Intended for healthcare professionals


International epidemiological and microbiological study of outbreak of Salmonella agona infection from a ready to eat savoury snack—I: England and Wales and the United States

BMJ 1996; 313 doi: (Published 02 November 1996) Cite this as: BMJ 1996;313:1105
  1. D Killalea, senior registrara,
  2. L R Ward,
  3. D Roberts, deputy directorb,
  4. J de Louvois, head, Environmental Surveillance Unita,
  5. F Sufi, statisticiand,
  6. J M Stuart, consultant epidemiologista,
  7. P G Wall, consultant epidemiologista,
  8. M Susman, consultant in communicable disease controle,
  9. M Schwieger,
  10. P J Sanderson, consultant microbiologistf,
  11. I S T Fisher, scientific coordinatorg,
  12. P S Mead, medical epidemiologisth,
  13. O N Gill, deputy director (information)a,
  14. C L R Bartlett, directora,
  15. B Rowe, directorc
  1. a Public Health Laboratory Service, Communicable Disease Surveillance Centre, London NW9 5EQ
  2. b Food Hygiene Laboratory
  3. c Laboratory of Enteric Pathogens
  4. d Public Health Laboratory Service, Statistics Unit, London NW9 5EQ
  5. e Barnet District Health Authority, Colindale Hospital, London NW9 5HG
  6. f Wellhouse NHS Trust, Edgware General Hospital, Middlesex HA8 0AD
  7. g European Union Salm-Net Surveillance Network, London NW9 5EQ
  8. h Centers for Disease Control and Prevention, 1600 Clifton Road, Atlanta, GA 30333, USA
  1. Public Health Laboratory Service, Central Public Health Laboratory, London NW9 5HT. Salmonella Reference Laboratory, Laboratory of Enteric Pathogens L R WU. Leeds District Health Authority, Leeds LS7 3JX M Schweiger, consultant in communicable disease control. Correspondence to: Dr Gill
  • Accepted 9 October 1996


Objectives: To identify the source of an international outbreak of food poisoning due to Salmonella agona phage type 15 and to measure how long the underlying cause persisted.

Design: Case-control study of 16 primary household cases and 32 controls of similar age and dietary habit. Packets of the implicated foodstuff manufactured on a range of days were examined for salmonella. All isolates of the epidemic phage type were further characterised by pulsed field gel electrophoresis.

Results: 27 cases were identified, of which 26 were in children. The case-control study showed a strong association between infection with S agona phage type 15 and consumption of a peanut flavoured ready to eat kosher savoury snack imported from Israel. S agona phage type 15 was isolated from samples of this snack. The combined food sampling results from the United Kingdom, Canada, the United States, and Israel showed that contaminated snacks were manufactured on at least seven separate dates during a four month period between October 1994 and February 1995. Voluntary recalls of the product successfully interrupted transmission.

Conclusions: Rapid international exchanges of information led to the identification of the source of a major outbreak of S agona in Israel and of associated cases in North America. The outbreak showed the value of the Salm-Net surveillance system and its links outside Europe, both for increasing case ascertainment and for improving the information on the duration of the fault at the manufacturing plant.

Key messages

  • International surveillance networks can strength- en infection control

  • When a foodstuff is contaminated it is important to examine as many batches as possible to show the duration of the production fault


Between 5 December 1994 and 30 January 1995, 27 isolates of Salmonella agona were identified in England and Wales (see fig 1) compared with 12 in the same period a year earlier. Many of the cases were in children, there was geographical clustering, and most of the children had Jewish surnames. We investigated the cause of this food poisoning by means of a case-control study and microbiological testing of the implicated foodstuff.

Fig 1
Fig 1

Distribution of cases of Salmonella agona infection, England and Wales, 1994–5

Subjects and methods

Parents of eight primary household cases were interviewed using a wide ranging semi-structured questionnaire asking about recent food consumption. Four parents each reported that their child had eaten a kosher, peanut flavoured ready to eat savoury snack made of maize and imported from Israel.

Case finding was extended outside England and Wales using Salm-Net, a network for the surveillance of salmonellosis in Europe.1 Information was shared with public health colleagues in other countries, including Israel, the United States, and Canada.

We performed a case-control study of primary household cases in England and Wales between 1 December 1994 and 9 February 1995. A case was defined as S agona of the outbreak strain isolated from a faecal specimen in a child under the age of 10 with an associated diarrhoeal illness. Cases were considered to be secondary if another household member had a diarrhoeal illness in the 24 to 72 hours before the onset of their own illness. Cases were excluded if they had been abroad within the week before the onset of diarrhoea. Parents of cases were asked to nominate three families living in their neighbourhood who were likely to eat kosher food and had a child aged within two years of the case. Controls were excluded if they had been abroad recently or had had a diarrhoeal illness since 1 December 1994. Cases interviewed in the preliminary enquiry were included in the study but separate analyses were undertaken both with and without these cases. A questionnaire requesting information on foods consumed by cases in the three days before illness and by controls in the three days before interview was administered by telephone to parents on 9 and 10 February. The food history included reference to a range of savoury snack products both from the manufacturer of the peanut flavoured snack and other manufacturers in the United Kingdom and Israel.

All isolates of S agona from human cases since 1 December 1993 have been phage typed. All isolates of the epidemic phage type were further characterised by pulsed field gel electrophoresis.2 Packets of the peanut snacks, purchased from retailers, were opened aseptically and the contents (25 g) were examined for Salmonella spp by a standard method.3 The importer and distributor of the implicated snack was approached to obtain details of batch numbers and manufacture dates.



The epidemic strain of S agona was identified as phage type 15. Between 1 December 1994 and the time when interviews in the case-control study were conducted 41 cases of S agona infection were identified. Twenty four were S agona phage type 15, 23 of them were in children aged 7 years or younger, six of them being under 1 year (fig 1). Fifteen of the 24 cases lived in London, eight lived in four different cities, and one lived in Israel. Three further cases infected with S agona phage type 15, all in children under 5 years old who had eaten the savoury snack, were reported by the end of March 1995. In the 13 months from November 1993 to November 1994, before the increase in incidence of S agona infection, 18 cases of infection with S agona phage type 15 were identified, of which six were in children. Pulsed field gel electrophoresis showed that these 18 strains were different from the outbreak strain. Interviews of a sample of these cases did not show any association with eating a kosher diet or the peanut snacks.


Of the 23 cases of infection with S agona phage type 15 in children, two children were abroad and not contactable and five others were excluded from the analysis because the organism had been isolated before 1 December (one), they were possible secondary cases (three), or there was a refusal to cooperate (one). Two potential controls were excluded—one had a diarrhoeal illness and the other a recent history of travel abroad. The median age of cases was 22 months (range 10 months to 5 years 11 months) and that of controls was 30 months (range 16 months to 6 years 4 months). Cases were ill for a median of six days (range 2 to 40), and two cases were admitted to hospital. Fifteen of the cases and all 32 controls were reported to always or sometimes eat kosher food.

When cases from the preliminary inquiry were excluded from the analysis a strong association was found between consumption of the savoury snack and infection with S agona phage type 15 (P=0.0002) (table 1). The association became stronger if all cases were included in the analysis. None of the other foods showed an independent association with illness.

Table 1

Association between illness due to Salmonella agona phage type 15 and consumption of the kosher savoury snack

View this table:


In total 83% (44/53) of the packets of the savoury snack that had been manufactured on 6 November 1994 with a best before date of the end of April 1995 and had been obtained from retailers contained S agona phage type 15. In eight packets subjected to quantitative studies the estimated salmonella count ranged from 2 to 45 organisms per packet (25 g). Three out of 28 packets manufactured on 6 October 1994 with a best before date of the end of March 1995 were also contaminated with S agona phage type 15. One hundred and eighty seven packets from the batch with a best before date of the end of May 1995, which was the batch most commonly available in shops at the time of sampling, were tested and no evidence of salmonella contamination was found. About 20 000 packets of the snack of the batch manufactured on 6 November 1994 were distributed in the United Kingdom.


A large outbreak of S agona was reported to be under way in Israel and is described in the accompanying paper.5 Although no increase was noted in the overall number of S agona infections in the United States, follow up interviews of 26 cases were conducted in four states. Ten cases reported eating the implicated snack before becoming ill, and all their strains isolated were phage type 15. Packets tested in the United States that were manufactured in Israel on 30 October 1994 were contaminated (D Morse, personal communication), as were those tested in Canada that were manufactured on 4 October, 18 November, and 19 December 1994 (J Guzewich and T Gleeson, personal communication), and those tested in Israel that were manufactured on 6 February 1995.5 S agona phage type 15 isolated in Israel, North America, and England from cases and snacks were indistinguishable when tested together.2 None of the 13 European countries contacted through Salm-Net reported any increase in numbers of S agona isolations.


A government food hazard warning identifying the contaminated batch of the savoury snack was issued on 10 February. The company subsequently agreed not to distribute other batches or to import further batches until investigations were completed. The product was also recalled in the United States and Canada. Several additional preventive measures were introduced to the manufacturing process at the production plant in Israel, as described in the accompanying paper.5


Timely national laboratory based surveillanc allowed prompt government action. Recognition of the outbreak and the source of infection was aided by the comparatively low numbers of S agona infections reported annually in England and Wales, the specific group infected, and the geographical clustering of cases. Within a week of the initial case interviews, subsequent epidemiological investigation and isolation of the organism from the product led to effective government control measures.

Young children are both a vulnerable population for infection with salmonella and major consumers of snack products. The peanut flavoured snack was eaten by children as young as 10 months old. Identification of such products as vehicles of infection in children may be difficult as snacks may be eaten outside the home and parents may not be aware of consumption.

Contaminated snacks were manufactured on at least seven separate dates over four months, indicating that the production fault was prolonged. The degree of contamination was expected and found to be low.6

The S agona outbreak demonstrates the importance of communication and collaboration between health officials locally, nationally, and internationally. Sharing of information led to voluntary recalling of the product in North America and in England and Wales, as well as to the identification of the source of a major outbreak in Israel.5 With the modern worldwide distribution of food products, the Salm-Net surveillance network has a crucial role in rapid case finding and information exchange within Europe and further afield.1 7

We acknowledge the contribution to this investigation made by Ms J Demsey, Mr D Wearing, and Mr C Carabine, Barnet Council Environmental Health Department; Ms L Smythers and Mr K Betts, Haringey Council Environmental Health Department; Dr EJ Threlfall of the Public Health Laboratory Service, Laboratory of Enteric Pathogens; Dr RJ Gilbert, director, and Mrs J Thirlwell of the Public Health Laboratory Service Food Hygiene Laboratory, Dr KF Barker, consultant microbiologist, Central Middlesex Public Health Laboratory; Dr N Peel, director, Leeds Public Health Laboratory; the staff of the Leeds Environmental Health Department; Dr O Olojugba, CCDC Bury; Dr L Robinson, Department of Health; Dr A Swan, director, Public Health Laboratory Service Statistics Unit; Dr M O'Mahony and Dr G Nichols, Public Health Laboratory Service Clinical, Epidemiological and Environmental Programmes Section; the staff members of the central, area, and regional public health laboratories; and the staff of the New York State Public Health Department.


  • Funding The Salm-Net surveillance network is a concerted action funded by the European Commission—Directorate General XII, under the Biomedical and Health Research Programme.

  • Conflict of interest None


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
View Abstract