ARCHIVED - Outbreak of Salmonella Typhimurium phage type U302 in Ontario, spring 2005

 

Canada Communicable Disease Report

1 April 2006

Volume 32
Number 07

Introduction

In April 2005, the Ontario Central Public Health Laboratory (CPHL) identified 55 cases of Salmonella Typhimurium, an increase from the average number of reports of 36 (range: 32 to 41) for the same month between 2002 and 2004 (unpublished data, CPHL). As well, the National Microbiology Laboratory (NML) identified more of phage type (PT) U302 than expected among S. Typhimurium cases in Ontario during this time. This particular phage type is relatively uncommon in Canada but has been reported sporadically in Ontario 1.

On 17 May, 2005, the Ontario Ministry of Health and Long-Term Care, with the assistance of the Canadian Field Epidemiology Program and the Foodborne, Waterborne and Zoonotic Infections Division of the Public Health Agency of Canada, launched an epidemiologic investigation into the increased reports of this uncommon S. Typhimurium phage type.

Methods

Case definitions were developed to assist in case finding.

  • Confirmed primary case: a person with laboratory-confirmed S. Typhimurium PT U302 reported in Ontario between 1 March and 31 May, 2005.

  • Confirmed secondary case: a person with laboratoryconfirmed S. Typhimurium PT U302 with symptom onset occurring more than 72 hours (i.e. longest incubation period of Salmonella) after the onset of symptoms in a confirmed case in the same household.

Demographic and clinical information on confirmed cases was provided by the health units in the areas where the cases resided.

Interviews were conducted with nine confirmed cases to generate hypotheses for source and transmission. Cases were selected for interview according to geographic and age distribution. Those whose symptoms began in April 2005 were selected to optimize food history recall. The interview questionnaire included items on demographic factors, symptoms and duration of illness, hospitalization, contact with persons with diarrheal illness, recent travel, and a detailed food history, including store of purchase and retail brand, for the 5 days before symptoms began. New food items reported by cases were added to subsequent interviews. When available, local health inspectors were present during in-home interviews to assist with refrigerator and cupboard inspections for possible food exposures.

A matched case-control study was designed to test hypotheses. Confirmed primary cases not previously interviewed for the hypotheses-generation phase were eligible for the case-control study. Controls were matched according to age and postal code of residence. One control per case was selected using random telephone numbers generated from all telephone numbers listed in the same forward sortation area (first three digits of postal code) as the case.

Telephone numbers were attempted up to four times at various times of the day and week before being discarded. The age groups selected for matching (1 to 4 years, 5 to 14 years, 15 to 19 years, ≥ 20 years) were based on the age distribution of the cases and likelihood of food preferences. Controls were excluded if they had experienced diarrheal illness in the 72 hours before the onset of symptoms of their matched case. Adult guardians were interviewed to collect data for cases and controls < 16 years of age. Control questionnaires excluded data on illness; food history applied to the previous 5 days or “usual” consumption.

Data entry, validation, and analysis for the case-control study were performed using EpiInfo(2). A matched analysis was used to identify food items significantly associated with illness. Variables that had a p value of < 0.2 in the univariate analysis were included in a conditional logistic regression analysis. A final model was obtained through a forward stepwise procedure, retaining only significant variables.

Results

Between 1 March and 31 May, 2005, 47 confirmed cases of S. Typhimurium PT U302 were identified in Ontario. Pulsed-field gel electrophoresis (PFGE) patterns were provided by CPHL for 87% of cases (41/47). Two cases were excluded from the investigation because their PFGE profiles were sufficiently different from the most common profile that they were likely not part of the outbreak.

Of the 45 cases with similar PFGE, two were secondary cases. The epidemic curve is shown in Figure 1. The pattern of the epidemic curve is suggestive of a common source exposure.

Figure 1. Cases of Salmonella Typhimurium PT U302 in Ontario by date of onset, 1 March to 31 May, 2005 (n = 45)

Figure 1. Cases of Salmonella Typhimurium PT U302 in Ontario by date of onset, 1 March to 31May, 2005 (n = 45)

Demographic information was available for 98% of the confirmed cases (44/45). Half were male (22/44), and the mean age was 20.2 years (median 14.0 years, range 1 to 75 years). Fifty-two percent (23/44 for whom this information was available) were < 18 years of age.

Clinical information was available for 93% of confirmed cases (42/45). All but one reported diarrhea, and 52% (22/42) reported bloody diarrhea. Other reported symptoms included fever (88%), abdominal cramping (79%), headache (45%), nausea (36%), and vomiting (24%). Nine cases were hospitalized for their illness (range of hospital stay 1 to 5 days). There were no deaths.

Twelve health units in southern Ontario reported confirmed cases. The largest proportion (17/45, 38%) was from York Region, followed by Toronto (10/45, 22%); between one and three cases were reported from each of the following health regions: Durham, Halton, Hamilton, Middlesex-London, Ottawa, Peel, Renfrew, Simcoe Muskoka, Waterloo, and Wellington-Dufferin.

Hypothesis-generating interviews conducted between 20 and 29 May revealed that eight of nine cases had been exposed to one of mortadella, salami, prosciutto, or capicollo in the 5 days preceding their illness. Although most of these foods were reportedly purchased at either large chain grocery stores or local delicatessens, several cases reported purchasing such products from the same manufacturer.

Between 30 May and 20 June, 32 case interviews and 30 control interviews were conducted. Three eligible cases could not be reached during the case-control study. Controls could not be recruited for two cases in the 2 to 4-year age group. Therefore a total of 30 case-control pairs were included in the analysis.

The matched odds ratios (MOR) for individual food exposures are shown in Table 1. Illness was significantly related to the consumption of salami (MOR = 3.75, 95% confidence interval [CI]: 1.2 to 11.3). Analysis using combined deli meat exposures showed that 87% of cases (26/30) reported eating either salami, mortadella, or prosciutto, compared with 40% of controls (12/30) (MOR = 8.0, 95% CI: 1.8 to 34.8).

Brand recall varied according to product. Forty-four percent of cases or controls (11/25) who reported consuming salami were unable recall the brand, although one-third (8/25) could recall consuming Genoa salami. In comparison, 83% of cases or controls (15/18) who reported eating mortadella could recall the brand. Ten of the 11 cases who were able to recall the brand of mortadella reported that it was made by Producer A. As shown in Table 1, exposure to salami, mortadella, or prosciutto from Producer A was significantly associated with illness (MOR = 11.0, 95% CI: 1.4 to 85.2). This combined variable remained significantly related to illness in a multiple logistic regression model. No individual food exposures were associated with illness when included in the same model as this combination variable.

Table 1. Matched odds ratios (univariate) for selected food exposures in a Salmonella Typhimurium outbreak, Ontario, March to May 2005

Individual food exposure

Cases
(n = 30)

Controls
(n = 30)

Matched
odds ratio

95% confidence
interval

p value

Romaine lettuce 16 17 0.83 0.25-2.73 0.55
Other lettuce type 17 17 1.00 0.35-2.85 0.79
Spinach 6 7 0.83 0.25-2.73 0.55
Bean sprouts 2 2 1.00 0.14-7.10 0.62
Fresh parsley 8 6 1.50 0.42-5.32 0.75
Fresh basil 4 4 1.00 0.20-4.95 0.68
Almonds 6 10 0.56 0.19-1.66 0.18
Peanuts 5 8 0.57 0.17-1.95 0.23
Cantaloupe 10 16 0.45 0.16-1.31 0.08
Raspberries 7 10 0.57 0.17-1.95 0.23
Raw or runny eggs 7 5 1.50 0.42-5.32 0.75
Items containing raw or runny eggs 1 2 0.50 0.05-5.51 0.25
Ground beef 19 21 0.71 0.23-2.25 0.39
Steak 12 16 0.60 0.22-1.65 0.21
Veal 12 6 2.20 0.76-6.33 0.21
Pork chop or loin 12 19 0.36 0.12-1.14 0.039
Fresh parmesan cheese 18 16 1.40 0.44-4.41 0.77
Provolone cheese 9 4 2.67 0.71-10.05 0.23
Ricotta cheese 4 2 Undefined    
Bocconcini 10 4 Undefined    
Kielbasa 6 4 1.50 0.42-5.32 0.75
Sausage 12 9 1.60 0.52-4.89 0.58
Ham 10 15 0.44 0.14-1.44 0.096
Pepperoni 8 7 1.17 0.39-3.47 1.00
Pepperettes 4 1 Undefined    
Capicollo 3 1 3.00 0.31-28.84 0.62
Salami 18 7 3.75 1.24-11.30 0.022
Mortadella 11 7 2.33 0.60-9.02 0.34
Prosciutto 10 4 4.00 0.85-18.84 0.11

Combined food exposures

         
Salami or mortadella 22 11 3.75 1.24-11.30 0.022
Salami, mortadella, or prosciutto 26 12 8.00 1.84-34.79 0.002
Salami, mortadella, prosciutto or capicollo 26 12 8.00 1.84-34.79 0.002
Any of salami, mortadella, or prosciutto 16 8 3.67 1.09-13.14 0.061
sliced at the deli counter          

Combined food exposures by brand

         
Salami from Producer A 3 0 Undefined    
Mortadella from Producer A 10 4 7.00 0.86-56.90 0.077
Prosciutto from Producer A 4 0 Undefined    
Capicollo from Producer A 0 0    
Any of salami, mortadella, or prosciutto from Producer A 14 4 11.00 1.42-85.20 0.009
Any of salami, mortadella, or prosciutto from Producer A sliced at the deli counter 11 4 8.00 1.00-63.82 0.050

Given the relatively low proportion of cases (14/30, 47%) reporting exposure to foods made by Producer A and the inability to identify a specific lot of product(s) for investigation, a formal traceback investigation was not conducted. A review by the Canadian Food Inspection Agency of Producer A found no consumer complaints or employee illness, and the three plants where the implicated foods were produced and prepared were found to have complied with existing regulations.

Discussion

This widespread outbreak of an uncommon phage type of S. Typhimurium involving 45 cases reported by 12 health units in southern Ontario occurring in a 3-month period (March-May 2005) was identified through laboratory-based surveillance using serotype and phage typing.

The pattern of the epidemic curve, the long period over which the reports of confirmed cases increased, the geographic distribution of cases, and the lack of an epidemiologic link between cases supports the hypothesis that the outbreak was due to a common source rather than a point source (e.g. food served at a large gathering). The common source could likely be a food product with a long shelf life that was widely distributed across southern Ontario. We suspect that it may have been a ready-to-eat item that did not require cooking, since food-borne infection with Salmonella species can usually be prevented with adequate refrigeration and cooking temperatures, and proper handwashing and food preparation practices 3.

The most likely hypothesis supported by the findings of this epidemiologic investigation was that illness was associated with salami, mortadella, and/or prosciutto made by a single producer. There was, however, no confirmatory microbiological evidence, and while cases were approximately 11 times more likely than controls to have eaten one of the three suspected deli meats from Producer A, only 14 of the 30 cases reported eating the suspected brand prior to onset of symptoms. This inconsistency may be related to cross-contamination at the retail level. For example, pathogenic bacteria from a contaminated product may remain on an infrequently cleaned slicing machine, thereby contaminating other meats subsequently sliced on the machine 4. Recall bias was also an important factor, as many study participants were unable to recall the brand of deli meat consumed, particularly for salami. This may occur in part because consumers receive products from the deli counter as labelled by the store, rather than with recognizable brand packaging.

Delays in the investigation may have also affected recall, as well as timely case finding and the availability of food items for microbiological testing. The increase in S. Typhimurium isolates was initially noted as a slight increase in the normal consecutive weekly totals for April 2005. The relatedness of the strains was not identified until 5 weeks after the peak of the outbreak, when typing results (phage typing and PFGE) became available. The initial identification of this outbreak relied on phage typing, which is not routinely performed in Ontario and is conducted at the NML. The time lag between outbreak identification and the investigation's hypotheses-generating interviews (range 7 to 16 days) compounded the investigative delay and made accurate recall difficult for cases. Case finding was impeded by the submission of incomplete case information to CPHL by some private laboratories.

These time constraints could be improved with the standardization of reporting formats of enteric pathogens for all private, hospital, and provincial laboratories in conjunction with the full implementation of real-time surveillance through the Integrated Public Health Information System (iPHIS) in Ontario, scheduled for fall 2005, and the implementation of a comprehensive laboratory information system at the CPHL. Increased resources for provincial and national laboratories might also improve the timeliness of molecular testing, which is a critical component of identifying outbreaks and separating outbreak-related cases from background cases.

S. Typhimurium is the most frequently isolated serovar in Canada, accounting for approximately 20% of all human Salmonella isolates reported to the National Enteric Surveillance Program between 2001 and 2004. However, the phage type particular to this outbreak generally circulates at low levels in Canada. Data collected through the Canadian Integrated Program for Antibiotic Resistance Surveillance for the years 2003 and 2004 show that PT U302 accounted for 3.5% of S. Typhimurium isolates (W. Demczuk, National Microbiology Laboratory, Winnipeg: personal communication, 2005). In comparison, PT U302 accounted for 36% of S. Tyhimurium cases (9/25) identified in Ontario in March and 62% of cases (34/55) reported in April 2005.

This is only the second outbreak of S. Typhimurium PT U302 reported in the published literature. In 2003, an ongoing outbreak comprising 67 confirmed cases from four European countries was traced to an assistant chef working at a restaurant in Copenhagen. The observed low attack rates and long incubation periods (median 4 days, range 1 to 27 days) indicated low-grade contamination of foods, and microbiological investigation revealed S. Typhimurium in food samples from the buffet. The outbreak strain had been previously identified only in unrelated sporadic cases in Denmark 5. The emergence of this phage type should be monitored because it is frequently associated with multi-drug resistance 1,5,6, which can adversely affect the burden of illness 7.

Dry fermented sausages, such as salami, contain raw, ground meat combined with various spices, curing agents, and salt. Through the processes of fermentation, which lowers the pH of the mixture, and drying, pathogenic bacteria are gradually replaced by non-pathogenic flora over a period of days or months. Salami is considered ready to eat and is not cooked before consumption 8,9. Mortadella is a dry sausage that does not undergo fermentation but is smoked at a high temperature before being air-dried. Prosciutto is salted ham that is air-dried to cure. Salami has been previously identified as the food vehicle for S. Typhimurium in two geographically widespread outbreaks in Northern Italy (PT 193) and England (definitive type 124). Investigation of the manufacturing plants in these outbreaks revealed that a lack of microbial starter cultures and/or insufficient ripening time likely contributed to the survival of Salmonella in the fermentation and drying stages 10,11. InMay 2005, there was a small outbreak of illness in Sweden due to S. Typhimurium PT U302 contamination of German mini salami, resulting in a product recall in Sweden, Finland, and the Netherlands 12. There have also been reported cases of salmonellosis related to the consumption of prosciutto in Italy 13, and Salmonella has been detected from samples of mortadella through routine post-production testing in the United States 14.

In Canada, a November 2004 amendment to the Meat Inspection Regulations, 1990 proposed the mandatory implementation of the Food Safety Enhancement Program at all federally registered meat and poultry establishments and storages, including sampling programs for Salmonella and Escherichia coli in order to verify that plant Hazard Analysis and Critical Control Point programs are effective in reducing contamination with these microorganisms. These sampling programs only apply to specific carcasses, ground meats, and fresh pork sausages, where as microbiological evaluation is not required for ready-to-eat sausages, which are dried fermented (e.g. salami, pepperoni, capicollo) or cooked (e.g. mortadella, bologna)15,16. Current recommended practices for ready-to-eat meats in Canada (e.g. monitoring of time/temperature logs, pH testing) may therefore not be sufficient to rule out contamination with pathogenic bacteria.

While deli meats from a single producer were identified as the most likely cause of this outbreak, the evidence obtained through the epidemiologic investigation was not strong enough to support further public health action, such as a food recall. The results of this outbreak investigation demonstrate how the strengths and weaknesses of the laboratory-based surveillance system in Ontario can affect the outcome of an outbreak investigation. Timely investigation of widespread food-borne outbreaks identified through molecular testing could provide public health authorities with sufficient evidence to respond appropriately and to make recommendations to further improve consumer protection.

Acknowledgements

The authors would like to thank the following for their assistance in the investigation: Dr. A. Ellis and M. Taylor, Foodborne, Waterborne and Zoonotics Division, Public Health Agency of Canada (PHAC), Guelph, Ontario; public health inspectors and communicable disease managers from the affected health units, particularly N. Wachowiak, I. Davis, E. Moteka, and M. Kroll, York Region, Ontario; M. Lombos, J. Labelle, A. Maki, and Dr. F. Jamieson, CPHL; R. Ahmed and W. Demczuk, NML; K. Marcynuk, Canadian Food Inspection Agency, Ottawa, Ontario; and S. Johnson, Ministry of Health and Long-term Care (MOHLTC), Toronto, Ontario.

References

  1. Lee MB, Middleton D, Ciebin B. Salmonella Typhimurium DT104 and non-DT104 in Ontario, 2000. Public Health Epidemiol Rep Ontario 2003;14:22-5.

  2. Centers for Disease Control and Prevention, USA. EpiInfo (version 3.3.2). February 2005. URL: <http://www.cdc.gov/epiinfo/>.

  3. Control of communicable diseases manual. 18th ed. Washington: American Public Health Association, 2005.

  4. Elson R, Burgess F, Little CL et al. Microbiological examination of ready-to-eat cold sliced meats and pâté from catering and retail premises in the UK. J Applied Microbiol 2004;96:499-509.

  5. Ethelberg S, Lisby M, Torpdahl M et al. Prolonged restaurant-associated outbreak of multidrug-resistant Salmonella Typhimurium among patients from several European countries. Clin Microbiol Infect 2004;10:904-10.

  6. Threlfall EJ, Teale CJ, Davies RH et al. A comparison of antimicrobial susceptibilities in nontyphoidal Salmonellas from humans and food animals in England and Wales in 2000. Microbial Drug Resistance 2003;9:183-9.

  7. Martin LJ, Fyfe M, Doré K et al. Increased burden of illness associated with antimicrobial-resistant Salmonella enterica serotype Typhimurium infections. J Infect Dis 2004;189: 377-84.

  8. Tilden J, Young W, McNamara AM et al. A newroute of transmission for Escherichia coli: Infection from dry fermented salami. Am J Public Health 1996;86:1142-45.

  9. HACCP generic model: Fermented smoked sausage. Ottawa: Canadian Food Inspection Agency, 1997.

  10. Pontello M, Sodano L, Nastasi A et al. A community-based outbreak of Salmonella enterica serotype Typhimurium associated with salami consumption in Northern Italy. Epidemiol Infect 1998;120:209-14.

  11. Cowden JM, O'Mahony M, Bartlett CLR et al. A national outbreak of Salmonella typhimurium DT 124 caused by contaminated salami sticks. Epidemiol Infect 1989;103:219-25.

  12. Salmonella Typhimurium in mini salami from Germany (Alert notification 2005.291-add01, May 10, 2005). Brussels: European Commission Health & Consumer Protection Directorate-General, 2005.

  13. Mete R, Fichera R, Chierchini P. Episodes of food poisoning caused by Salmonella C2 transmitted via prosciutto [Italian]. Nuovi Ann Ig Microbiol 1987;38:219-24.

  14. New York firm recalls mortadella for possible Salmonella contamination: US Food Safety & Inspection Service Recall Release 00-010. URL: <http://www.fsis.usda.gov/OA/recalls/prelease/pr010-2000.htm>.
    Date of access: 24 Aug 2005.

  15. Regulations amending the meat inspection regulations, 1990, Vol. 138, No. 25 - December 15, 2004. Can Gazette 2004.

  16. Food Safety Enhancement Program: HACCP generic models for fermented smoked sausage and cooked sausage. URL: <http://www.inspection.gc.ca/english/fssa/polstrat/haccp/modele.shtml#1> Date of access: 24 Aug 2005.

Source: C Navarro, MSc, Canadian Field Epidemiology Program, PHAC; D MacDonald, MSc, Foodborne, Waterborne and Zoonotic Infections Division, PHAC; DMiddleton, DVM, MSc, Ontario Ministry of Health and Long-Term Care (MOHLTC), Public Health Branch; L Landry, MSc, Foodborne, Waterborne and Zoonotic Infections Division, PHAC; L Vrbova, MSc, Ontario MOHLTC, Public Health Branch; LY Lior, MD, MSc, Canadian Field Epidemiology Program, Public Health Agency of Canada, Ottawa, Ontario.

Page details

Date modified: