due to logistic and practical reasons. In our outbreak one of the suppliers to the restaurant turned out to be a packaging firm. During our investigation, we found that there were problems with the distribution system, which prevented us from pin pointing the origin of the contaminated eggs. A possibility always exists that we missed a few subjects from this investigation and not all could be persuaded to provide a stool sample. There was no single list of all the patrons who ate/purchased food on the evening. However, all efforts were made to contact the patrons and the outbreak caused considerable publicity in the local media. Hence, we are confident that we have included most patrons. Although the precise number of patrons who were not included will never be known we are confident that their number is small and might be in the region of 10 to 15.
Palmer  has pointed out the need to undertake outbreak investigations rapidly but at the same time with sound methodology. We tried to adopt the standard approach to investigating an outbreak including a retrospective cohort study. However, we did not attempt multivariate analysis due to the small number of subjects involved in the investigation.
S. Enteritidis PT4 was isolated from one of the subjects. Further investigation revealed that this subject had recently returned from holiday in continental Europe and had suffered mild symptoms before the meal.
In response to this and other outbreaks associated with eggs, a Public Health Investigation was launched in October 2002 in the UK to determine the rate of Salmonella contamination in eggs. Tests of nearly 4000 eggs showed that Salmonella was recovered from 5.3% of pooled eggs . The Food Standards Agency has also produced a leaflet titled "Eggs – what caterers need to know"  which emphasises the importance of thoroughly cooking the eggs, buying eggs from reputable suppliers and use of pasteurised eggs when serving a vulnerable individuals.