An Outbreak of Salmonella Typhimurium Phage Type 42 Associated with the Consumption of Raw Flour

Abstract

A cluster of salmonellosis cases caused by Salmonella Typhimurium phage type 42 (STM42) emerged in New Zealand in October 2008. STM42 isolates from a wheat-based poultry feed raw material (broll; i.e., product containing wheat flour and particles of grain) had been identified in the 2 months prior to this cluster. Initial investigations indicated that eating uncooked baking mixture was associated with illness. A case-control study was conducted to test the hypothesis that there was an association between STM42 cases and consumption of raw flour or other baking ingredients. Salmonella isolates from human and non-human sources were compared using pulsed-field gel electrophoresis (PFGE) and multiple-locus variable number tandem repeat analysis (MLVA). Environmental investigations included testing flour and other baking ingredients from case homes, unopened bags of flour purchased from retail stores, and inspection of an implicated flour mill. A case-control study of 39 cases and 66 controls found cases had 4.5 times the odds of consuming uncooked baking mixture as controls (95% confidence interval [CI] 1.6–12.5, p-value 0.001). Examination of individual baking ingredients found that, after adjusting for eggs, flour had an odds ratio (OR) of 5.7 (95% CI 1.1–29.1, p-value 0.035). After adjusting for flour, eggs had an OR of 0.8 (95% CI 0.2–3.4, p-value 0.762). PFGE patterns were identical for all STM42 isolates tested; however, MLVA distinguished isolates that were epidemiologically linked to the cluster. STM42 was recovered from flour taken from four cases' homes, two unopened packs purchased from retail stores and packs from three batches of retrieved (recalled) product. This outbreak was associated with the consumption of uncooked baking mixture containing flour contaminated with STM42. The implicated flour mill initiated a voluntary withdrawal from sale of all batches of flour thought to be contaminated. Media releases informed the public about implicated flour brands and the risks of consuming uncooked baking mixture.

Introduction

C haracterization of Salmonella is vital for identification, tracking, and intervention during outbreaks. Traditionally, the characterization of Salmonella Typhimurium has relied on phage typing (Anderson et al., 1977; Rabsch, 2007) and pulsed-field gel electrophoresis (PFGE) (Fisher and Threlfall, 2005; Gerner-Smidt et al., 2006). These methods have provided valuable epidemiologic data and assisted in outbreak investigations for many years, but they have recognized limitations such as limited discriminatory power. Multiple-locus variable number tandem repeat analysis (MLVA) is a molecular typing method characterizing bacterial isolates based on the variation in tandem repeat sequences at multiple loci (Lindstedt, 2005) and has been successfully applied to many bacterial species, including Salmonella Typhimurium (Lindstedt et al., 2004; Lindstedt et al., 2007).

Surveillance of Salmonella in New Zealand is achieved through laboratory and public health collaboration. The Enteric Reference Laboratory (ERL) at the Institute of Environmental Science and Research (ESR) provides a national Salmonella reference service where presumptive Salmonella species isolated from human, animal, food, and environmental sources are confirmed then characterized through serotyping and phage typing. Non-human Salmonella isolates are submitted to ERL from food testing laboratories and include samples from primary producers that are analyzed as part of the National Microbiology Database. As salmonellosis is a notifiable disease, doctors and laboratories are required to notify all cases to local public health units, who are responsible for the public health management of human salmonellosis. Case information is recorded on the national surveillance database EpiSurv and analyzed both regionally (by individual public health units) and nationally (by ESR).

The ERL issued an alert to public health units in November 2008 regarding a cluster of 10 Salmonella Typhimurium phage type 42 (STM42) isolates that had been identified from the South Island of New Zealand. STM42 is endemic in New Zealand though infection is not commonly confirmed in humans. Less than one human case of STM42 infection per week is expected at the same time of year as these cases occurred. Historically, non-human STM42 have been isolated predominantly from bovine sources; however, STM42 had been isolated from poultry feed raw material in the 2 months prior to the first cases (ESR, 2008). A review of the literature did not identify any previous outbreaks of STM42 in New Zealand or elsewhere.

As cases were reported from geographically distinct parts of the South Island, a multi-jurisdictional outbreak investigation was initiated. Food histories were obtained from initial cases, and additional questions were asked specifically about the consumption of chicken and eggs. These data were compared with control data from a previous outbreak investigation, and the consumption of uncooked cake and baking mixture was found to be associated with illness. We describe the laboratory, environmental, and case investigations that confirmed that the source of this outbreak was flour consumed in uncooked home baking mixtures.

Materials and Methods

This multi-jurisdictional outbreak was investigated through the collaboration of public health units, the New Zealand Food Safety Authority (now part of the Ministry for Primary Industries), and ESR food and reference laboratories, with the ESR public health surveillance team coordinating activities and providing epidemiologic services. Regular meetings by teleconference enabled information to be shared between all agencies involved.

Hypothesis generation

In the 2 weeks prior to the first case notifications, STM42 was isolated from poultry feed raw materials. This was a similar pattern to that seen prior to a Salmonella Mbandaka outbreak in early 2008, which was (tentatively) linked to poultry and egg consumption (ESR, unpublished data). Due to the finding of STM42 in poultry feed, our initial hypothesis was that consumption of chicken or egg products may have been associated with STM42 infection. The case questionnaire developed for a case-control study to investigate the outbreak of Salmonella Mbandaka was used initially to interview STM42 cases. Control data from the Salmonella Mbandaka outbreak case-control study was compared with data collected from the initial STM42 cases.

Case definition

Cases were identified through notifications received by public health units following the laboratory diagnosis of salmonellosis in a patient and typing of the isolate by the ERL. A case was defined as a person in New Zealand with symptoms of vomiting or diarrhea reported to a public health unit from October 13, 2008 to January 28, 2009 from which STM42, MLVA type 3-15-NA-NA-311, had been isolated.

Case-control study

A STM42-specific case-control study was conducted to test the hypothesis that cases were no more likely to have been exposed to raw baking mixture than controls. The case-control study questionnaire was administered in addition to standard questionnaires used by public health units to investigate salmonellosis cases. Cases included in the hypothesis generating interviews were also included in the case-control study.

Our initial control recruitment protocol aimed to recruit three controls per case matched by age group and, for adult cases only, also by sex. Controls were recruited through progressive digit dialing on the telephone; however, recruitment was slow, and the impending holiday period and urgent nature of the investigation required recruiting friends of cases to act as controls.

Controls were interviewed by telephone using the same questionnaire as that used with cases. All interviewers attended a training teleconference prior to conducting interviews. If the interview was being conducted within 3 weeks of onset of the corresponding case, the control was asked about exposures during the 5 days prior to their case's onset date. If the control interview was being conducted greater than 3 weeks after the onset date of their corresponding case, the control was asked about exposures in the previous 5 days. Controls interviewed using the previous 5-day exposure period were asked an additional question to determine if exposures were likely to have been significantly different during their corresponding case's exposure period. Controls were excluded if they had been overseas during the period of interest, had diarrhea or vomiting during the 5-day period, or had been diagnosed with salmonellosis in the 28 days prior to the period of interest.

Environmental investigation

The New Zealand Food Safety Authority traced-back the source of poultry feed raw material from which STM42 had been isolated. Authorized officers conducted an inspection of the implicated grain mill and collected surface swabs and flour samples. Samples of flour and other baking ingredients were collected from cases' homes where available. Packages of flour of the same brands and batch dates as those used by cases were purchased from retail outlets by Food Act Officers. A random selection of packages of flour from the recalled product was also tested. All environmental and food samples were submitted for testing at the Public Health Laboratory at ESR.

Laboratory investigations

Food samples were analyzed for Salmonella by the Public Health Laboratory at ESR using standard cultural methods (APHA, 2001). Analytical unit size varied, depending on the amount of sample available for testing, but was generally 25–300 g. In order to estimate the concentration of Salmonella present in the flour, one sample of flour from a case's home and two retail packs in which Salmonella was detected were retested using a Most Probable Number (MPN) technique.

Salmonella isolates received by ERL were serotyped using the White-Kauffmann-Le Minor scheme (Grimont and Weill, 2007) and then subtyped by phage typing (Callow, 1959; Anderson et al., 1977). Isolates from the first eight cases, poultry feed samples from October 2008 (n=3) and a bovine source isolated in January 2008 (n=1) were characterized by single-enzyme PFGE using the standardized laboratory PulseNet protocol (Ribot et al., 2006; CDC, 2011). These initial 12 isolates and all isolates from cases and food samples collected during the investigation were also characterized by MLVA (Lindstedt et al., 2004). All MLVA results were reported as a string of five numbers (STTR9-STTR5-STTR6-STTR10pl-STTR3) using nomenclature described previously (Larsson et al., 2009).

Statistical analysis

Data were entered into EpiData (version 3.1) in duplicate, and entries were compared and corrected as appropriate. Odds ratios (ORs) were calculated using STATA (version 9), and logistic regression was used to calculate adjusted ORs. Charts and tables were generated using MSExcel. Where a participant answered unsure or did not answer a question, the participant was excluded from the analysis of that question only. Ethnicity was collected and prioritized to a single ethnicity using the New Zealand health sector protocol (Ministry of Health, 2004).

Results

Summary of outbreak cases

From October 13, 2008 to January 28, 2009, a total of 75 cases of salmonellosis caused by STM42 were notified, of which 67 met the case definition. The epidemic curve of cases investigated as part of this outbreak is shown in Figure 1. Cases meeting the case definition had onset dates ranging from October 12, 2008 to January 20, 2009 (onset date was unknown for six cases). A total of 12 cases were hospitalized; no cases died. Cases were aged from 11 months to 86 years, with a median age of 19 years; 16 were male, and 51 were female. Case symptoms continued for 2–21 days (median 7 days) and included diarrhea (98%, 58 cases), abdominal pain (78%, 46 cases), fever (58%, 34 cases), headache (49%, 29 cases), chills (49%, 28 cases), nausea (41%, 24 cases), vomiting (32%, 19 cases), muscle aches (32%, 19 cases), and bloody diarrhea (5%, 3 cases). Symptoms were unknown for eight cases. The majority of outbreak-associated cases were from the South Island of New Zealand (82%), compared with half of non-outbreak-associated STM42 cases occurring at the same time.

FIG. 1.

Epidemic curve of outbreak- and non-outbreak-associated Salmonella Typhimurium phage type 42 cases notified in New Zealand, October 13, 2008 to January 28, 2009 by onset date, showing key events. Onset date was unknown for six outbreak cases and four non-outbreak cases. MLVA, multiple-locus variable number tandem repeat analysis.

Hypothesis generation

A comparison of 19 STM42 case interviews with 63 controls recruited for a Salmonella Mbandaka case-control study conducted in March–April 2008 found that cases had 12.5 times the odds of eating uncooked cake or pancake batter than controls (adjusted for age, sex, and public health unit, 95% confidence interval [CI] 0.8–201.6, p-value 0.08). Cases and controls were also asked a range of questions regarding chicken and egg consumption, none of which came close to significance (adjusted ORs of 0.5–2.2, p-values 0.31–0.93).

STM42 case-control study

A total of 39 cases and 66 controls were included in the case-control study; there were no significant differences between cases and controls with respect to age, sex, ethnicity, or geographic location (data not shown). Exposures related to home baking or contact with uncooked flour were examined; univariate associations between exposure and illness are shown in Table 1. As we suspected that either eggs or flour were the likely source of contamination, we calculated ORs for each of these, adjusting for the other. After adjusting for eggs, flour had an OR of 5.7 (95% CI 1.1–29.1, p-value 0.035). After adjusting for flour, eggs had an OR of 0.8 (95% CI 0.2–3.4, p-value 0.762). As the other baking ingredients were not suspected sources and were highly correlated, we did not pursue further multivariable analysis. Brand A flour, flour purchased from supermarket A, and plain flour were significantly associated with illness (Table 1).

Table 1.

Comparison of Case and Control Responses to Questions Regarding Baking and the Consumption of Uncooked Baking Mixture and Baking Ingredients, During an Investigation into an Outbreak of Salmonella Typhimurium Phage Type 42 in New Zealand, October 2008 to January 2009

Exposure	Cases exposed	Controls exposed	Odds ratio (95% CI)	p-value
Did you or someone in your household do any baking?	25 (78%)	35 (55%)	3.0 (1.0–9.2)	0.025
Did you eat lick or taste any uncooked baking mixture?	18 (58%)	15 (23%)	4.5 (1.6–12.5)	0.001
Ingredients
Flour	30 (86%)	33 (54%)	5.1 (1.6–18.8)	0.002
Eggs	25 (74%)	30 (50%)	2.8 (1.0–7.9)	0.026
Baking powder	27 (79%)	25 (42%)	5.2 (1.8–16.3)	0.001
Sugar	28 (82%)	31 (53%)	4.2 (1.4–14.1)	0.004
Corn flour	9 (26%)	6 (10%)	3.1 (0.9–11.8)	0.043
Wheat bran	2 (6%)	1 (2%)	3.6 (0.2–213.8)	0.278
Cocoa powder	19 (58%)	19 (32%)	2.9 (1.1–7.6)	0.018
Butter	24 (71%)	25 (44%)	3.1 (1.1–8.5)	0.013
Milk	24 (71%)	29 (50%)	2.4 (0.9–6.6)	0.054
Flour purchased from
Supermarket A	15 (44%)	9 (15%)	4.4 (1.5–13.2)	0.002
Supermarket B	12 (35%)	16 (27%)	1.5 (0.5–4.0)	0.408
Supermarket C	5 (15%)	5 (8%)	1.9 (0.4–8.8)	0.350
Supermarket D	1 (3%)	1 (2%)	1.8 (0.0–140.3)	0.690
Brand of flour
Brand A	18 (62%)	8 (14%)	10.0 (3.1–33.3)	<0.001
Brand B	2 (7%)	4 (7%)	1.0 (0.1–7.4)	0.983
Brand C	5 (17%)	3 (5%)	3.8 (0.7–25.7)	0.071
Brand D	1 (3%)	4 (7%)	0.5 (0.0–5.1)	0.504
Other brand	2 (7%)	8 (14%)	0.5 (0.0–2.5)	0.329
Type of flour
Self-raising	4 (12%)	6 (10%)	1.2 (0.2–5.3)	0.833
Plain	21 (62%)	18 (31%)	3.6 (1.4–9.6)	0.004
High grade	9 (26%)	8 (14%)	2.3 (0.7–7.5)	0.131
Wholemeal	2 (6%)	3 (5%)	1.1 (0.1–10.5)	0.885

CI, confidence interval.

Laboratory investigations

PFGE patterns from 12 STM42 isolates from both human and non-human sources were indistinguishable. By contrast, MLVA data from all STM42 isolates from both human and non-human sources collected from January to November 2008 demonstrated differences in the STTR5 locus, with the majority of isolates from cases in the South Island from October 2008 onwards and isolates from the poultry feed raw material isolated in September and October 2008 having allele 15 and isolates from other parts of New Zealand and from other non-human sources having a variety of different alleles. Additional details of MLVA profiles from isolates from this outbreak are published elsewhere (Dyet et al., 2011)

Environmental investigation

Trace-back of the poultry feed raw material determined that it contained broll (i.e., product containing wheat flour and particles of grain) that had come from a grain mill that also produced products for human consumption. STM42 was not detected in environmental samples taken from the mill premises but another Salmonella serovar (Salmonella Typhimurium phage type 160) was detected in a swab taken from the broll load-out shoot.

A total of 63 food samples taken from cases' homes (complaint samples), 41 flour samples from retail premises and 63 samples of retrieved product, were tested. Flour from the homes of four cases tested positive for the STM42 outbreak strain. The outbreak strain was also isolated from two packs of unopened flour from retail stores and three batches of retrieved product.

Estimates of the level of contamination of flour in unopened packs ranged from approximately 1 per 300 g (0.0036 MPN/g [95% CI 0.00051–0.025]) to 1 per 50 g (0.024 MPN/g [95% CI 0.0067–0.085]), and in one of the complaint samples, 1 per 100 g (0.012 MPN/g [95% CI: 0.0017–0.081]). All contaminated flour batches were produced over a 4-week period during September and October 2008.

Discussion

This outbreak of STM42 infection was caused by consumption of contaminated flour (in the form of raw baking mixture). This appears to be the first time contaminated flour has been conclusively identified, by both laboratory and epidemiologic investigation, as a vehicle for Salmonella infection in New Zealand and internationally. Though the case-control study results indicated that Brand A flour, flour purchased from supermarket A, and plain flour were significantly associated with illness, most cases did not have packaging available to determine the batch dates of implicated flour and limited the usefulness of this information.

Foods containing uncooked flour have also been implicated in previous bacterial foodborne illness outbreaks. Raw cake mix used as an ingredient in ice cream was epidemiologically implicated in a Salmonella Typhimurium outbreak in the United States in 2005 (Zhang et al., 2007), and uncooked pre-packaged cookie dough was identified as the vehicle for a multistate outbreak of Escherichia coli O157:H7 illness in the United States in 2009 (Neil et al., 2012). No specific ingredient was identified as the definitive source in these outbreaks; however, contaminated flour was suspected. Outbreaks associated with the consumption of chemically contaminated flour have also been reported (Davies and Lewis, 1956; Weeks, 1967; Stewart et al., 1969; Richter et al., 2000; Kakar et al., 2010).

Microbial surveys of flour and other wheat-based end products have rarely identified Salmonella (Berghofer et al., 2003); no published data is available for New Zealand milled flour. Dry flour (less than 12% moisture) is microbiologically stable and will not support the growth of pathogenic or spoilage microorganisms (ICMSF, 2005). Salmonella can, however, survive for long periods in low-moisture environments and, if present in flour or other low-–water activity foods, may remain viable for several months (ICMSF, 1996, 2005).

Methods used to detect Salmonella in food typically sample 25 g of food. The concentration of Salmonella present in flour implicated in this outbreak was estimated at between 1 organism per 50 g to 1 organism per 300 g. Larger analytical units may be required when investigating outbreaks associated with low-level contamination that may be clinically relevant.

For the first time in New Zealand, MLVA typing was used to aid in an outbreak investigation where standard methods failed to discriminate between outbreak-related and sporadic STM42 case isolates. No amplicons were detected for STTR6 or STTR10, which was not unexpected as STTR6 is missing from approximately 16% of all Salmonella Typhimurium and STTR10 is missing from approximately 48% of isolates tested (Lindstedt et al., 2004). The significance of differences in MLVA profiles has not yet been agreed on internationally. In this study, a difference at the STTR5 locus compared to the outbreak profile was significant. Epidemiologic investigations concluded that isolates that had a MLVA profile that differed from the outbreak profile at the STTR5 locus were not related to the outbreak. The ability to discriminate between isolates that were part of the outbreak using MLVA was critical to the investigation, as it allowed us to narrow our hypothesis at an early stage. Isolates from poultry feed raw materials that had been identified prior to the first human cases had the outbreak MLVA profile, and STM42 isolates from other environmental sources had different MLVA profiles. As we were also able to restrict the investigation of STM42 cases to those with the outbreak MLVA profile, we were able to minimize the impact of sporadic cases diluting the effect of exposures of interest in the case-control study. MLVA typing was usually performed within 48 h of the identification of an STM42 isolate, reducing the burden of unnecessary investigation of cases and control recruitment. As MLVA typing was performed by a different laboratory, ERL did not suffer increased work pressure due to the additional typing requirements of this investigation.

There are some limitations to this study. The case-control study used two control recruitment methods as initial attempts at using progressive digit dialing were very slow and the impending holiday period necessitated timely control recruitment. The use of friend controls may bias the OR estimates towards the null, as friend controls would be more similar with respect to exposures than randomly selected controls. We believed that friends of cases were unlikely to have similar behaviour with respect to eating uncooked cake and baking mixture so decided that the risk of bias was acceptable for this study. Salmonella isolates from non-human sources are submitted to the ERL on a voluntary basis and therefore are not representative of all non-human sources.

Conclusion

This outbreak was associated with the consumption of uncooked baking mixture containing flour contaminated with STM42. The mill voluntarily withdrew from sale potentially affected batches of flour covering a 2-month period. A media statement giving details of the implicated flour brands was released. Food safety advice was given to consumers regarding the consumption of uncooked food, in particular raw baking mixtures.

Footnotes

Acknowledgments

We wish to acknowledge the work of the staff of the Enteric Reference and Public Health at ESR who conducted the laboratory investigations, the staff at the New Zealand Food Safety Authority who coordinated the environmental investigation, staff at the public health units that conducted the case investigations and control interviews and collected samples, and the cases and controls that provided the information that was vital in identifying the cause of this outbreak.

Disclosure Statement

No competing financial interests exist.

References

Anderson

, Ward

, Saxe

, de Sa

. Bacteriophage-typing designations of Salmonella Typhimurium. J Hyg (Lond), 1977; 78:297–300.

[APHA] American Public Health Association. Compendium of Methods for the Microbiological Examination of Foods. Washington, DC: APHA, 2001.

Berghofer

, Hocking

, Miskelly

, Jansson

. Microbiology of wheat and flour milling in Australia. Int J Food Microbiol, 2003; 85:137–149.

Callow

. A new phage-typing scheme for Salmonella Typhimurium. J Hyg (Lond), 1959; 57:346–359.

[CDC] Centers for Disease Control and Prevention. Standardized molecular subtyping of foodborne bacterial pathogens by pulsed-field gel electrophoresis. Atlanta: The Centers, 2011. http://www.cdc.gov/pulsenet/protocols.htm. 2012 December 1.

Davies

, Lewis

. Outbreak of food-poisoning from bread made of chemically contaminated flour. Br Med J, 1956; 2:393–398.

Dyet

, Turbitt

, Carter

. Multiple-locus variable-number tandem-repeat analysis for discriminating within Salmonella enterica serovar Typhimurium definitive types and investigation of outbreaks. Epidemiol Infect, 2011; 139:1050–1059.

[ESR] Institute of Environmental Science and Research. Non-human Salmonella isolates, 2008. Public Health Surveillance. 2008. http://www.surv.esr.cri.nz/enteric_reference/nonhuman_salmonella.php?we_objectID=1950. 2012 December 1.

Fisher

, Threlfall

. Enter-net; Salm-gene. The Enter-net and Salm-gene databases of foodborne bacterial pathogens that cause human infections in Europe and beyond: An international collaboration in surveillance and the development of intervention strategies. Epidemiol Infect, 2005; 133:1–7.

10.

Gerner-Smidt

, Hise

, Kincaid

, Hunter

, Rolando

, Hyytiä-Trees

, Ribot

, Swaminathan

. Pulsenet Taskforce. PulseNet USA: A five-year update. Foodborne Pathog Dis, 2006; 3:9–19.

11.

Grimont

PAD

, Weill

F-X

. Antigenic Formulae of the Salmonella Serovars, 9 ^th. Institut Pasteur/World Health Organization (WHO) Collaborating Centre for Reference and Research on Salmonella, 2007. http://www.pasteur.fr/ip/portal/action/WebdriveActionEvent/oid/01s-000036-089. 2012 December 1.

12.

[ICMSF] International Commission on Microbiological Specifications for Foods. Microorganisms in Foods 5: Characteristics of Microbial Pathogens. London: Blackie Academic & Professional, 1996.

13.

[ICMSF] International Commission on Microbiological Specifications for Foods. Microorganisms in Foods 6. Microbial Ecology of Food Commodities, 2 ^nd. New York: Kluwer Academic & Plenum Publishers, 2005.

14.

Kakar

, Akbarian

, Leslie

, Mustafa

, Watson

, van Egmond

, Omar

, Mofleh

. An outbreak of hepatic veno-occlusive disease in Western Afghanistan associated with exposure to wheat flour contaminated with pyrrolizidine alkaloids. J Toxicol, 2010; 2010:313280.

15.

Larsson

, Torpdahl

, Petersen

, Sorensen

, Lindstedt

, Nielsen

. Development of a new nomenclature for Salmonella Typhimurium multilocus variable number of tandem repeats analysis (MLVA) Euro Surveill, 2009; 14:pii–19174.

16.

Lindstedt

. Multiple-locus variable number tandem repeats analysis for genetic fingerprinting of pathogenic bacteria. Electrophoresis, 2005; 26:2567–2582.

17.

Lindstedt

, Vardund

, Aas

, Kapperud

. Multiple-locus variable-number tandem-repeats analysis of Salmonella enterica subsp. enterica serovar Typhimurium using PCR multiplexing and multicolor capillary electrophoresis. J Microbiol Methods, 2004; 59:163–172.

18.

Lindstedt

, Torpdahl

, Nielsen

, Vardund

, Aas

, Kapperud

. Harmonization of the multiple-locus variable-number tandem repeat analysis method between Denmark and Norway for typing Salmonella Typhimurium isolates and closer examination of the VNTR loci. J Appl Microbiol, 2007; 102:728–735.

19.

Ministry of Health. Ethnicity Data Protocols for the Health and Disability Sector. Wellington, New Zealand: Ministry of Health, 2004. http://www.fmhs.auckland.ac.nz/faculty/tkhm/tumuaki/_docs/ethnicity-data-protocols.pdf. 2012 December 1.

20.

Neil

, Biggerstaff

, MacDonald

, Trees

, Medus

, Musser

, Stroika

, Zink

, Sotir

. A novel vehicle for transmission of Escherichia coli O157:H7 to humans: Multistate outbreak of E. coli O157:H7 infections associated with consumption of ready-to-bake commercial prepackaged cookie dough—United States, 2009. Clin Infect Dis, 2012; 54:511–518.

21.

Rabsch

. Salmonella Typhimurium phage typing for pathogens. Methods Mol Biol, 2007; 394:177–211.

22.

Ribot

, Fair

, Gautom

, Cameron

, Hunter

, Swaminathan

, Barrett

. Standardization of pulsed-field gel electrophoresis protocols for the subtyping of Escherichia coli O157:H7, Salmonella, and Shigella for PulseNet. Foodborne Pathog Dis, 2006; 3:59–67.

23.

Richter

, El-Sharif

, Fischbein

, Konijn

, Gorodetsky

, El-Sharif

, Kaul

, Hershko

, Grauer

, Foner

, Al-Baba

, Dweik

, Lihsounat

. Re-emergence of lead poisoning from contaminated flour in a West Bank Palestinian village. Int J Occup Environ Health, 2000; 6:183–186.

24.

Stewart

, Sherman

, Politzer

. An outbreak of food-poisoning due to a flour improver, potassium bromate. S Afr Med J, 1969; 43:200–202.

25.

Weeks

. Endrin food-poisoning. A report on four outbreaks caused by two separate shipments of endrin-contaminated flour. Bull World Health Organ, 1967; 37:499–512.

26.

Zhang

, Ma

, Patel

, Swaminathan

, Wedel

, Doyle

. Isolation of Salmonella Typhimurium from outbreak-associated cake mix. J Food Prot, 2007; 70:997–1001.