- Email: [email protected]
Food poisoning in a hospital staff canteen
Journal of Infection (I986) I3, I95-I98
EPIDEMIOLOGY
F o o d p o i s o n i n g in a h o s p i t a l staff c a n t e e n P h i l i p p a M. B. W h i t e
Public Health Laboratory, Coventry & Warwickshire Hospital, Coventry Accepted for publication 2 April z986 Summary Eighty-two confirmed cases of salmonella food poisoning arose among hospital staff due to consuming contaminated tartare sauce served in the staff canteen. Many key personnel were affected and the hospital was closed to non-urgent admissions. In order to maintain the accident and trauma services, the normal policy of excluding infected persons from work had to be modified. Staff returned to work 48 h after they had become asymptomatic provided that they did not have contact with patients' mouths, food or drink. There were no secondary cases. During the investigation of the outbreak, lack of national guidelines for the preparation and handling of mayonnaisebased food products became apparent. Introduction F o o d - b o r n e outbreaks a m o n g hospital staff m a y have major implications for the continued provision o f services. T h e hospital involved in this episode of food poisoning has I86 beds, p r e d o m i n a n t l y for orthopaedic and trauma patients. It provides accident and emergency services for a wide area, including a b u s y stretch of m o t o r w a y , and houses the outpatient clinics for the District. N i n e h u n d r e d and fifty-four full-time m e m b e r s of staff are based at the hospital. T h i s report describes the investigation and m a n a g e m e n t of the episode as well as modifications to the normal policy of excluding salmonella excretors. T h e s e changes were perceived as necessary for maintaining the accident services to a local population of approximately 33oooo.
The outbreak T h e incriminated meal was served for lunch in two staff dining rooms on T u e s d a y I9 F e b r u a r y i985. D u r i n g the afternoon of T h u r s d a y 2I F e b r u a r y the infection control nurse was informed that three of the junior medical staff were ill with abdominal pain, fever, influenza-like s y m p t o m s and diarrhoea. By that evening, the n u m b e r of those ill had risen to eight with five requiring admission to hospital. W h e n a further four m e m b e r s of staff reported similar s y m p t o m s the next morning, it appeared likely that there was a major outbreak of food poisoning and the local D e p a r t m e n t o f Environmental Health was invited to assist in the investigation. Environmental H e a l t h Officers ( E H O s ) visited victims who lived away from the hospital site in order to collect specimens and to take food histories. At that stage there was a widely held belief among staff that fish oi63-4453/86/o5oi95+o4 $02.00/0
© I986 The British Society for the Study of Infection
I96
P.M.B.
WHITE
served for lunch on Tuesday 19 February was the likely cause of the incident. T h e rest of the main course had consisted of chips, peas and tartare sauce. None of these had been served to the patients who were unaffected. T h e tartare sauce, which had been prepared in the hospital kitchen, seemed to be the most likely vehicle of infection. T h r e e jars of the same batch were about to be served for lunch when these were retrieved together with other foods and submitted to the laboratory for microbiological examination. Later in the day study of the food histories taken by the EHOs and hospital staff confirmed that the fish or, more likely, the tartare sauce was the probable cause. Victims included members of staff who ate in the canteen only on Tuesday lunchtime and one unfortunate person who liked tartare sauce so much that she put it on her homemade sandwiches. A meeting of interested parties (including microbiologists, the Medical Officer for Environmental Health, Environmental Health Officers, Infection Control and Occupational Health Nurses, catering staff, clinicians, senior nursing staff and administrators) was held on Friday 22 February. T h e Communicable Disease Surveillance Centre of the Public Health Laboratory Service was informed of the outbreak. It had become obvious that because of shortage of staff the hospital could not continue to provide a full service so it was decided to close the hospital to non-urgent admissions. It was known that I3o portions of fish had been served. In an effort to identify all potential victims, a circular was sent to each department requesting details of those who had eaten the fish or were absent from work. By the end of the day, 4o members of staff had reported ill. A presumptive salmonella, later identified as Salmonella typhimurium, was isolated late on Friday afternoon from the first patients to present. T h e same organism was isolated later in very small numbers from the tartare sauce which had been retrieved from the kitchen. T h e last report of a new case was received on Thursday 28 February. Virtually all departments in the hospital were affected with Io4 members of staff reporting ill. Of these, 82 were culture positive. This represents an attack rate of 80 %. Salmonella typhimurium was isolated from 79 % of the victims. T h e major long-term problem created by this outbreak was how to maintain orthopaedic and trauma services w i t h so much sickness among key staff and the possibility of prolonged excretion of salmonellae after symptomatic recovery. It was agreed at a Control of Infection Committee meeting that, with the exception of food handlers who were to follow current recommendations, and after the importance of personal hygiene had been emphasised, staff could return to work 48 h after they had become asymptomatic provided that they did not take patients' temperatures, attend to oral hygiene or give food, drink or oral medication to patients. Faecal samples were collected in the usual manner until three consecutive samples were negative. Secondary cases did not arise among hospital patients. T h e strain of S. typhimurium isolated was later found to belong to phage type 197, a type uncommon in this country and which had not been reported in human beings or animals for some time. Since this outbreak, several other sporadic cases have been reported to the Communicable Disease Surveillance Centre of the Public Health Laboratory Service (personal communication).
Food poisoning among hospital staff
I97
It proved impossible to identify the ingredient that introduced the organism into the food since it was not isolated from those items remaining in the kitchen (gherkins, capers and pepper). Egg yolk would seem to be the most likely candidate but investigations by the EHOs did not reveal a salmonella outbreak in the egg producers' flocks. Tartare sauce
T h e preparation of the tartare sauce was examined in detail. It had begun on the afternoon before the lunch with the production of a mayonnaise sauce which contained 60 egg yolks, 6.82 1 gallons of warmed oil, mustard, vinegar, seasoning and lemon juice. T h e resulting product was stored in a refrigerator overnight. T h e following morning o'57 1 of gherkins and 0.28 1 of capers were minced and added to the mayonnaise. T h e total volume of sauce produced was IO 1 and the pH of the final product was 3"9. A test batch of tartare sauce was made under conditions similar to those of the affected sauce. T h e p H of the final product was 4"0, but the pH at the mayonnaise stage was 4"7. A series of inoculation experiments was designed to study the ability of both tartare sauce and mayonnaise to support the growth and replication of the organism responsible for the outbreak. These showed that, with an inoculum of lO6/1, S. typhimurium did not multiply in either test material. Over a period of 24 h there was little change in the bacterial counts of the material kept at 4 °C or room temperature but the count fell approximately Io-fold when the sauces were kept at 37 °C. Discussion
. Two important aspects of this outbreak need to be considered. First is the microbiology of tartare sauce; second the problems caused by a high rate of sickness among hospital staff. Several outbreaks of salmonella food poisoning due to contaminated mayonnaise have been reported. 1 T h e microbiology of mayonnaise was reviewed by Smittle. 2 T h e low pH, salt concentration and presence of organic acids, in particular acetic acid, make it an unfavourable environment for the survival or growth of most bacteria. Salmonellae, however, may survive but not multiply in mayonnaise for at least 24 h. This was confirmed by inoculation experiments performed in this laboratory and suggests that the original inoculum in the sauce must have been large. T h e very high attack rate and proportion of positive specimens suggest heavy contamination. This, in the absence of bacterial multiplication in the food, makes it unlikely that the outbreak had been caused by a carrier among the catering staff. T h e ability of mayonnaise to kill salmonellae more rapidly at room temperature than in the refrigerator is well recognised. T h e United States Food and Drug Administration recommends that the p H of mayonnaise should be 4"I or less and that the product should be held at 18-22 °C for at least 72 h before use unless pasteurised egg is used). 4 Following an outbreak in I955 of IOOOO cases of salmoneUosis due to contaminated mayonnaise the Danish Administration recommended a holding period of 4 days at room temperature
198
I,. M. B. WHITE
or 2 h at 4o °C. Such recommendations are contrary to general guidance on food handling practices in the U . K . although pasteurised egg is used in the commercial production of mayonnaise. In the light of this and other outbreaks of salmonella food poisoning associated with contaminated mayonnaise or mayonnaise-based products similar guidelines should be considered for this country. Following a salmonella infection it has been normal practice to exclude from work those who pose a special risk of spreading infection until three negative faecal specimens have been obtained. 5 T h i s group includes food handlers and health care staff. I f this r e c o m m e n d a t i o n had been applied during and following this outbreak it would have been impossible to maintain the only accident and t r a u m a service available to one-third of a million people. It was decided that closure of the service was not acceptable and the compromise described was agreed. S u p p o r t for this policy comes from a study by Pether and Scott 6 who looked at contamination of the hands of convalescent salmonella excretors. T h e y showed that, although hands were contaminated after defaecation, detectable salmonellae were easily removed by washing the hands with soap and water. T h e success of so m o d i f y i n g standard control measures confirms the view that prolonged exclusion f r o m work of asymptomatic salmonella excretors in some special risk groups m a y be unnecessary provided that adequate personal hygiene is maintained. (I wish to acknowledge the work of the hospital infection control team, led by Dr R. M. Hutton, and the Coventry Environmental Health Department in the investigation and control of this outbreak. ! am most grateful to Dr P. R. Mortimer for help in preparing this paper.) References I. Davies RF, Wahba AH. Salmonella Infections of Charter Flight Passengers, Report on a Visit to Spain (Canary Islands) 26 February - 2 March 1976. Copenhagen: WHO Reg. Off. Eur. 2. Smittle RB. Microbiology of Mayonnaise and Salad Dressing: A ReviewJ Food Protection 1977; 4o: 415-422. 3. U.S. Department of Health, Education and Welfare. Dressings for Food. Mayonnaise. 21 CFR 25.1. Washington D.C., U.S. Gov. Printing Off., 1975. 4. U.S. Department of Health, Education and Welfare. Dressings for Food. Salad Dressing. 21 CFR 25.3. Washington D.C., U.S. Gov. Printing Off., 1975. 5. P.H.L.S. Salmonella Sub-Committee. Notes on the control of human sources of gastrointestinal infections, infestations and bacterial intoxications in the United Kingdom. Community Med 1963; 5:152-169 • 6. Pether JVS, Scott RJD. Salmonella carriers; are they dangerous ? A study to identify finger contamination with Salmonellae by convalescent carriers. J Infect 1982; 5:81-88.
EPIDEMIOLOGY
F o o d p o i s o n i n g in a h o s p i t a l staff c a n t e e n P h i l i p p a M. B. W h i t e
Public Health Laboratory, Coventry & Warwickshire Hospital, Coventry Accepted for publication 2 April z986 Summary Eighty-two confirmed cases of salmonella food poisoning arose among hospital staff due to consuming contaminated tartare sauce served in the staff canteen. Many key personnel were affected and the hospital was closed to non-urgent admissions. In order to maintain the accident and trauma services, the normal policy of excluding infected persons from work had to be modified. Staff returned to work 48 h after they had become asymptomatic provided that they did not have contact with patients' mouths, food or drink. There were no secondary cases. During the investigation of the outbreak, lack of national guidelines for the preparation and handling of mayonnaisebased food products became apparent. Introduction F o o d - b o r n e outbreaks a m o n g hospital staff m a y have major implications for the continued provision o f services. T h e hospital involved in this episode of food poisoning has I86 beds, p r e d o m i n a n t l y for orthopaedic and trauma patients. It provides accident and emergency services for a wide area, including a b u s y stretch of m o t o r w a y , and houses the outpatient clinics for the District. N i n e h u n d r e d and fifty-four full-time m e m b e r s of staff are based at the hospital. T h i s report describes the investigation and m a n a g e m e n t of the episode as well as modifications to the normal policy of excluding salmonella excretors. T h e s e changes were perceived as necessary for maintaining the accident services to a local population of approximately 33oooo.
The outbreak T h e incriminated meal was served for lunch in two staff dining rooms on T u e s d a y I9 F e b r u a r y i985. D u r i n g the afternoon of T h u r s d a y 2I F e b r u a r y the infection control nurse was informed that three of the junior medical staff were ill with abdominal pain, fever, influenza-like s y m p t o m s and diarrhoea. By that evening, the n u m b e r of those ill had risen to eight with five requiring admission to hospital. W h e n a further four m e m b e r s of staff reported similar s y m p t o m s the next morning, it appeared likely that there was a major outbreak of food poisoning and the local D e p a r t m e n t o f Environmental Health was invited to assist in the investigation. Environmental H e a l t h Officers ( E H O s ) visited victims who lived away from the hospital site in order to collect specimens and to take food histories. At that stage there was a widely held belief among staff that fish oi63-4453/86/o5oi95+o4 $02.00/0
© I986 The British Society for the Study of Infection
I96
P.M.B.
WHITE
served for lunch on Tuesday 19 February was the likely cause of the incident. T h e rest of the main course had consisted of chips, peas and tartare sauce. None of these had been served to the patients who were unaffected. T h e tartare sauce, which had been prepared in the hospital kitchen, seemed to be the most likely vehicle of infection. T h r e e jars of the same batch were about to be served for lunch when these were retrieved together with other foods and submitted to the laboratory for microbiological examination. Later in the day study of the food histories taken by the EHOs and hospital staff confirmed that the fish or, more likely, the tartare sauce was the probable cause. Victims included members of staff who ate in the canteen only on Tuesday lunchtime and one unfortunate person who liked tartare sauce so much that she put it on her homemade sandwiches. A meeting of interested parties (including microbiologists, the Medical Officer for Environmental Health, Environmental Health Officers, Infection Control and Occupational Health Nurses, catering staff, clinicians, senior nursing staff and administrators) was held on Friday 22 February. T h e Communicable Disease Surveillance Centre of the Public Health Laboratory Service was informed of the outbreak. It had become obvious that because of shortage of staff the hospital could not continue to provide a full service so it was decided to close the hospital to non-urgent admissions. It was known that I3o portions of fish had been served. In an effort to identify all potential victims, a circular was sent to each department requesting details of those who had eaten the fish or were absent from work. By the end of the day, 4o members of staff had reported ill. A presumptive salmonella, later identified as Salmonella typhimurium, was isolated late on Friday afternoon from the first patients to present. T h e same organism was isolated later in very small numbers from the tartare sauce which had been retrieved from the kitchen. T h e last report of a new case was received on Thursday 28 February. Virtually all departments in the hospital were affected with Io4 members of staff reporting ill. Of these, 82 were culture positive. This represents an attack rate of 80 %. Salmonella typhimurium was isolated from 79 % of the victims. T h e major long-term problem created by this outbreak was how to maintain orthopaedic and trauma services w i t h so much sickness among key staff and the possibility of prolonged excretion of salmonellae after symptomatic recovery. It was agreed at a Control of Infection Committee meeting that, with the exception of food handlers who were to follow current recommendations, and after the importance of personal hygiene had been emphasised, staff could return to work 48 h after they had become asymptomatic provided that they did not take patients' temperatures, attend to oral hygiene or give food, drink or oral medication to patients. Faecal samples were collected in the usual manner until three consecutive samples were negative. Secondary cases did not arise among hospital patients. T h e strain of S. typhimurium isolated was later found to belong to phage type 197, a type uncommon in this country and which had not been reported in human beings or animals for some time. Since this outbreak, several other sporadic cases have been reported to the Communicable Disease Surveillance Centre of the Public Health Laboratory Service (personal communication).
Food poisoning among hospital staff
I97
It proved impossible to identify the ingredient that introduced the organism into the food since it was not isolated from those items remaining in the kitchen (gherkins, capers and pepper). Egg yolk would seem to be the most likely candidate but investigations by the EHOs did not reveal a salmonella outbreak in the egg producers' flocks. Tartare sauce
T h e preparation of the tartare sauce was examined in detail. It had begun on the afternoon before the lunch with the production of a mayonnaise sauce which contained 60 egg yolks, 6.82 1 gallons of warmed oil, mustard, vinegar, seasoning and lemon juice. T h e resulting product was stored in a refrigerator overnight. T h e following morning o'57 1 of gherkins and 0.28 1 of capers were minced and added to the mayonnaise. T h e total volume of sauce produced was IO 1 and the pH of the final product was 3"9. A test batch of tartare sauce was made under conditions similar to those of the affected sauce. T h e p H of the final product was 4"0, but the pH at the mayonnaise stage was 4"7. A series of inoculation experiments was designed to study the ability of both tartare sauce and mayonnaise to support the growth and replication of the organism responsible for the outbreak. These showed that, with an inoculum of lO6/1, S. typhimurium did not multiply in either test material. Over a period of 24 h there was little change in the bacterial counts of the material kept at 4 °C or room temperature but the count fell approximately Io-fold when the sauces were kept at 37 °C. Discussion
. Two important aspects of this outbreak need to be considered. First is the microbiology of tartare sauce; second the problems caused by a high rate of sickness among hospital staff. Several outbreaks of salmonella food poisoning due to contaminated mayonnaise have been reported. 1 T h e microbiology of mayonnaise was reviewed by Smittle. 2 T h e low pH, salt concentration and presence of organic acids, in particular acetic acid, make it an unfavourable environment for the survival or growth of most bacteria. Salmonellae, however, may survive but not multiply in mayonnaise for at least 24 h. This was confirmed by inoculation experiments performed in this laboratory and suggests that the original inoculum in the sauce must have been large. T h e very high attack rate and proportion of positive specimens suggest heavy contamination. This, in the absence of bacterial multiplication in the food, makes it unlikely that the outbreak had been caused by a carrier among the catering staff. T h e ability of mayonnaise to kill salmonellae more rapidly at room temperature than in the refrigerator is well recognised. T h e United States Food and Drug Administration recommends that the p H of mayonnaise should be 4"I or less and that the product should be held at 18-22 °C for at least 72 h before use unless pasteurised egg is used). 4 Following an outbreak in I955 of IOOOO cases of salmoneUosis due to contaminated mayonnaise the Danish Administration recommended a holding period of 4 days at room temperature
198
I,. M. B. WHITE
or 2 h at 4o °C. Such recommendations are contrary to general guidance on food handling practices in the U . K . although pasteurised egg is used in the commercial production of mayonnaise. In the light of this and other outbreaks of salmonella food poisoning associated with contaminated mayonnaise or mayonnaise-based products similar guidelines should be considered for this country. Following a salmonella infection it has been normal practice to exclude from work those who pose a special risk of spreading infection until three negative faecal specimens have been obtained. 5 T h i s group includes food handlers and health care staff. I f this r e c o m m e n d a t i o n had been applied during and following this outbreak it would have been impossible to maintain the only accident and t r a u m a service available to one-third of a million people. It was decided that closure of the service was not acceptable and the compromise described was agreed. S u p p o r t for this policy comes from a study by Pether and Scott 6 who looked at contamination of the hands of convalescent salmonella excretors. T h e y showed that, although hands were contaminated after defaecation, detectable salmonellae were easily removed by washing the hands with soap and water. T h e success of so m o d i f y i n g standard control measures confirms the view that prolonged exclusion f r o m work of asymptomatic salmonella excretors in some special risk groups m a y be unnecessary provided that adequate personal hygiene is maintained. (I wish to acknowledge the work of the hospital infection control team, led by Dr R. M. Hutton, and the Coventry Environmental Health Department in the investigation and control of this outbreak. ! am most grateful to Dr P. R. Mortimer for help in preparing this paper.) References I. Davies RF, Wahba AH. Salmonella Infections of Charter Flight Passengers, Report on a Visit to Spain (Canary Islands) 26 February - 2 March 1976. Copenhagen: WHO Reg. Off. Eur. 2. Smittle RB. Microbiology of Mayonnaise and Salad Dressing: A ReviewJ Food Protection 1977; 4o: 415-422. 3. U.S. Department of Health, Education and Welfare. Dressings for Food. Mayonnaise. 21 CFR 25.1. Washington D.C., U.S. Gov. Printing Off., 1975. 4. U.S. Department of Health, Education and Welfare. Dressings for Food. Salad Dressing. 21 CFR 25.3. Washington D.C., U.S. Gov. Printing Off., 1975. 5. P.H.L.S. Salmonella Sub-Committee. Notes on the control of human sources of gastrointestinal infections, infestations and bacterial intoxications in the United Kingdom. Community Med 1963; 5:152-169 • 6. Pether JVS, Scott RJD. Salmonella carriers; are they dangerous ? A study to identify finger contamination with Salmonellae by convalescent carriers. J Infect 1982; 5:81-88.