A lArge point-source outbreAk of Salmonella
typhimurium phAge type 9 linked to A bAkery in
sydney, mArch 2007
cases? or? their? guardians? and? active? surveillance?
319? cases? of? gastroenteritis? were? identified,? of?
which? 221? cases? (69%)? presented? to? a? hospital?
emergency? department? and? 136? (43%)? required?
that? foodborne? illness? outbreaks? can? cause? sub-
Foodborne illness accounts for an estimated 32% of
gastroenteritis illness in Australia, causing around
5.4 million episodes of illness, 18,000 hospitalisations
and approximately 120 deaths annually.1 Salmonella
infection (mainly Salmonella Typhimurium) is
predominantly a foodborne illness in Australia with
annual seasonal peaks in the summer months.2,3 In
early 2007, Salmonella notifications in New South
Wales increased to levels well above previous years4
in part due to the outbreak reported in this paper.
On the evening of Monday 26 March 2007 a hospital
emergency department (ED) in Sydney notified the
Sydney South West Area Health Service (SSWAHS)
Public Health Unit (PHU) of 5 people (three from
the same family) with severe acute gastroenteritis
who had presented to this ED reporting that they
ate food from the same bakery, five to 24 hours prior
to becoming ill. In the ensuing days this evolved to
be one of the largest Salmonella outbreaks reported
in New South Wales. The pathogen responsible
was subsequently determined to be Salmonella
Typhimurium phage type 9. This paper describes
the public health investigation and response to the
Active case findings were conducted by requesting
that other public health units in New South Wales,
EDs and general practitioners (GP) notify the PHU
of any possible cases. Laboratories and EDs close
to the premises were contacted for updated possible
case listings on a daily basis for 1 week.
Cases were also identified through routine
laboratory notification of Salmonella species and
Salmonella Typhimurium between 19 March and
10 April 2007. Case details were verified against
an electronic clinical information system for prob-
able cases who presented to facilities within the
SSWAHS (n = 125). Where a person was unable
to contacted directly, details were verified by discus-
sion with the treating clinician or the notifier.
Details were collected on all food eaten in the 3 days
preceding onset of illness for the first 20 cases notified
to the PHU to confirm that food from the bakery
was the only common exposure, hypothesising that
this was the likely source of gastroenteritis illness.
Other cases were interviewed with a standardised
questionnaire to characterise exposure and obtain
demographic, illness and exposure details. A case was
defined as ‘any person who had a consistent illness
with symptoms of either diarrhoea and/or vomiting
AND who ate food prepared at the bakery between
19 and 27 March 2007, prior to the onset of their
illness’. Cases were confirmed by interview with the
case or by discussions with treating clinicians.
All 5 symptomatic food handlers were interviewed.
All symptomatic food handlers working at the bak-
ery had stool cultures for Salmonella and arrange-
ments were made for asymptomatic food handlers
to be screened.
Six weeks after the commencement of the outbreak
the SSWAHS PHU conducted a sub-investigation.
Forty-five cases were randomly selected for follow-
up interview to obtain information on illness dura-
tion, illness severity and contact with health care
facilities. The subset was randomly select using a
random number generator matched with case iden-
Data were entered into Epi Info 3.3.2 for Windows
(CDC, Atlanta, GA, USA), and analysed using SAS
System® version 8.02 and Microsoft® Excel 2000.
The NSW Food Authority was notified of the initial
cluster of gastroenteritis cases by the PHU on the
evening of Monday 26 March and inspected the
bakery the following day. Food handling and clean-
ing practices were reviewed as part of the environ-
mental investigation. Food items were sampled for
microbial testing and environmental samples were
taken from surfaces and equipment where food was
prepared. A trace-back of all foods served at the
premises was initiated, which included an inspec-
tion of an egg farm. The farm inspection involved
a review of egg handling procedures and the collec-
tion of egg and environmental samples for microbial
testing. Environmental samples included chicken
faeces specimens, meat meal, stock feed and drag
swabs collected from 4 laying sheds. Samples were
tested by the Division of Analytical Laboratories.
Salmonella isolates from stool and blood specimens,
were cultured at laboratories throughout Sydney,
and collated by the PHU. Serotyping of human
Salmonella isolates and isolates from food and envi-
ronmental sources collected at the food premises was
performed by the Institute of Clinical Pathology and
Medical Research and collated by the PHU. Multiple
locus variable number of tandem repeats analysis
(MLVA) was performed on these isolates.5,6 These
isolates were then tested by multiplex polymerase
chain reaction-based reverse line blot (mPCR/RLB)
phage-type (PT) prediction assay.
The SSWAHS PHU was notified of 365 pos-
sible cases, from which 319 cases were identified.
Of these, 283 were interviewed. The remaining
36 cases were confirmed by contacting the treating
clinician (Figure 1). Five cases were food handlers
who worked at the bakery and reported eating food
from the bakery prior to becoming unwell. Onset
dates for the food handler cases were between
23 and 27 March.
Demographic details are contained in Table 1.
Cases were aged between one and 74 years
Figure 1: Case verification process
365 possible cases
Unable to contact
Excluded – did
not eat at
bakery or did
not get sick =
Not a case
Case (283, 76%)
Unable to contact
or verify case
details = Not a
case (28, 8%)
clinician = Case
Total cases = 319
CDI? Vol?34?No?1? 2010?43
(median = 31 years) and most cases were male.
Figure 2 contains the epidemic curve and the date
of purchase of food from the bakery. The peak of
the outbreak occurred on 25 and 26 March with
67% (189/283) of cases reporting disease onset on
these 2 days. Incubation periods ranged from one to
118 hours (median 10 hours). The median reported
incubation period decreased as the outbreak pro-
gressed (Figure 2); ranging from 104 hours for cases
who purchased food on 22 March (n = 2) to 8 hours
for those who purchased food on 27 March (n = 44).
Food was purchased by cases between Thursday
22 March and Tuesday 27 March (Figure 3). All
but 7 cases purchased food over a 4-day period,
with 40% (113/283) being purchased on a single
day. Forty-four cases reported purchasing food on
27 March, after the outbreak was notified. Of these,
28 were admitted to hospital. All cases interviewed
reported that they ate a chicken, pork or salad roll
from the bakery prior to becoming unwell (Table 2).
While most cases reported that they ate a whole roll
prior to becoming unwell, 14 cases reported that
they ate only a very small amount of the roll, as little
as a single bite.
Attack rates could not be calculated, however, the
owners of the food premises estimate that 320 pork
rolls would be sold over a 5 day period (40 on week-
days and 100 weekend days). On 3 days at the peak
of the outbreak (Sunday, Monday and Tuesday) the
number of cases who purchased food exceeded the
estimated number of rolls sold. The attack rate on
these days therefore is likely to be close to 100%,
although it may be much less on other days.
There were 8 food handlers at the bakery. Of these,
five met the case definition and six had positive stool
cultures for Salmonella Typhimurium phage type 9,
with one initially being reported as asymptomatic.
Later in the investigation this food handler admit-
ted to having some stomach pain after consuming a
small amount of a pork roll on 27 March.
Table 1: Demographic and clinical characteristics for cases, Salmonella Typhimurium outbreak,
Sydney, March 2007 (n = 283) and a subset of 45 cases interviewed a second time to determine
illness severity and duration
Full outbreak Illness severity subset
Male 171 60.4 2146.7
Female 11239.624 53.3
0–910 3.52 4.4
40–49 48176 13.3
50–5928 9.94 8.8
60+9 3.22 4.4
Unknown 144.91 2.2
Figure 2: Number of cases and incubation
period (hours), Salmonella Typhimurium
outbreak, Sydney, March 2007, by date of
onset and date of purchase
Number of cases
Median incubation period (hours)
Date of purchase
Date of onset
Median incubation period
Table 2: Foods eaten by cases, Salmonella Typhimurium outbreak, Sydney, March 2007 (n = 283)
Yes NoUnknown Percentage
Bread roll 28012 98.9
Salad vegetables 26310 1092.9
Pate 1388659 48.8
Chicken103 15228 36.4
Ham9495 94 33.2
Salad only, no meat14 2690 04.9
During the outbreak, most cases (94%, 301/319) pre-
sented to a health care provider; 80 of these presented
only to their GP and 69% (221/319 cases) presented
to a hospital ED. A large number of the cases (62%,
136/221) who presented to EDs required hospital
admission. Median duration of hospital admission
was 24 hours. One case reported being admitted for
8 days. Of those cases who presented to a hospital
within the SSWAHS boundary (n = 125), 65% of
cases required hospital admission, with a median
duration of 31 hours. Cases presented to 26 different
hospitals across New South Wales. Seventy-six cases
(24%) related to this outbreak presented to a single
hospital ED over a 4-day period; 42 (13%) cases
presented to a second hospital. These 2 hospitals
are located within 5 kilometres of the bakery. Most
presentations occurred on a single day.
Sub-investigation of illness severity
Of the 45 cases in the sub-analysis group (demo-
graphic characteristics are shown in Table 1), illness
duration (defined as number of days before the case
reported their health returned to normal) ranged
from three to 45 days (median 14 days); at least
5 people reported still being unwell at the time of
second interview. Eighteen per cent of this subset
purchased food on 24 March, 40% on 25 March, and
20% on either 26 or 27 March. Thirty-nine of these
cases (87%) reported taking time off work or school
due to the illness; 20 cases reported taking less than
a week off work or school, 14 between one and
2 weeks, one between two and 3 weeks and three
greater than 3 weeks. Fifty-six per cent of this subset
of cases indicated on questioning that they required
intravenous fluid replacement at the time of the
acute illness. Figure 3 displays a dose-response
curve for this subset of cases, showing that duration
of illness increased with the estimated quantity of
the food eaten.
Salmonella Typhimurium was cultured on stool
from blood from 173 cases (54%). Thirty-nine
human isolates from cases were typed using MLVA
and had MLVA type 01-03-20-04-06. These isolates
exhibited a phage reaction pattern on mPCR/RLB
that was consistent with Salmonella Typhimurium
phage type 9, subsequently confirmed by phage
typing. No other types of Salmonella were isolated
The bakery is a busy premises located near a train
station in the inner west of Sydney. The bakery
had previously been inspected by the NSW Food
Authority in September 2005 during a food han-
dling survey. The premises consisted of a shop
fronted by a serving counter and a rear preparation
area. The front counter was fitted with a glass pastry
display unit and an adjoining takeaway food bar
unit used to display ingredients for the preparation
of pork and chicken rolls. The rolls contained any
combination of ingredients including sliced hams,
sliced pork, marinated cooked diced chicken, pate,
raw egg mayonnaise, shredded carrot, coriander
Figure 3: Dose response curve: Amount of
roll eaten, Salmonella Typhimurium outbreak
subset of 45 cases, Sydney, March 2007, by
number of days sick
A bite Quarter rollHalf rollThree
One rollOne and
Amount of roll eaten
Number of days sick
The inspection and interviews with the owners of
the premises revealed that there was a lack of effec-
tive sanitation of food handling equipment and
surfaces. It was also noted that there was a refrigera-
tion malfunction in the early hours of the morning
of 25 March (after the outbreak had commenced).
Inadequate refrigeration was also recorded for the
raw egg mayonnaise in the display unit.
Salmonella Typhimurium phage type 9 was detected
in 15 isolates from food and environmental samples
taken at the source premises. Salmonella Typhi-
murium was isolated in the raw egg mayonnaise,
ham, pork, chicken, pate and shell eggs and from
swabs of the preparation bench, tongs, meat slicer,
floor drain and display tray. All isolates from food and
environmental samples were MLVA type 01-03-20-
04-06, which is identical to that isolated from cases.
A quantitative analysis of the raw egg mayonnaise
sample yielded a count in excess of 1.1 x 107 colony
The eggs used in the raw egg mayonnaise were
traced to a farm in outer Sydney. The farm was
inspected on 4 April 2007. The drag swab from a
laying shed on the egg farm and from a meat meal
sample were positive for Salmonella although sub-
sequent serotyping and phage typing revealed that
the egg farm isolates did not match the human and
food isolates linked to the outbreak.
During the inspection of the premises on the
morning of 27 March the NSW Food Authority
removed the raw egg mayonnaise from sale. By
5 pm 27 March, a prohibition order was issued to
the premises by the NSW Food Authority to pre-
vent sale of chicken and pork rolls and associated
ingredients. Upon receipt of analytical results a sub-
sequent prohibition order was issued on 30 March
preventing the business from operating.
This describes one of the largest point source out-
breaks of Salmonella in Australia for many years.
Similar outbreaks linked to Vietnamese pork rolls
in Victoria in 1997 resulted in 774 and 154 cases.7
Other large outbreaks in recent years have related to
commercially available products, including chicken
meat8 and fruit juice.9 In the outbreak described
in this report, routine surveillance through statu-
tory reporting and active case finding identified
319 cases, about half of which were laboratory
confirmed. This outbreak is notable for its scale, the
severity of illness experienced by the cases and the
degree of contamination at the point source.
A large proportion of cases required hospital admis-
sion, with many requiring prolonged stays. Hospital
admission for salmonellosis generally only occurs in
more severe cases, with most cases (estimated to be
over 90%) being treated in the community.1 In this
outbreak, two-thirds of all identified cases presented
to an emergency department at some point during
their illness and half were admitted to hospital.
In addition, the average duration of illness in this
outbreak was long. Salmonellosis usually results
in diarrhoea lasting 3–7 days,3 whereas an average
duration of diarrhoea of 14 days was observed in
this outbreak, and many experienced illnesses of
longer duration of up to 45 days. Over half of the
cases in this sub-investigation reported requiring
intravenous fluid replacement due to dehydration,
further indicating that illness was severe.
It is estimated that the attack rate was close to 100%,
which is consistent with or higher than similar out-
breaks reported elsewhere.10–13 The incubation period
decreased as the outbreak progressed, indicating
significant contamination at the point source, and
likely bacterial growth in foods and spread to surfaces
as the outbreak progressed. Inadequate temperature
controls when storing foods, as was found in the
environmental investigation, would have contrib-
uted to bacterial proliferation. The environmental
investigation revealed widespread contamination
throughout the premises; Salmonella was detected
in most foods used in the preparation of the chicken
and pork rolls, on the food slicer, preparation bench,
serving tongs, display tray and floor drain. As a
result, a specific source of contamination was not
established. It is clear, however, that there were defi-
ciencies in food handling and sanitation contributing
to the proliferation of the organism and its spread
throughout the premises after the initial contamina-
tion occurred. Similar deficiencies in handling and
cross contamination were noted in 2 other outbreaks
linked to similar premises in Victoria.7
The 2 most likely sources of the initial contamina-
tion are the eggs used for the raw egg mayonnaise
or, less likely, an asymptomatic food handler.
Salmonella was identified in the stool of a food han-
dler with only mild symptoms of stomach pain after
the outbreak had subsided, but it was not possible
to confirm whether carriage commenced during or
prior to the outbreak. Positive stool specimens were
found in 2 asymptomatic food handlers working at
a bakery associated with an outbreak of Salmonella
Typhimurium phage type 9 in Victoria in 1997.7 The
significance of detection of Salmonella in this previ-
ous outbreak was also not established. Under the
NSW Food Act 2003 (with reference to the Australia
New Zealand Food Standards Code covering health
and hygiene requirements for food handlers) 14
food handlers with symptoms of foodborne illness,
or carriage, are prohibited from handling food if
there is a possibility they may contaminate it. NSW
Health recommends that food handlers who have
diarrhoea are excluded from work for 48 hours after
symptoms resolve. Neither the NSW Food Act 2003
or the Food Standards Code allow for this and con-
sideration should be given to strengthen, and clarify,
the restriction of food handlers who are ill.
Salmonella Typhimurium phage type 9 was identi-
fied on the shells of one open tray of eggs kept on the
premises. However, no Salmonella was found on eggs
contained in 2 closed cartons also on the premises.
Traceback to the farm identified other Salmonella
serovars but not Salmonella Typhimurium phage
type 9, however it is possible that Salmonella serovars
may move in and out of poultry flocks with the intro-
duction of contaminated feed, or other inputs and
excretion of Salmonella may be intermittent.11 Raw
eggs are frequently implicated in large outbreaks of
Salmonella11 and in this outbreak it is believed that
heavily contaminated mayonnaise, and subsequent
cross contamination was the most the likely cause.
Egg related outbreaks have increased in Australia in
recent years. OzFoodNet data show that there have
been 63 outbreaks in the last 3 years where egg or
egg based dishes were suspected to be the source;
14% of all foodborne illness outbreaks in 2005 were
considered to be related to eggs, compared with 23%
in 2008.15–20 While eggs are the most likely vector for
the majority of these outbreaks, poor hygiene and
food handling at retail level is also a major contrib-
uting factor and influences the size and severity of
There are several approaches to reduce the occur-
rence of such large outbreaks resulting from
service of high risk foods, including raw egg based
products. Starting with a raw product that has the
lowest level of contamination possible is key. The
reduction of pathogens on egg farms is a priority
and many countries have developed schemes and/
or regulations to reduce human pathogens such
as Salmonella and Campylobacter in poultry.21,22 In
Australia, Food Standards Australia New Zealand
(FSANZ) is developing a Primary Production and
Processing Standard for Eggs and Egg Products.23
This will impose requirements on producers and
users of eggs at the primary production and retail
level. The draft standard addresses specific food
safety risks associated with cracked and dirty eggs
and prohibits sale of unpasteurised egg pulp for
retail or catering purposes. In New South Wales,
the NSW Food Authority will be responsible for
implementation and enforcement of the national
standard. Additional cooperation between govern-
ment regulators and industry is vital to ensure that
food safety measures are of the highest standard to
protect public health. Stronger, proactive engage-
ment by government with industry, including the
egg industry, will assist with better sharing of infor-
mation and earlier identification of hazards in the
food production chain.
The NSW Food Authority is working with all New
South Wales local councils, who are responsible for
routine inspection of retail businesses such as bak-
eries and cafés, to educate businesses on the risks
of using products containing raw eggs. In 2005 the
bakery at the source of this outbreak was presented
with information about the risks of raw egg based
foods and also advised to use a commercial, pasteur-
ised mayonnaise product. This advice was ignored
and a large fine was subsequently imposed by the
courts as a result of prosecution action undertaken
by the NSW Food Authority. This is a powerful
deterrent for businesses to avoid using raw egg
products where possible. Standard fact sheets and
warning letters have been developed for distribution
by local council officers to businesses where these
are found to be using raw egg products.
The widespread contamination in the bakery indi-
cates that food handling skills were not adequate.
A national food handler survey undertaken by
FSANZ in 2007 indicated that bakeries were less
proficient in food handling activities compared
with other businesses surveyed.24 Protecting public
health through food safety is primarily the respon-
sibility of jurisdictional food regulators. The NSW
Food Authority is responsible for regulating the
food industry through the NSW Food Act 2003. This
Act, through the Food Standards Code, requires food
handlers to have adequate skills and knowledge in
food safety and food hygiene if employed in food
service. The owner of the business is responsible for
ensuring staff have adequate skills and knowledge
appropriate for their activities within the business.
However, no formal training is currently required,
even for the owner. In New South Wales this is
being addressed through implementation of a Food
Safety Supervisor Initiative. This will require every
retail food business in New South Wales to have a
dedicated food safety supervisor with responsibility
for staff training. The scheme will be implemented
in the middle of 2010.
Outbreaks, even in small premises, can result in high
morbidity, high cost to the community and have a
significant impact on health care provision. The
scale and impact of this outbreak augurs for contin-
ued investment in risk assessment and food safety
measures at all stages throughout the food chain and
regulation of primary producers and food premises.
The most notable feature of this outbreak is its scale,
the health and economic consequences for those
affected and the resultant burden on the health
system of a preventable foodborne illness. This
outbreak is likely to have resulted in considerable
costs to the healthcare system and the community.
Given the number of people presenting to EDs
and requiring admission to hospital, the cost to
the hospitals involved would have been significant.
CDI? Vol?34? No?1?2010?47
This outbreak was initially reported to local health
authorities by an astute ED clinician. Early notifica-
tion by clinicians requires the development of good
relationships with their local public health author-
ity. Continued co-operation between NSW Health
and the NSW Food Authority will provide for early
identification of outbreaks and timely public health
action to protect public health. These authorities
must seek to continually improve outbreak response
procedures to reduce the impact on public health
through, among other avenues, regular evaluation
of outbreak response. Based on the estimated attack
rate and the degree of environmental contamination,
it is likely that the closure of the bakery prevented
many more cases occurring. The co-operation and
timely communication between clinicians at the
coal face, public health officials and the regulators
provided an opportunity for urgent and responsible
public health action.
We wish to acknowledge the assistance of the fol-
Public Health Units in New South Wales, in par-
ticular Sydney West Public Health Unit for assisting
with interviews and case finding;
Sydney South West Public Health Unit staff;
NSW Food Authority, specifically Marianne Tegel;
Emergency Department staff from Concord and
Division of Analytical Laboratories Lidcombe for
food and environmental testing, and others;
Dr Bradley Forssman for advice on the manuscript.
At the time of the outbreak, Alexander Rosewell was
a Master of Applied Epidemiology Scholar at the
Australian National University, a program funded
by the Australian Government Department of
Health and Ageing.
4.? Australian? National? University? Master? of? Applied?
5.? Communicable? Diseases? Branch,? New? South? Wales?
This? work? was? conducted? by? the? Sydney? South? West? Area?
Corresponding? author:? Ms? Trish? Mannes,? Consultant? in?
Health? Protection,? Thames? Valley? Health? Protection? Unit,?
1.? Hall?G,?Kirk?M.?Foodborne Illness in Australia: Annual
Incidence circa 2000.?Australian?Government?Department?
The health of the people of New South Wales – Report
of the Chief Health Officer.?Sydney:?New?South?Wales.?
3.? Hawker? J,? Begg? N,? Blair? I,? Reintjes? R,? Weinberg? J.?
Communicable Disease Control Handbook.? 2005,?
4.? NSW? Department? of? Health,? Communicable? Diseases?
Branch.? Communicable? Diseases? Report,? New? South?
Wales,?for?March?and?April?2007.?N S W Public Health
5.? Gilbert? GL.? Using? MLVA? to? type? strains? of? Salmonella?
Typhimurium?in?New?South?Wales.?N S W Public Health
6.? Torpdahl? M,? Sørensen? G,? Lindstedt? BA,? Nielsen? EM.?
Tandem? repeat? analysis? for? surveillance? of? human?
Salmonella Typhimurium? infections.? Emerg Infect Dis?
1997.? Victorian? Government? Department? of? Human?
chasing? of? chicken? meat? from? a? supermarket? chain.?
Commun Dis Intell?2006;30(4):449–455.
break,?South?Australia.?Commun Dis Intell?1999;23(3):73.
ucts?containing?raw?egg?in?Tasmania.?Commun Dis Intell?
Williams? C.? Salmonella? Typhimurium? U290? outbreak?
linked?to?a?bakery.?Commun Dis Intell?2003;27(4):514–
48? CDI? Vol?34?No?1?2010 Download full-text
13.?Tribe? I,? Hart? S,? Ferrall? D,? Givney? R.? An? outbreak? of?
Salmonella? Typhimurium? phage? type? 99? linked? to?
contaminated? bakery? piping? bags.? Commun Dis Intell?
15.?The? OzFoodNet? Working? Group.? Monitoring? the? inci-
dence? and? causes? of? diseases? potentially? transmitted?
Network,? 2006.? Commun Dis Intell? 2007;31(4):345–
Australia:? Annual? report? of? the? OzFoodNet? Network,?
2007.?Commun Dis Intell.?2009;33(4):389–413.
report,?1?January?to?31?March?2008.?Commun Dis Intell?
report,? 1? April? to? 30? June? 2008.? Commun Dis Intell?
report,?1?July?to?30?September?2008.?Commun Dis Intell?
21.?European? Union.? Regulation? (Ec)? No.? 2160/2003?
Of? The? European? Parliament? And? Of? The? Council? of?
other? specified? food-borne? zoonotic? agents.? Official?
Journal? of? the? European? Union.? 2003.? Accessed? on?
22.?New? Zealand? Food? Safety? Authority.? Campylobacter?
in? poultry:? Risk? Management? Strategy? 2007–2010.?
September? 2007? Accessed? on? 2? March? 2009.?
24.?Food? Standards? Australia? New? Zealand,? 2007?
National? Food? Handling? Survey,? November? 2008,?
Evaluation? Report? Series? No.? 19.? Accessed? December?