Volume:09
Issue:02
Year:
2009
Dr Julian Elston1 BSc MSc (Dist) PhD MPHM, Mike J Wade2 BSc MA Dip SW, Dr Anders Wallensten3,4 MD PhD, Katherine O’Connor5 BSc MCIEH
1 Peninsula Technology Assessment Group (PenTAG), Peninsula College of Medicine and Dentistry, Barrack Road, Exeter
2 Public Health Development Unit, NHS Plymouth, 18 Catherine Street, Plymouth
3 Health Protection Agency South West, The Wheelhouse, Bond's Mill, Stonehouse, Gloucester
4 European Programme for Intervention Epidemiology Training (EPIET), European Centre for Disease Prevention and Control (ECDC),
Stockholm, Sweden
5 Environment and Planning Department, Plymouth City Council, Plymouth
Correspondence: Dr Julian Elston, Honorary Research Fellow and Academic Specialty Registrar in Public Health, PenTAG, Peninsula College of
Medicine and Dentistry, 3rd Floor, Noy Scott House, Royal Devon and Exeter NHS Foundation Trust, Barrack Road, Exeter EX2 5DW, England.
Telephone: +44 (0)1392 406914. Email: julian.elston@nhs.net
Abstract
Background
Salmonella enteritidis infection can cause severe diarrhoea and dehydration, and, in vulnerable people, may result in death. Although fatality rates are generally low, in nursing and residential homes they can be considerably higher. Eggs and poultry are the main source of infection. Preventative measures have reduced the risk posed by eggs over the past decade, although outbreaks continue, often via imported eggs. The susceptibility of residents relates to their generally poorer health, intimate domestic environment and dependence on shared services, all of which can facilitate transmission. This paper reports on an outbreak of Salmonella enteritidis (phagetype 6) at a nursing home in South West England, and identifies the source and mode of transmission.
Methods
Following laboratory notification of salmonellosis at a nursing home, environmental health and Health Protection Agency staff undertook a joint site visit. Environmental and microbiological investigations were conducted and preventative measures put in place. A cohort study was conducted with staff and residents to identify the possible route of transmission, with resident exposure assessed by staff in all but two cases.
Results
Microbiological investigations confirmed PT6 in uncooked eggs in the fridge, and on a dishcloth and tea towel in the kitchen. Nine people (five residents and four staff) had
positive stool samples for Salmonella enteritidis (PT6); two were hospitalised. The epidemic curve suggested either an ongoing source or a point source of infection with personto person transmission. All 32 staff and residents participated in a cohort study. Relative risks for the point source did not support a strong association with any
individual food item, supporting the hypothesis that more than one route was involved.
Conclusion
Although measures to control and prevent the contamination of eggs with salmonella have reduced the risk of infection, this study highlights the ongoing need for promoting best practice in food handling and ensuring catering standards, especially in institutions
caring for vulnerable populations, if serious outbreaks are to be prevented.
Key words:
Cohort study; environmental health; nursing
Introduction
Non-typhoidal salmonella infection is usually a selflimiting gastroenteritis, marked by symptoms of diarrhoea, nausea and vomiting, abdominal pains, fever and headache. It can lead to hospitalisation in dehydrated patients and occasionally result in death. Although in Salmonella enteritidis, the most common foodborne infection in the UK, the case fatality rate is low (0.4-0.6%), it can be considerably greater in susceptible groups such as infants, the immuno-compromised and the elderly. Nursing homes and hospitals, in particular, have high mortality rates, with rates 70 times higher than the wider population (Gil et al., 2009; Mishu et al., 1994).
Salmonella enteritidis is commonly acquired by ingesting uncooked or raw eggs either directly or by contaminated fomites. Person-to-person transmission can also occur when symptomatic individuals have poor hygiene (Health Protection Agency, 2009b), especially if involved in food preparation or if they have learning or cognitive difficulties.
In the 1980s and early 1990s, salmonellosis was a leading cause of infectious intestinal disease in residential homes, leading to more hospital admissions and deaths compared with other outbreaks of intestinal infections (Levine et al., 1991; Strausbaugh et al., 2003). Residents of nursing homes are a particularly vulnerable to infection because of their generally poor physical health, age-related achlorhydria and their dependence on shared facilities like bathrooms and services such as laundry and catering (Strausbaugh et al., 2003).
Over the past decade, a number of preventative measures have been introduced to reduce the level of community infections in the UK. In the mid-1990s, vaccination programmes against S. enteritidis phage-type 4 (PT4), then the predominant cause of human salmonella infection in the UK, were established in broiler breeder and commercial laying poultry flocks. In 2002, UK guidance on food-handling and catering standards, particularly relating to eggs, was produced (Food Standards Agency, 2002). Subsequently, the incidence of S. enteritidis cases in England and Wales fell by 66%, primarily through a reduction in PT4 cases. However, this has been accompanied by a doubling of the incidence of lower prevalence non-PT4 infections (Gillespie, 2004). Nevertheless, egg-borne outbreaks continue (Calvert et al.,2007), suggesting that, despite the inherently lower risk of infection from eggs, other control measures to reduce transmission are not being adhered to. home; Salmonella enteritidis PT6
This paper describes an egg-borne outbreak of Salmonella enteritidis phage-type 6 (PT6) in a nursing home in the South West region of England. The investigation raises a number of issues relating to infection control in eggs, particularly in vulnerable populations.
Methods
Following the notification of two confirmed cases of salmonella infection at a nursing home, environmental and microbiological investigations were promptly instigated by local environmental health and Health Protection Agency (HPA) staff in order to identify the possible source of infection and to implement measures to prevent further cases. A cohort study was subsequently conducted to establish the infection’s source and mode of transmission.
Environmental investigation
Communal areas, particularly the kitchen, were thoroughly inspected, and cleaning and storage facilities such as the refrigerator were checked. Cleaning and cooking rotas were obtained and reviewed. Kitchen surfaces were swabbed and general purpose cloths and three batches of eggs in the refrigerator, some without printed codes on their shell, were taken for microbiological analysis. The eggs were purchased from a local shop.
Microbiological
Stool samples were requested and obtained from all staff (who worked during the outbreak period) and symptomatic residents, and their GPs notified. All samples were tested for enteric pathogens using standard laboratory methods. These include direct inoculation of a one in 10 suspension of the faeces onto Xylose Lysine Deoxycholate (XLD Oxoid number POO164A) and enrichment using Selenite F Broth (Oxoid number EBO354E), which was then inoculated onto XLD after 24 hours’ incubation (BSOP 30, BSOPID 24 [FSA, 2008]). Suspected salmonellas were identified biochemically and serologically using ‘O’ and ‘H’ antisera, before being submitted on solid nutrient agar to the Laboratory for Enteric Pathogens, HPA Centre for Infections, for confirmation and phage typing (Ward et al., 1987).
Swabs, eggs and cloths were examined for salmonella at the Food, Water and Environmental Laboratory at Plymouth Hospitals NHS Trust using HPA Standard Operating Procedure 214 – isolation of salmonella from foodstuffs.
Cohort study
An adapted HPA food poisoning questionnaire was administered by laptop to staff and residents present during the outbreak period – defined by the period three days before the first confirmed case (to allow for the incubation period (Hawker et al., 2005)) to the day of symptom onset reported in the last two cases (in total one week). Questionnaires included items on food consumption (using the daily menus during the outbreak period), contact with people with gastroenteritis, recent travel, and any symptoms of salmonellosis. Residents’ questionnaires were completed by senior care workers as cognitive limitations meant all but two were unable to remember their food consumption two weeks after the outbreak period. Where available, answers were corroborated using entries in residents’ files or with other members of staff. Two clients were excluded from the study as they were fed special processed food, not prepared on site, via gastrostomy tubes. All staff handled food with no responsibility for serving specific residents. Records of interaction between staff and residents and residents themselves were not kept.
For the analysis, probable cases were defined as residents or staff working during the outbreak period and reporting diarrhoea (three or more loose stools in 24 hours). Other symptoms of salmonellosis (fever, nausea, vomiting, bloody stool, abdominal pain) were not used as these were not commonly shared. Confirmed cases were defined as individuals with a positive stool sample for salmonella. Cases were included only if symptoms occurred within the defined outbreak period. Relative risks and attack rates with 95% confidence intervals (95% CI) were calculated using EpiData v3.1 (Lauritsen, 2000-2008). Cases (confirmed or confirmed and probable) and non-cases (staff, residents or both) included in the analysis depended on the item under investigation and the time period for likely exposure. The null hypothesis – no increase in relative risk – was rejected when p<0.05 using the Fisher’s exact test (two-sided).
Results
The findings from the environmental and microbiological investigation informed the analysis of the cohort study.
Figure 1.0 Epidemiological curve of Salmonella Enteritidis outbreak at a nursing home. Note: three confirmed cases were not included in the curve as they did not report symptoms so their date of onset was not known.
Environmental investigation
The kitchen inspection identified a number of problems. The sink for hand washing was difficult to reach and dry when examined. Surfaces and cloths were dirty and the deep-cleaning rota had not been completed for three months. The temperature of the fridge was documented as 6 ºC but when measured was 11 ºC. A number of eggs found in the fridge were visibly contaminated with faeces and feathers. The manager reported that on the eve of the first case, residents’ food was prepared by a non-catering worker due to staff shortages, although the worker was previously employed as a cook at the home.
Microbiological investigation
Laboratory investigation and subsequent typing confirmed that stool samples from nine individuals at the home were positive for Salmonella enteritidis (phage-type 6). Five residents and one member of staff were symptomatic (100% diarrhoea, 57% abdominal pain, 43% vomiting, 29% nausea and 29% fever). The remaining three cases, who did not report symptoms, were identified through microbiological screening of staff. Salmonella enteritidis phage-type 6 (PT6) was also identified on a tea towel, dish cloth and in stored eggs.
Environmental health investigations revealed that the eggs were supplied by a registered wholesaler. Two of the three batches had a UK origin code printed on the shell, one of which was from a large, reputable company with a vaccinated flock. The code on the other batch did not match any UK producer. This batch was infected with Salmonella enteritidis PT6. Checks with the Health Protection Agency did not identify other outbreaks of S. enteritidis PT6 around the period of enquiry.
Cohort study
A total of 32 questionnaires was completed (17 by clients and 15 by staff), giving a response rate of 74% (32/43). The mean age was 63.3 years (range 17.2-96.8 years), and 85% were female.
1) The epidemic curve (Figure 1.0) suggested that a point source was unlikely to be responsible for all cases as the time-lapse between the first and latter cases exceeded the typical incubation period of 12-72 hours (Health Protection Agency 2009a). Given environmental and microbiological findings showing the same phage-type of salmonella in eggs and on contaminated cloths, and that the staff and residents reported no recent travel-associated diarrhoea or contact with people with gastroenteritis, Figure 1.0 suggested four possible routes of transmission:
2) Two point sources; one affecting staff between 13th-15th September and one affecting mainly residents between 17th-20th September.
3) A member of staff became a case from handling eggs before the 15th September, followed by a point source outbreak on the 17th-20th September.
Table 1.0 Relative risk of infection for confirmed cases presenting in residents between18th-20th September
Key: inf – infinity: – infinity Note: As all residents had the same main dishes on 17th September this meant an analysis could not be undertaken. No analysis for performed for items consumed on the 19th September because of insufficient cases (two) for a robust statistical analysis.
4) On-going source: cases may have become ill after eating egg-containing foods on one or more occasions.
5) Cases may have been caused by person-toperson transmission and/or cross-contamination
of food items.
Analysis tested the first three hypotheses in turn. Hypothesis 4 could not be tested directly as person-toperson contact data was not available.
Investigation of hypothesis 1 and 2
No analysis was conducted for the period 13-15th September, as there were insufficient confirmed cases for a statistically robust analysis. Table 1.0 shows the relative risk of infection for confirmed cases in residents only, presenting between18th-20th September. The member of staff and food handler (17th September) was excluded from the analysis so as to ensure transmission could be attributed to food consumption rather than direct environmental contamination in the kitchen.
Table 2.0 The relative risk associated with consumption of any egg-containing meal (including and excluding mousse and bread and butter pudding and for mousse alone)
Key: inf – infinity: – infinity
Note: Seven confirmed cases were used in the analysis of residents and staff, as we included the asymptomatic confirmed case in a staff member.
Table 1.0 shows that six food items had a relative risk greater than one, suggesting their consumption was associated with an increase in risk of infection, but none of these associations reached statistical significance.
Stewed fruit and ice cream and semolina had an infinite relative risk, indicating that only residents who ate these items became infected. Other supper (e.g. soup – the only additional item to be consumed by all cases) and breakfast cereal on both 17th and 18th September were associated with relative risks of infection of 2-3, albeit with wide confidence intervals which included unity. None of these foods contained eggs, the likely source of PT6 in this outbreak. If these relative risks are real, it is possible that these foods may have been crosscontaminated during preparation or when being served.
One explanation for seeing an elevated risk on two consecutive days could be the habitual nature of residents’ diets, ensuring that any breakfast item showing an elevated risk on one day is likely to show a similar increased risk the following day. However, none of our results were statistically significant, possibly due to a lack of power as there were only a small number of cases.
Investigation of hypothesis 3
Table 2.0 shows the relative risk associated with consuming different egg-containing meals. Individuals were defined as being exposed if they had consumed any one or more of the following items during the outbreak period: fried, poached or boiled eggs, mousse, quiche, omelette and bread and butter pudding. Associated risks were analysed for the whole cohort and for residents only, as there were insufficient cases in staff cases to analyse this group separately.
All seven cases were among the 25 residents and staff who ate egg-containing foods. As there were no cases in the six who did not eat egg-containing foods, this suggested that these may be the route of transmission. However, when egg-containing puddings were excluded from the analysis, consumption of egg-containing breakfasts and main courses appeared to protect against infection, suggesting puddings alone might be the mode of transmission. As the mousse was made commercially, it was investigated as a single food item. This revealed an increase in relative risk for staff and residents but not residents alone. These inconclusive results suggested bread and butter pudding might have been the route of transmission, but as this dish was served on the evening of 19th September, it could only have caused the last two cases. However, none of the results were statistically significant, despite a larger cohort being used in some analyses.
Discussion
The environmental investigation found a disorganised kitchen with faulty and un-used equipment and irregular cleaning and staffing practices. Microbiological investigations subsequently revealed that the outbreak strain of Salmonella enteritidis PT6, isolated from affected residents and staff, was indistinguishable from isolates found in a batch of eggs (kept in a malfunctioning refrigerator) and on kitchen cloths. This strongly suggested that eggs were the source of the outbreak.
The epidemic curve suggested four possible transmission modes including point or on-going source, cross-contamination and person-to-person spread. However, despite an apparent increase in risk from consuming four non-egg containing meals on more than one day, no results were statistically significant. Similarly, the consumption of one or more cooked meals containing egg during the outbreak period appeared to be associated with an increased risk, but these results were not statistically significant. Furthermore, the risk of infection reduced when egg-containing puddings were not included. Although this suggested these puddings might be associated with an increased risk, the mousse did not contain eggs brought into the home and the bread and butter pudding could only have accounted for two cases of infection. In addition, none of these results were statistically significant. The isolation of salmonella on the tea towel and dishcloth a week after the outbreak started and evidence of poor cleaning practice and maintenance of kitchen facilities, together with the results of the cohort study, suggested crosscontamination was the most likely mode of transmission. Person-to-person transmission was also possible but could not be explored in the analysis because of lack of data.
Strengths and weaknesses
The main limitation of our investigation related to the cohort study. The elapsed time between consumption of food items and completion of study questionnaires may have introduced bias. In the case of residents, the need to rely heavily on staff to complete their questionnaires may have compromised their validity. However, residents were reported by staff to be consistent in their menu choices, and in a few cases residents’ notes were also used, which would have improved the accuracy of responses. Nevertheless, it was possible that recall of food consumption in cases (compared to non-cases) may have been better, biasing the results.
Although the response rate was good, one confirmed case could not be interviewed as he was on annual leave. This limited the analysis relating to staff alone. Generally, the small number of cases limited the power of the study to detect effects. Finally, difficulties in establishing the time (not just the date) of onset of symptoms limited the ability to more accurately define those at risk of infection in the analysis, weakening ability to find associations between infection and specific food items.
Policy implications
The incident is a pertinent reminder that despite the number of egg-related infections falling significantly over the past decade, salmonella can still be found in eggs purchased in the UK. A recent survey of raw shell eggs used in UK catering premises revealed that one in 335 batches (0.3%) were infected with salmonella, 88% with S. enteritidis, 37% of which were PT6 (Elson and Little, 2004). Although salmonella infection rates are at lower levels following widespread poultry vaccination against PT4, the relative proportion of infections from non-PT4 subspecies has doubled since 1990 (possibly through increased colonisation of breeding, laying and broiler flocks). However, a recent HPA report has implicated imported eggs in continuing outbreaks, particularly from Spain, the largest supplier of eggs to the UK (Pugh 2009). Evidence suggests that between 4.4% and 6.7% of imported Spanish eggs are contaminated – PT6 is one of the more common isolates (Elson and Little, 2004; Little et al., 2007).
The source of the contaminated batch of eggs was investigated by the local authority and the local DEFRA (Department for the Environment, Food and Rural Affairs) egg inspector. There remains some doubt about the provenance of the batch infected with salmonella.
The presence of shell-soiled eggs suggested poor quality control mechanisms at the home, or irreverence or lack of awareness of the low but persistent risk of salmonella infection from eggs. However, the presence of salmonella in egg alone is not sufficient to cause an outbreak. The chain of infection between source and susceptible host needs to be completed. In this respect, residential or nursing homes provide an ideal environment for acquisition and spread of infection, as susceptible residents share food and water sources, and health care in a crowded setting (Strausbaugh et al.,2003). Inadequate food handling practices and general kitchen hygiene, including the lack of refrigerated storage of dairy-containing food products, are key contributory factors to making the link (Frank et al., 2007). In this incident these failings were exacerbated by an undersized kitchen and staffing problems. Poor practice regarding egg storage and usage appears widespread in the UK. A recent survey identified over half of salmonella-contaminated catering premises as not storing eggs under refrigeration, and a fifth storing eggs that had passed their ‘best-before’ dates or using eggs older than three weeks (Food Standards Agency, 2007). Even when storage practices are good, egg-containing meals need to be cooked thoroughly and pasteurised eggs used for recipes requiring raw or lightly cooked eggs to minimise the risk of transmission, especially when catering for vulnerable populations (Food Standards Agency, 2002).
Local authorities have a key role in protecting the health and well-being of clients of residential or nursing homes. This may involve, among other things, monitoring against national guidance or standards, inspection, training on food hygiene and education on good infection control practices.
This study underlines the continuing need for environmental health practitioners to work closely with the management of institutions looking after vulnerable populations to ensure actions aimed at stemming the sources of infection are adequate, and in particular:
Source of infection
Regular and ad hoc inspection of registered egg suppliers and distributors to ensure compliance with regulations and prevent illegally imported eggs entering the supply network.
Encouraging institutions only to purchase Lionmarked eggs or eggs from a vaccinated flock, and that any soiled, feathered or unstamped eggs should be disposed of. The use of eggs from vaccinated flocks should be a material consideration in the food safety management systems of a food business. All food handlers should be aware of the persistent risk of salmonella infection.
Encouraging the thorough cooking of food dishes containing egg. Institutions housing vulnerable people should have a policy of using pasteurised eggs if recipes require raw or uncooked eggs.
Emphasising the importance of maintenance and monitoring of refrigeration equipment and the types of food that should be subject to cool or cold storage.
Mode of transmission
Ensuring that managers are aware of importance of;
- Regular, planned deep cleans of the kitchens and that all staff handling food have been on, at least, the basic level food hygiene course, in order to reduce the risk of infection foods entering into the wider kitchen environment.
- Daily laundering of tea and dish cloths on a hot wash to prevent any bacterium persisting or spreading within the kitchen environment, or the use of disposable cloths.
Conclusion
Although national policy focus is required to support the long-term control of salmonella-infected imports, this must not detract at the local level from the continued
efforts to reduce transmission using basic infection control measures. This outbreak provides evidence to support the need for sustained food handling education and awareness-raising, particularly in institutions caring for vulnerable individuals.
Acknowlegements
Thanks to Dr Geoffrey Thould, Dr Isabel Oliver, Brenda Dale and Joy Slaughter who actively contributed to this investigation and, in the case of former, also commented on the draft paper.
References
Calvert N, Murphy L, Smith A and Copeland D (2007). A hotel-based outbreak of Salmonella enterica subsp. Enterica serovar Enteritidis (Salmonella Enteritidis) in the United Kingdom (2006), Eurosurveillance,12(3): 222. Available online at: www.eurosurveillance.org/ViewArticle.aspx?ArticleId=689 [accessed 20/07/09]
Elson R and Little C (2004). First United Kingdomwide study of raw shell eggs and their use in catering premises. Eurosurveillance, 8(12). Available online: www.eurosurveillance.org/ViewArticle.aspx?ArticleId=2422 [accessed 20/07/09]
Food Standards Agency (2007). UK-wide survey of salmonella in raw shell eggs used in catering premises – final report. Available online: www.food.gov.uk/multimedia/pdfs/eggsurvey2007.pdf [accessed 20/07/09]
Food Standards Agency (2002). Eggs – what caterers need to know. Available online: www.food.gov.uk/multimedia/pdfs/publication/eggscaterers1007.pdf [accessed 20/07/09]
Frank C, Buchholz U, Maass M, Schroder, Bracht K H, Domke P G, Rabsch W, Fell, G (2007). Protracted outbreak of S.Enteritidis PT 21c in a large Hamburg nursing home, BMC Public Health, 7: 243
Gil P R, Alejandre C G, Meca A A, Barrera V H and de Miguel A G (2009). Epidemiology of hospitaltreated salmonella infection; data from a national cohort over a ten-year period, Journal of Infection,58(3):175-181.
Gillespie I (2004). Salmonella enteritidis non-phage type 4 infections in England and Wales: 2000 to 2004: Report from a multi-agency national outbreak control team. Eurosurveillance, 8(42). Available online at: www.eurosurveillance.org/ViewArticle.aspx?ArticleId=2569 [accessed 20/07/09]
Hawker J, Begg N, Blair I, Reintjes R and Weinberg J (2005). Communicable Disease Control Handbook, 2nd Edition, Oxford: Blackwell Publishing.
Health Protection Agency (2009). Salmonella in humans (excluding S. typhi; S. Paratyphi) Faecal and lower gastrointestinal tract isolates reported to the Health Protection Agency Centre for Infections. England and Wales, 1990-2008. Available online at: www.hpa.org.uk/webw/HPAweb&HPAwebStandard/HPAweb_C/1195733760280?p=1191942172078 [accessed 20/07/09]
Health Protection Agency (2008). National Standard Method: Investigation of faecal samples for bacterial pathogens, BSCOP 30, Standards Unit, Department for Evaluations, Standards and Training. Available online at: http://www.hpa-standardmethods.org.uk/documents/bsop/pdf/bsop30.pdf [accessed 20/07/09]