D e t e c t i o n , o c c u r r e n c e , g r o w t h a n d i n a c t i v a t i o n o f C r o n o b a c t e r s p p . M . C .

K a n d h a i 2 0 1 0
D
e
t
e
c
t
i
o
n
,

o
c
c
u
r
r
e
n
c
e
,
g
r
o
w
t
h

a
n
d

i
n
a
c
t
i
v
a
t
i
o
n
o
f

C
r
o
n
o
b
a
c
t
e
r
s
p
p
.
(
E
n
t
e
r
o
b
a
c
t
e
r

s
a
k
a
z
a
k
i
i
)
C
h
a
n
t
a
l

K
a
n
d
h
a
i
k
a
f
t
_
d
e
f
.
q
x
d


2
0
/
0
3
/
2
0
1
0


1
1
:
3
6


P
a
g
e

1
Propositions

1. In the last decennium, detection methods for Cronobacter spp. have been
improved, which has led to an increase in the isolation of Cronobacter spp.
from a wide variety of food products and environments. This fact in itself is
not a reason to consider Cronobacter spp. as an emerging pathogen.
2. The use of sterile liquid formulae is an effective control measure to reduce
the probability of infections in low and very low birth weight infants due to
Cronobacter spp..
3. As, generally, pathogens are relevant in foods only at such low numbers
that multiplication is required to allow their detection, the enrichment step
must be regarded as the most important step in the entire detection
procedure. Yet, relatively little research effort is dedicated to optimizing
this procedure.
4. Improvements in food safety management can only successfully be
achieved in a food enterprise when there is a true food safety culture in the
food enterprise.
5. Children are an extraordinary gift and enhance their mother's career by
improving her multitasking skills.
6. Patience is the most important virtue in joint scientific writing.






Propositions belonging to the thesis, entitled:

“Detection, occurrence, growth, and inactivation of Cronobacter spp.
(Enterobacter sakazakii)”

M.C. Kandhai
Wageningen, 23 April 2010.




Detection, occur r ence, gr owth
and inactivation
of Cronobacter spp.
(Enterobacter sakazakii)
























M.C. Kandhai



















Thesis committee

Thesis super visor s
Prof. dr. L.G.M. Gorris
Chairholder of the European Chair in Food Safety Microbiology
Wageningen University

Prof. dr. ir. M.H. Zwietering
Professor of Food Microbiology
Wageningen University

Thesis co-super visor
Dr. ir. M.W. Reij
Researcher and Lecturer at Laboratory of Food Microbiology
Wageningen University

Other member s
Prof.dr.ir. M.A.J.S. van Boekel, Wageningen University
Prof. S. Fanning, UCD Center for Food Safety, Ireland
Prof. dr. ir. A.H. Havelaar, University of Utrecht
Dr. ir. A.E. Heuvelink, Food and Consumer Product Safety Authority, Zutphen




This research was conducted under the auspices of the Graduate School of Voeding,
Levensmiddelentechnologie, Agrobiotechnologie en Gezondheid (VLAG).


Detection, occur r ence, gr owth and inactivation
of Cronobacter spp.
(Enterobacter sakazakii)

M.C. Kandhai




























Thesis
Submitted in fulfilment of the requirements for the degree of doctor
at Wageningen University
by the authority of the Rector Magnificus
Prof. dr. M.J. Kropff,
in the presence of the
Thesis Committee appointed by the Doctorate Board
to be defended in public
on Friday 23 April 2010
at 4 PM in the Aula.




































M.C. Kandhai
Detection, occurrence, growth and inactivation of Cronobacter spp. (Enterobacter
sakazakii),
240 pages.

Thesis, Wageningen University, Wageningen, NL (2010)
With references, with summaries in Dutch and English

ISBN 978-90-8585-556-9
















To Fons, Aditya,
Sarada & Ishaan


“Life is constantly changing.
Whatever time you have is yours,
for tomorrow might never come”


Abstr act
The genus Cronobacter consists of Gram-negative, motile, non-spore forming, facultative
anaerobic bacteria, and was originally defined as one species “Enterobacter sakazakii”
within the genus Enterobacter in 1980. Cronobacter spp. have been documented as a rare
cause of outbreaks and sporadic cases of neonatal meningitis, necrotizing enterocolitis, and
sepsis in infants with a high mortality. Among these infants, those at greatest risk are
infants less than 2 months of age, particularly pre-term infants, low birth weight (LBW)
infants (< 2500 g), and immuno-compromised infants.
At the onset of the work for this thesis, Cronobacter spp. had been isolated from milk-
based powdered formulae which have a direct link to the sub-population at greatest risk.
However, there was a need to more closely investigate whether and where Cronobacter
spp. occurs in environments in which these powdered are manufactured and packed but also
to investigate other sources which could lead to exposure of vulnerable sub-populations.
The main objective of this study was to develop isolation and detection methods that would
allow quick and reliable investigation into the occurrence of the micro-organism in
potential sources. Furthermore, more insight into the growth behavior of Cronobacter spp.
in reconstituted infant formula was necessary in order to provide data to be used in
Microbiological Risk Assessment (MRA) dedicated to this particular food product.
A selective enrichment method was developed for the rapid and reliable enrichment and
detection of Cronobacter spp. in environmental samples. The detection method which was
developed is based on two features of Cronobacter spp. combined: their yellow pigmented
colonies when grown on tryptone soy agar and their constitutive o-glucosidase, which can
be detected in a 4-h colorimetric assay. The initially developed method and refinements
thereof were applied for routine screening for the presence of Cronobacter spp. in
environmental samples and a variety of food products manufactured or marketed in The
Netherlands. The detection method described in this thesis has been the basis for a series of
media for Cronobacter spp. that have recently been commercialized.
Quantitative data on product contamination at manufacture, during preparation, and also
growth after reconstitution are required in order to assess the risk associated with


Cronobacter spp. exposure. Next to that, tools are needed to asses the micro-organisms
growth potential as well as its inactivation (thus, its survival) due to specific control
measures applied. In this thesis, predictive growth models were developed that capture key
growth parameters. Minimum– and maximum temperatures estimated with the Secondary
Rosso equation were 3.6 ºC and 47.6 ºC, respectively. The estimated lag time of the micro-
organisms was found to vary from 83.3 ± 18.7 h at 10 ºC to 1.73 ± 0.43 h at 37 ºC and
could be described with the hyperbolic model and reciprocal square root relation. The
models for growth rates and lag times as a function of temperatures obtained during this
study allow estimating the potential growth of Cronobacter spp. in reconstituted infant
formula stored at any temperature below 47 °C.
As growth rates of Cronobacter at refrigeration temperatures are relatively small,
caregivers are advised to store reconstituted infant at low temperature as a control measure
to prevent microbial growth. It is evident that storage of reconstituted formula in a
refrigerator may require a significant amount of the time before the formula reach the
targeted refrigeration temperature. Therefore, a mathematical model was built to predict the
temperature profile and the resulting growth of Cronobacter spp. during cooling, i.e. under
dynamic temperature conditions. Predictions showed that proliferation of Cronobacter spp.
during cooling strongly depends on the size of the container used for storage and that it may
be prevented by limiting the volume to be cooled to portion-size only or by reconstituting at
temperatures of 25 °C or lower.
The survival of two Cronobacter strains in dry powdered infant formula (PIF) was tested
and compared to the survival of six other bacterial strains after inoculation and storage at
several temperatures between 7 and 42 ºC. The effect of temperature on survival in PIF,
was described using both the Weibull distribution model and the log-linear model.
Differences were found in the rate of survival that can be due to difference in the resistance
to inactivation in dry environments between Cronobacter species, which could be relevant
to consider when establishing quantitative risk assessments on consumer risks related to
PIF.
The research described in this thesis contributes to the existing knowledge on the natural
habitat of Cronobacter spp. and its occurrence and behavior in PIF. The models developed


for quantifying the growth of Cronobacter spp. in reconstituted formulae under various
conditions can be applied in risk assessments set-up to estimate the probability of
vulnerable sub-populations becoming ill after consuming infant formulae. International and
national governmental bodies may use these predictive models in risk assessments and to
establish guidelines for health care professionals to provide effective hygiene training to
parents and professional caregivers to ensure that PIF is prepared handled and stored
appropriately.


Contents

Abstr act 6

Chapter 1 Introduction 11

Chapter 2 A new protocol for the detection of Cronobacter spp.
applied to environmental samples 43

Chapter 3 A simple and rapid cultural method for enrichment
of Cronobacter spp. in environmental samples 55

Chapter 4 Occurrence of Cronobacter spp. in food production
environments and households 69

Chapter 5 A study into the prevalence of Cronobacter spp. in
The Netherlands 77

Chapter 6 The effect of pre-culturing conditions on the lag time
and specific growth rate of Cronobacter spp. in
reconstituted infant formula 109

Chapter 7 Growth of Cronobacter spp. under dynamic
temperature conditions occurring during cooling
of reconstituted powdered infant formula 133

Chapter 8 Survival of Cronobacter spp. in comparison to
other bacteria in dry infant formulae 165

Chapter 9 Discussion, and concluding remarks 193

Samenvatting 227

Acknowledgement 231

List of publications 233

Cur r iculum vitae 237

Over view of completed tr aining activities 238



Introduction

11
Chapter 1
12
Gener al
Cronobacter spp. (Enterobacter sakazakii) is an opportunistic food-borne pathogen, that
has been linked with serious infections in infants, causing bacteraemia and meningitis and
associated with necrotizing enterocolitis (NEC) (26, 27, 54). Most cases of Cronobacter
disease have been detected among newborn and very young infants in hospital nurseries
and neonatal intensive care units (28). The first reported cases attributed to this organism
occurred in 1958 in England (See Figure 1) and resulted in the death of two infants (74). At
the time the micro-organism was referred to as “the yellow pigmented Enterobacter
cloacae”. In most outbreaks, in which a source could be established, the source of
Cronobacter spp. was proven to be (dry) powdered infant formulae (PIF) (5, 6, 69, 75, 76).
Interest has been increasingly raised in the public domain with notable activities of Codex
Alimentarius, national governments, scientific communities, health care providers and the
infant food industry, since an outbreak involving powdered infant formula (PIF) in the USA
in 2002 (36).


Figur e 1. First report of an outbreak related to Cronobacter spp. in scientific literature.

In 2002, the International Commission on Microbiological Specifications for Food
(ICMSF) categorised Cronobacter spp. as a severe hazard for restricted populations,
The Lancet, Volume 277, Issue 7172, 11 February 1961, Pages 313-315


Introduction

13
resulting in life-threatening diseases or substantial chronic sequelae or effects of illnesses of
long duration (37). There is evidence that consumer groups vulnerable to Cronobacter
infections are premature and low-birth-weight infants and those aged < 28 days (8, 22).
More recently, the Codex Committee for Food Hygiene (CCFH) of Codex Alimentarius
and the BIOHAZ scientific expert panel of the European Food Safety Authority (EFSA)
established that all infants at or below 12 months are at particular risk for Cronobacter
infections (21, 22). Among these infants, those at greatest risk are neonates (< 28 days),
particularly pre-term infants, low birthweight (LBW) infants (< 2500 g), and immuno-
compromised infants, and those infants that are less than 2 months of age. Infants of HIV-
positive mothers are also at risk for Cronobacter infection, as they may specifically require
infant formula and may be more susceptible to infection (12, 21, 26-28).

Micr obiological Risk Assessment appr oach
Microbiological risk assessment (MRA) is widely used to characterise health risks
associated to the potential presence of microbial hazards in foods. An MRA consists of four
stages: 1) hazard identification, 2) exposure assessment, 3) hazard characterization and 4)
risk characterization (49). A hazard is defined as a biological, chemical or physiological
agent in food (or condition of food) with the potential to cause an adverse health effect (11).
The magnitude of the associated risk is obtained as well as insight in risk-contributing
factors, and possible risk-mitigating control measures.
To determine the magnitude of the risk through MRA predictive models are often used as
tools in such assessments. Suitable predictive models capture insights in the dynamics of
the target micro-organisms, such as inactivation, survival and growth. Resulting parameters
can be applied to estimate the consumer risk and to give directions to which preventive
measures should be advised (51). Most value is derived from being able to quantify
individual parameters. In this thesis the aspects that have been investigated are presented
within the framework of the MRA. Information relevant for compiling a quantitative MRA
relevant to Cronobacter spp. in PIF is presented. In Figure 2, the various steps of this MRA
Chapter 1
14
approach are given with sub-topics indicated in the various boxes. Grey boxes indicate the
topics that were investigated in more detail in this thesis.


Figur e 1. Schematic view of elements of microbiological risk assessment for Cronobacter spp. in
powdered infant formula. All boxes are described in this introduction section. The grey boxes indicate
the research topics focused on in this thesis.
Introduction

15
Epidemiology
The epidemiology of Cronobacter spp. is poorly understood. Cronobacter infections are
very rare and often underreported, especially in developing countries (23). Since 1961 and
up to July 2008, 156 documented cases of Cronobacter spp. infections from all parts of the
world have been reported in the published literature and in reports submitted by public
health organizations and laboratories. Of these 156 cases, at least 29 cases (19%) resulted in
death (28).
Specific international attention has been given to the safety of food for infants and young
children as it relates to the possible presence of Cronobacter spp. in powdered formulae
intended for this consumer group. The Food and Agriculture Organization of the United
Nations (FAO) and the World Health Organization (WHO) have established two
international risk assessments on the topic (26, 27) that have provided a scientific basis for
creating a new code for hygienic practices by CCFH (12), aimed at providing risk
management guidance at the international level. In this code, products for infants up to 12
months of age are considered.
According to Codex Alimentarius, an infant is a person not more than 12 months of age,
while young children are persons from 12 months up to the age of three years (36 months)
(12). A neonate is a newborn infant, less than four weeks old, and a premature is an infant
that was born prior to 37 weeks of gestation. An immunocompromised infant may have an
immunodeficiency of any kind, and may therefore be particularly vulnerable to
opportunistic infections in addition to normal infection that could affect anyone. A number
of factors could contribute to an infant’s immune status, i.e. nutritional status such as
vitamin status, HIV status, clinical conditions, pharmaceutical treatment, low birth weight,
and premature birth. Because the prevalence of these factors varies among countries, there
is a wide variation in the prevalence of immunocompromised infants globally.
Diseases caused by Cronobacter spp. may affect the intestines (causing necrotizing
enterocolitis), invade the blood stream (bacteraemia and/or sepsis) and/or invade the central
nervous system, causing celebritis and/or meningitis. Meningitis has a high mortality rate
(42%) and many of the survivors (74%) suffer from neurological complications (66).
Chapter 1
16
Premature infants, especially those with a low birth weight (< 2500 g) or a very low birth
weight (< 1500 g), are thought to be at greater risk for severe infection than more mature
infants, children or adults (26, 33, 48). The incidence of Cronobacter spp. invasive
infections seems to be much higher among infants than older age groups such as young
children. Cases of Cronobacter spp. meningitis are reported exclusively among infants,
while Cronobacter bacteremia has occurred in all age groups (8). The group of infants at
risk for necrotizing enterocolitis, however, seems to be similar to the group of neonates
which is defined to be at risk for bacteraemia and meningitis (66).
Many infections in newborns are transmitted from mother to child. For several years
passage of the organism through the mother’s birth canal was therefore suspected to be the
source of Cronobacter spp. infection. In most cases that infection has occurred, however
both the route of exposure and the incubation period are generally unclear. Only in two
outbreaks, occurring in neonatal intensive care units, a clear relationship was shown
between Cronobacter spp. isolates from patients and isolates from unopened cans of
powdered dry infant formulae of the same batch as consumed by the patient (6, 15). It has
been reported that in outbreaks associated with powdered infant formulae, the illness started
as soon as 3 to 4 days after the initial exposure to the implicated formula (2, 69, 75).
Though Cronobacter spp. infections have been associated with contaminated powdered
formulae, environmental sources of contamination should not be excluded (58). While
relationships between sources of Cronobacter spp. other than from powdered infant
formula and human illness are less well understood, documented Cronobacter spp.
infections in immunocompromised adults (34, 48) may indicate potential other sources such
as home environments, food and food manufacturing environments, as well as insect or
animal reservoirs.
Table 1 shows an overview of documented outbreaks and cases of infections in infants and
children per year per country caused by Cronobacter spp. as published by the Food and
Agriculture Organization of the United Nations (FAO) and the World Health Organization
(WHO) (28). In various reports mortality rates between 40 and 80% are reported (48). A
recent survey showed a mortality rate of invasive neonatal Cronobacter infections of 27%.
The lethality was also calculated for Cronobacter meningitis (42%), Cronobacter
Introduction

17
septicaemia (< 10%), Cronobacter necrotising enterocolitis (19%) (31). Considering the
overview of invasive Cronobacter spp. outbreaks from 1961 to 2008 (28) as presented in
Table 1, an average mortality rate of 19% can be calculated (i.e. 29 death in 156 cases).
Analyzing the tota; number of cases did not yield consistent information about the cases per
age group in the population, so currently it is not possible to quantify existing potential
differences in the susceptibility of infants of different age groups. In the age group 6-11
months six cases are well documented, five of which were invasive. From the five invasive
cases, three had other medical problems. In the age group 12–35 months only two cases
have been well described. Cronobacter spp. infections have occurred in both hospital and
home settings, but the incidence rate appears to be low. In the first expert meeting of the
FAO/WHO on Cronobacter spp. in PIF, an annual incidence rate of 1 per 100,000 infants
(i.e. children < 12 months of age) for Cronobacter spp. invasive infections was estimated
for the United States of America (26), whereas the annual incidence rate among the low
birth weight infants (< 2500 g) was found to be higher at 8.7 per 100,000 infants (27).
Similarly, an another study estimated an incidence rate of 9.4 per 100,000 very low birth
weight infants (< 1500 g) (72).

Table 1. Overview of the number of invasive Cronobacter spp. outbreaks and cases in infants and
childeren reported until July 2008 (adapted from reference (28)). In all cases the age at illness onset
was < 1 year, except for *, where the child was 13 months of age
Publication Year Countr y No. cases Deaths
1961 UK 2 2
1965 Denmark 1
1979 USA 1
1981 USA 2
1983 Netherlands 8 6
1984 Greece 1
1985 USA 1
1987 Greece 11 4
1988 USA 2
1989 Portugal 1 1
1989 Iceland 3 1

Chapter 1
18
Table 1. Continued
Publication Year Countr y No. cases Deaths
1989 USA 4
1990 USA 1
1991 USA 1
1994 Germany 1
1997 Scotland 1
2000 USA 1
2000 Brazil 5
2001 USA 3
2001 Belgium 12 2
2001 Israel 2
2002 Israel 3
2002 USA 20 2
2002 Belgium 1 1
2003 Brazil 1
2003 Hungary 1
2003 USA 5 1
2004 USA 2
2004 Hungary 1
2004 New Zealand 5 1
2005 USA 2
2005 Hungary 1
2006 USA 5
2006 Hungary 1
2006 France 9 2
2007 USA 9 1*
2007 Canada 1
2007 India 2 1
2007 Spain 1
2007 France 18 4
2008 USA 3
2008 Japan 1
Total 156 29
Introduction

19
Pathogenicity
Enterobacter species are widely distributed in nature, occur in the intestinal tract of humans
and animals, and are frequently a cause of nosocomial infections. The dose-response
relationship for Cronobacter spp. causing illness in humans is not known and represents a
challenging aspect requiring further investigation. To date, relatively little scientific
attention has been given to unravelling the mechanism of pathogenicity or to potential
virulence factors of Cronobacter spp. (63). Some strains of C. sakazakii have been shown
to produce an enterotoxin (64) however, its relevance to pathogenesis has not been
established yet. There is no data available that any one of the Cronobacter species is more
pathogenic than another; each of the species have been linked retrospectively to clinical
cases of infection in either infants or adults (28). It is known that Cronobacter spp.
infection may cause a highly lethal syndrome of bacteraemia and meningitis with
involvement of the central nervous system in neonates. In several cases and outbreaks,
powdered infant formula has been identified as the source of infection, therefore the entry
for the infection can be gastrointestinal (64). Pagotto et al., (2003) used a suckling mice
model to determine the pathogenicity of 18 Cronobacter isolates (9 clinical, 8 food and 1
Type strain), showing that a dose of 10
8
CFU/mouse was lethal to 58 out of 69 (84%)
suckling mice within 3 days after intraperitoneal administration. A minimal number of 10
5

CFU of certain Cronobacter strains was lethal to 5 out 20 (25%) suckling mice, indicating
that the pathogenicity between strains tested may differ. Orally dosed suckling mice,
showed lethality at a dose of 10
7
CFU per mouse (2 out 8 suckling mice (20%) died ) (64).
These data have been extrapolated to estimate the minimum infectious dose for human
infants to be 10
5
CFU in a number of studies.
Currently, very little information is available on the virulence factors of Cronobacter spp.
and its pathogenic mechanisms.



Chapter 1
20
Taxonomy
Cronobacter spp. are Gram-negative, motile, peritrichous non-spore forming, straight rods,
within the family Enterobacteriaceae and originally belonging to the genus Enterobacter.
Until 1980, Cronobacter spp. was referred to as the “yellow pigmented Enterobacter
cloacae”. The reclassification to Enterobacter sakazakii was based on differences from E.
cloacae in DNA-DNA hybridization, biochemical reactions, pigment production and
antibiotic susceptibility. The micro-organism was named after the Japanese microbiologist
Riichi Sakazaki in honor of his work on the enteric bacteriology. Based on DNA-DNA
hybridization, 15 biogroups of 57 strains of E. sakazakii were described, with the wild type
Biogroup 1 being the most common (29). In 2007, there was a proposal for a novel genus
Cronobacter, comprising at least five genomospecies, including three subspecies. Also a
new biogroup (biogroup 16) was identified (41). Figure 3 gives an overview of the place of
the genus Cronobacter within the Enterobacteriaceae family. Throughout this thesis the
name Cronobacter spp. is used, covering the various strains which were earlier named
Enterobacter sakazakii.
The latest classification into Cronobacter gen. nov. is based on a polyphasic taxonomic
approach, employing full-length 16S rRNA gene sequencing, ribotyping, fluorescent-
amplified fragment length polymorphism (f-AFLP) and DNA-DNA hybridisation. Based
on this approach and the different phenotypic profiles, the novel Cronobacter species are
now divided in 16 biogroups.
The following five species are distinguished:
- Cronobacter sakazakii gen. nov., comb. nov. (Biogroup 1- 4, 7, 8, 11 and 13),
- Cronobacter malonaticus sp. nov. (Biogroup 5, 9, 14),
- Cronobacter dublinensis sp. nov. (Biogroup 6, 10, 12),
- Cronobacter muytjensii sp. nov. (Biogroup 15),
- Cronobacter turicensis sp. nov. (Biogroup 16).
Two Cronobacter strains appear to be a separate genomospecies and are indicated as
Cronobacter genomospecies I. This genomospecies has not been associated with a specific
biogroup.
Introduction

21
A number of subspecies of Cronobacter dublinensis are: Cronobacter dublinensis subsp.
dublinensis subsp. nov. (Biogroup 12), Cronobacter dublinensis subsp. lausannensis subsp.
nov. (Biogroup 10), and Cronobacter dublinensis subsp. lactaridi subsp. nov. (Biogroup 6).
For the full taxonomic description on the genus Cronobacter, including (sub) species, the
reader is referred to Iversen et al., 2008 (42). Table 2 provides a list of the Cronobacter
strains used throughout this research, including their new names.

Figur e 2. Schematic view of the place of Cronobacter spp. and Salmonella spp. within the
Enterobacteriaceae family. Adapted from B. Healy (personal communication).
Chapter 1
22
Table 2. Overview of the Cronobacter spp.
a
strains investigated in this thesis
Old str ain taxonomy and cultur e collection code(s) New str ain taxonomy
Enterobacter sakazakii NCTC 8155 Cronobacter sakazakii
Enterobacter sakazakii NCTC 9238 Cronobacter sakazakii
Enterobacter sakazakii NCTC 11467 = ATCC 29544 =
DSM 4485
Cronobacter sakazakii
Enterobacter sakazakii NCTC 9844 Cronobacter dublinensis
Enterobacter sakazakii NCTC 9846 Cronobacter dublinensis
Enterobacter sakazakii NCTC 9529 Cronobacter genomospecies 1
Enterobacter sakazakii DSM 18705 Cronobacter dublinensis subsp.
dublinensis
Enterobacter sakazakii ATCC 51329 Cronobacter muytjensii
Enterobacter sakazakii DSM 18703 Cronobacter turicensis
Enterobacter sakazakii DSM 18702T Cronobacter malonaticus
Enterobacter sakazakii DSM 18706T Cronobacter dublinensis subsp.
lausannensis
Enterobacter sakazakii DSM 18707T Cronobacter dublinensis subsp.
lactaridi
Enterobacter sakazakii MC10 Cronobacter spp.
Enterobacter sakazakii MM9 Cronobacter spp.
a
According to Iversen et al., 2008 (42).

Methodology
Detection and isolation
Cronobacter spp. can be present in low numbers in powdered infant formulae. Prevalences
varying from 2 to 14% have been reported (35, 56, 60). Concentrations of the organism
found in powdered infant formulae were between 0.2 and 92 CFU/100 g; levels greater than
1 CFU/g have not been reported (25).
Because of the low prevalence of Cronobacter spp. in PIF and powdered formula
producing facilities, highly specific methods are required for enumeration/detection and
identification that can be used for both product and environmental samples and that allow
quick isolation of Cronobacter spp.. At the start of the studies documented in this thesis,
there was a special need to improve the existing methods for isolation and detection of
Introduction

23
Cronobacter spp. shown in Table 3 (56, 57, 73). Many methods and media have been
described since then to isolate Cronobacter spp. (Chapter 9).

Table 3. Methods available in 2001 for detection of Cronobacter spp. in powdered infant formulae
Pr e-
enr ichment
Enr ichment
(usually selective)
Isolation Identification Ref.
Buffered
Peptone
Water
(BPW)
Enterobacteriaceae
enrichment (EE)
broth
Violet Red Bile
Glucose Agar
(VRBGA) plates
API 20
E
and yellow
colonies on Tryptone
Soy Agar (TSA)
plates
(56)
De-ionized
Water
EE broth VRBGA plates API 20
E
(60)

Methods for detecting Cronobacter spp. are still under development to increasingly better
meet the need to detect levels of approximately 1 cell per 100 g of PIF. There are three
basic steps in the detection of micro-organism in foods. The first one is pre-enrichment in a
non-selective medium allowing recovery of sub-lethally damaged cells. The second is
selective enrichment containing one or more compounds that are inhibitory to the majority
of micro-organisms, but significantly less to the species or group of species to be isolated.
The third is streaking of (selective) enrichment broth onto selective solid media.
These procedures may be used in different combinations, depending on the number of cells
expected in the sample. After the selective enrichment step, presumptive Cronobacter spp.
isolated on agar need to be confirmed. The success rate of the above mentioned basic
detection and isolation protocols depends on:
1) the number and the state of the micro-organisms in the sample,
2) the selectivity of the media (the balance between the inhibition of the competitors
and inhibition of the target organism),
3) the conditions of the incubation (temperature, time, oxygen availability),
4) the visual distinction of the isolation medium (the distinction between the
competitive flora and the target organism) (4).

Chapter 1
24
Confirmation methods
To confirm presumptive Cronobacter spp. detected on selective plating media, biochemical
tests, such as API 20
E
can de used that distinguish Cronobacter spp. from other species by
biochemical differentiation. A specific feature among Enterobacter species, namely the
presence of α-glucosidase activity, described by Muytjens et al., (57), was firstly employed
in the research conducted in this thesis as a suitable confirmation method for Cronobacter
spp. isolated from PIF – producing facilities (Chapter 2). The biochemical characteristic of
α-glucosidase production by Cronobacter spp. has remained the basis for many detection
media (22).
Much effort has been put on the development of molecular sub-typing techniques, like
Pulsed Field Gel Electrophorese (PFGE) (71), ribotyping and random amplification of
polymorphic DNA (RAPD) (15, 62). Molecular subtyping of bacteria by profiling either
proteins or nucleic acids could be a useful tool to investigate for example epidemiological
relationships of isolates (25). Furthermore, DNA fingerprints obtained with standerized
DNA- based protocols could be applied for direct comparison of isolates in outbreaks (70).
Recently, also several genotypic methods including the Polymerase Chain Reaction (PCR)
have been developed. These are based on amplifying specific fragments of DNA, allowing
detection of a single copy of the target sequence (18, 52, 53, 65, 68, 77). The PCR
confirmation method depends on the availability of two short oligonucleotide primer
sequences that will hybridize to opposite strands of heat-denatured DNA at either end of the
region which eventually be probed. A DNA polymerase then catalyses the extension of the
primers to produce two double-stranded copies of the region of interest. Recently, a number
of PCR probes for conserved genes such as ompA have been published (59). A potential
disadvantage of molecular techniques can be their sensitivity that can be influenced by food
components. Furthermore, molecular methods can not distinguish living from dead cells
and organisms killed for example by a heat process can still be detected. Also these
methods does not allow for isolation of the strains. Nevertheless, such methods can still be
useful in rapid detection of the presence of very low numbers of target micro-organisms.

Introduction

25
Production
Food production aims at delivering safe, nutritious and wholesome foods with an adequate
shelf-life and at reasonable costs to the consumer. By taking adequate control measures and
deploying sound assurance systems to manage food production, the food industry helps to
minimize risks to consumers associated with the potential presence of foodborne hazards in
foods. Food production processes often rely on control methods such as pasteurization or
sterilization to reduce the number of vegetative pathogens and/or spore-formers in the final
food product. Implementation of general food safety management systems such as Good
Hygiene Practices (GHP) and Good Manufacturing Practices (GMP) are considered in
many countries as pre-requisite conditions for facilities preparing or manufacturing foods.
To better assure safe food production, specific food safety management systems such as the
Hazard Analysis Critical Control Points (HACCP) system, have been deployed
successfully. Using these food safety management systems, the effectiveness of various
control measures is monitored to determine whether the process parameters for the control
measures remain within predetermined limits designed to achieve food safety (50). To
design adequate control measures and set-up effective food safety management systems it is
important to understand the different possible contamination routes of relevant hazards as
well as the efficacy of specific control measures to control them.

Production of Powdered Infant Formulae
Whereas it is generally accepted that breastfeeding is the best option for the nutrition and
the health of infants and young children, there may be situations where a mother can not
breastfeed her baby or chooses no to do so. In such cases, powdered formulae represent
useful alternatives to replace breast feeding either partially or totally and these are therefore
formulated to meet the nutritional needs of infants (78). Powdered formulae are divided in
powdered infant formula (PIF), follow-up formula (FUF), and dietary foods for special
medical purposes. PIF is defined as a breast milk substitute specially manufactured to
satisfy, by itself, the nutritional requirements of infants during the first months of life up to
the introduction of appropriate complementary feeding. FUF is defined as a food intended
Chapter 1
26
for use as a liquid part of the weaning diet for infants from the 6
th
month onwards and for
young children. Dietary foods for special medical purposes are defined as foods for
particular nutritional uses, intended for dietary management of patients and to be used
under medical supervision (12). The most significant difference between PIF and FUF is
that FUF may contain a wider variety of dry-mix ingredients.
The composition, quality and labelling requirements of powdered infant formulae are
clearly defined in national, regional as well as international standards or regulations (12).
Production of powdered formulae can follow three procedures: the dry-mix-, the wet-mix-
and the combined process. In the wet-mixing process, all unprocessed raw materials as well
as separately processed ingredients are handled as a liquid product that is heat treated, dried
and then further handled up to the filling stage. In the dry-mix process, all separate
ingredients are dry blended to obtain the final product, which is further handled up to the
filing process. The dry mixing is done when heat sensitive ingredients such as vitamins,
minerals, starch, carbohydrates and others need to be added, according to the formulation.
Ingredients as such have all been submitted to some thermal inactivation step during their
manufacture and must fulfil the same microbiological requirements as the final product.
The combined process may include and combine different mixing steps to obtain the final
formulation. In the combined process the unprocessed raw material and part of the
ingredients are processed according to the wet-mix process to obtain base powder, which is
further used for the manufacture of different finished products. The type of process used
depends mostly on the product manufactured and the processing facility. For a detailed
description of the production process the reader is referred to Cordier (2008) (16). A
schematic overview of PIF production is given in Figure 4.
Powdered formulae are produced with stringent control measures and strategies to prevent
post-process contamination. Guidance and recommendations concerning control and
elimination of Salmonella spp. and Cronobacter spp. in powdered infant formulae and
reconstituted infant formula have been issued in joint reports by the Food and Agriculture
Organization of the United Nations (FAO) and the World Health Organization (WHO) (12,
26, 27). The microbiological criteria, applied by the European Commission, EC No
1441/2007, for PIF are shown in Table 4. Monitoring for Enterobacteriaceae is used as
Introduction

27
process criterion that indicates the hygienic status of the production. For PIF for infants of 6
months and dietary products for special medical purposes parallel testing for
Enterobacteriaceae and Cronobacter spp. is advised, unless a correlation between micro-
organisms has been established at an individual plant level. If Enterobacteriaceae are
detected in any of the products samples tested in such a plant, the batch must be tested for
Cronobacter spp.. The detection of Enterobacteriaceae is a process hygiene parameter,
while detection of Salmonella spp. and Cronobacter spp. include food safety parameters
(24).

Table 4. Current European Union microbiological criteria for dried infant formulae for infants up to 6
months of age and for formulae for special medical purposes (24)

Micr obiological limit

Micr o-or ganism(s) n c m
Process hygiene
criteria
Enterobacteriaceae
(in 10-g samples)
10 0 Absence in 10 g

Food safety
criteria
Cronobacter spp.
(in 10-g samples)
30 0 Absence in 10 g

Salmonella
(in 25-g samples)
30 0 Absence in 25 g
n is the number of units comprising the sample.
c is number of acceptable sample units with values above m.

The recent review regarding the relationship between Cronobacter spp., Salmonella and
other Enterobacteriaceae in PIF and FUF undertaken by the scientific Panel on Biological
Hazards (BIOHAZ Panel) of the European Food Safety Authority (EFSA) concluded that
there is no correlation between Salmonella spp. and other Enterobacteriaceae, and that
there is also no apparent relationship between Cronobacter spp. and other
Enterobacteriaceae. However, at individual plant level such a correlation between
Cronobacter spp. and other Enterobacteriaceae may be established (22).
Chapter 1
28

Figur e 3. Schematic view of a production process for PIF, including both wet- and dry
mixing. Contamination during manufacture could in particular occur in the steps
highlighted in grey.
Introduction

29
Contamination routes relevant for PIF production
Severe illness and sometimes death of infants has been attributed to PIF that has been
contaminated with Cronobacter spp.. PIF has been found to be possibly contaminated
during addition of ingredients after the pasteurization process, during the packing process
or during reconstitution of the powder at homes or in the hospitals (14, 16, 54). Even
though Good Practices (GMP and GHP) and correct HACCP plans are adhered to (26, 27),
sporadically, PIF may be contaminated with this micro-organism. The relevant hazards in
PIF requiring appropriate control measures, as pointed out by the FAO and WHO
consultations, are Cronobacter spp. and Salmonella spp. (26, 27). Micro-organisms
belonging to the family of Enterobacteriaceae (which includes Salmonella spp. and
Cronobacter spp.) are known to be present in powdered infant formulae processing
environments and in homes or other places where PIF is reconstituted, such as hospitals
(14, 39, 45). The level of Enterobacteriaceae has been used for decades as an indicator for
process hygiene. Literature data have shown that the levels of Salmonella spp. in the
processing of PIF are much reduced in properly maintained high hygiene zones, and its
occurrence generally is rare. Contrary to Salmonella spp., complete elimination of
Cronobacter spp. during manufacture of PIF is, due to its ubiquitous nature, currently not
considered to be feasible in the high-hygiene zone (27). Indeed, Cronobacter spp. has been
detected in factory environments (46).
Following the production scheme depicted in Figure 4, Cronobacter spp. could enter
powdered infant formulae via the processing environment or the processing line and
through the addition of (contaminated) ingredients such as vitamins and starch after the
heat-inactivation step (16, 27). In this Figure, the production steps where contamination
could occur are depicted in grey. Next to contamination via ingredients, contamination
from the environment could occur via aerosols, through air containing dust or skin particles
of food handlers (17). The presence of Cronobacter spp. in the processing environment
implies that there may always be a chance of product contamination. As the micro-
organism occurs in homes and hospitals, powdered formulae can also be contaminated
during preparation of the feed at homes and at hospitals (27).
Chapter 1
30
Prevalence in factories
An adequate heating step is very important to significantly decrease the bacterial number in
raw milk. Since Cronobacter spp. is a vegetative micro-organism, the pasteurization step
should be sufficient to inactivate this micro-organism (9, 19, 40, 44, 61). This processing
step is particularly important because Cronobacter spp. may survive the spray drying
process (1). As air can be contaminated with bacteria, due to adherence to dust particles,
water droplets or even skin particles, product–air contact could thus cause contamination
after the pasteurization step. Cronobacter spp. may be introduced during dry cleaning of the
spray dryer and also via air that is in contact with the product after the spray drying step.
Cronobacter cells are able to survive in dry environments (45) such as powdered formulae
producing facilities (55). In these factories the numbers of Cronobacter cells and the
incidence of product contamination could increase after a wet cleaning process (16).

Prevalence outside factories
In order to identify potential vehicles for transmission, it is important to investigate
potential environmental reservoirs of Cronobacter spp.. The micro-organism seems to be
ubiquitously present as it has been isolated from a wide spectrum of environmental sources
and food products (22). It has been detected in raw and fresh products of animal and
vegetable origin as well as in processed and prepared foods such as dried, smoked, frozen,
fermented, cooked or fried products, ready-to-eat and street foods. It was also isolated from
fresh vegetables and spices (3). Primary Cronobacter spp. contamination may occur
intrinsically, e.g. due to its endophytical presence in plants or through contact with water,
soil and living vectors like insects or small vertebrates. Raw foods of animal origin may be
contaminated additionally via the (faecal) micro-flora of the food source animal itself (30).
In several investigations Cronobacter spp. was detected on kitchen equipment used for
preparing powdered infant formulae and also on apparently clean equipment (58, 71).
Furthermore, literature data showed that Cronobacter spp. was able to survive in a dish
brush (58). In homes, Cronobacter spp. was isolated from household clothes (47),
indicating that good hygiene practices are necessary to avoid contamination at homes and in
Introduction

31
hospital kitchen facilities. Because of its ubiquitous nature, it is necessary to consider
possible contamination during preparation and handling of prepared formulae.

Pr oduct handling
As the manufacture of commercially sterile PIF is not feasible using current processing
technologies, there is a low but potential risk of infection to infants through consumption of
PIF. This risk is particularly relevant when prepared food is handled or stored incorrectly.
Growth in PIF is not possible after the spray drying step because the water activity is too
low to allow bacterial growth. Although the micro-organism may be able to survive in PIF,
depending on the storage time and conditions, generally there is a gradual die-off.
However, reconstituted PIF forms an excellent growth medium for Cronobacter spp. and
other micro-organisms that may be present in such products. To avoid growth and further
contamination, leading to infection and illness in infants, specific preparation and handling
instructions are mentioned on the packaging label of powdered infant formulae.
Nevertheless, several investigations showed that professional and home caregivers have
their own way of preparing the formula. In hospitals, the preparation practices vary
according to local arrangements and the availability of adequately trained personnel (23).
Immediate consumption or rapid cooling and storage at low temperature are critical control
measures to prevent microbial growth once PIF is reconstituted.

Inactivation of Cronobacter spp. in dry powder
Cronobacter spp. has an unusual surviving ability under dry conditions compared to other
Enterobacteriaceae (9, 13, 20, 32), but there is a difference in the thermal tolerance
between Cronobacter strains (19, 61). Previous studies showed the occurrence of
Cronobacter spp. cells in the powdered infant formulae (39, 56, 60), and also that
Cronobacter spp. cells were able to survive for at least twelve months in powdered infant
formulae under favourable conditions (13, 20). Acidification could reduce the concentration
of Cronobacter spp. cells in different types of powdered formulae and vegetable based
Chapter 1
32
products (43, 67). Quantification of survival rates of Cronobacter spp. isolates in dry PIF
may provide useful data for microbiological risk assessments and for product/process
design studies, as it allows determining likely levels of the micro-organism in PIF
dependent on for instance different scenarios of time and temperature during storage and
distribution.

Growth of Cr onobacter spp. in reconstituted infant formulae
As acknowledged by the Food and Agricultural Organization and the World Health
Organization (26, 27), PIF is not a sterile product and several investigations have shown the
occurrence of cells of Cronobacter spp. in powdered infant formulae (35, 39, 56) leading to
it being associated to outbreaks after consumption of reconstituted infant formula (5, 8, 15,
36, 69, 75). Most probably, multiplication of the organism has taken place in these cases, as
only low numbers of the micro-organism have been detected in powdered infant formulae
(25). Literature data about the growth potential of Cronobacter spp. in reconstituted infant
formulae at various temperatures is scarce. To determine its effect on risk it is therefore
necessary to quantify the growth of Cronobacter spp. in reconstituted PIF. This information
could lead to recommendations to health professionals and parents on proper preparation,
handling, and storage of reconstituted PIF to ensure the product is safe when fed to the
infants.
In order to predict the microbial behaviour in a food product and to determine the effect of
growth on the product’s safety and quality, it is a necessity to have good insights in the
production, storage and handling conditions of PIF and to relate this understanding to
quantitative data regarding the micro-organisms ability to survive, recontaminate and grow.
Bacterial growth is defined as an orderly increase in the quantity of cellular constituents,
depending on the ability of cells to form new protoplasm from available nutrients. Bacterial
growth can be divided in four phases, being the lag phase, the exponential phase, the
stationary phase and the death phase. In the lag phase, bacteria can adapt to their new
environment and injured bacterial cells can recover. A wide variety of factors such as
inoculum size, time necessary to recover from physical damage or shock after transfer from
Introduction

33
other conditions, time needed for synthesis of essential (co-)enzymes or division factors and
time required for synthesis of new (inducible) enzymes that are necessary to metabolize the
substrates present in the medium, can influence the length of the lag phase. In the
exponential phase, cells divide at a constant rate depending on the composition of the
growth medium and incubation conditions such as temperature. In the stationary phase,
initially the cell number no longer increases but is kept in balance as a result of cells dying
(due to exhaustion of available nutrients and/or accumulation of inhibitory metabolites or
end products or exhaustion of space) and of cells multiplying. However, later in the
stationary phase a gradual reduction in cell numbers may be seen as the balance is lost.
The duration of the lag phase is of particular interest, as an extended lag phase can be a
means to prevent growth and multiplication of bacteria (10). Temperature is one of the
major environmental factors influencing microbial growth, and fluctuations in temperature
during storage and preparation may significantly affect the potential for outgrowth.
Mathematical models are useful tools to describe the growth (7). Quantification of growth
parameters, such as the specific growth rate as a function of temperature, help calculating
possible outgrowth of Cronobacter spp. in reconstituted infant formulae at various time-
temperature scenarios.

Hazar d Char acter ization
In the hazard characterization step the possible impact of an identified hazard on consumer
is determined, either qualitative or quantitative, using available dose-response relationships.
The likelihood and severity of illness is directly related to the number of micro-organisms
ingested, though in most cases this is a very variable relationship across consumer groups
and individuals as well as possibly across different strains of micro-organisms. In the case
of Cronobacter infections, the infectious dose relationship for humans has not been clearly
determined yet. The minimum infectious dose (MID) of Cronobacter spp. for infants, is
extrapolated from animal models (64): it estimated that high levels of the organism (> 10
5

CFU/feeding) are necessary to cause illness. In other studies it is speculated that a
reasonable estimate for infection might be close of that postulated for Escherichia coli
Chapter 1
34
O157:H7 and Listeria monocytogenes 4b, i.e. circa 1000 CFU (38). Retrospective studies
on epidemiological data and animal models could help to better estimate the dose-response
relationship of Cronobacter spp..

Risk char acter ization
In the risk characterization step the results of the previous three phases are integrated to
give an estimate of the probability of becoming ill after the food product, i.e reconstituted
infant formula is consumed. Results from risk characterization and especially comparisons
of the impact of different scenarios of exposure and possible mitigation could be used by
risk managers to decide on appropriate controls, including regulatory measures.
Furthermore, risk manager could decide whether to inform the consumer to implement
actions to reduce the risk on infection.

Outline of this thesis and the r esear ch conducted
This research project aimed to get more insight into the habitat and growth characteristics
of Cronobacter spp. and to study the factors that might ultimately affect consumer risk, in
particular the survival of Cronobacter spp. in dry powder and growth after reconstitution of
the powdered formula (Figure 2). Prior to this study, no specific enrichment method was
available and therefore an enrichment method was developed as described in Chapter 2. To
further improve enumeration and detection of Cronobacter spp., a more specific detection
method than available at the onset of the study was designed based on the presence of α-
glucosidase activity in the micro-organism (described in Chapter 3). The newly developed
methods were then applied to dry environmental samples from households and food
producing facilities (Chapter 4). To better explore the varied occurrence of the micro-
organism in food products, an extensive survey on the occurrence of Cronobacter spp. in
foods marketed in The Netherlands was conducted. In this survey, isolates were confirmed
using a Polymerase Chain Reaction (PCR) method and typed with Pulse Field Gel
Electrophoresis (PFGE) fingerprinting analysis (Chapter 5). As reported in Chapter 6, the
effect of the pre-culturing conditions on key growth parameters of Cronobacter spp. was
Introduction

35
quantified. Growth of Cronobacter spp. was then studied in reconstituted infant formulae
cooled in refrigerators with stagnant air (normal household refrigerators) and in a
refrigerator equipped with a fan (circulating air); growth parameters obtained earlier were
used in different scenarios to predict the possible growth of Cronobacter spp. in
reconstituted infant formula (Chapter 7). Finally, the survival and gradual inactivation of
Cronobacter spp. in dry powdered infant formula under different storage conditions was
investigated in detail and mathematically described (Chapter 8). In the general discussion,
the scientific insights from the thesis work are put into the perspective of general progress
in the field (Chapter 9).

Refer ences
1. Arku, B., N. Mullane, E. Fox, S. A. Fanning, and K. Jordan. 2008. Enterobacter
sakazakii survives spray drying. Int. J. Dairy Technol. 61:102-108.
2. Bar Oz, B., A. Preminger, O. Peleg, C. Block, and I. Arad. 2001. Enterobacter
sakazakii infection in the newborn. Acta Paediatr. 90:356-358.
3. Baumgartner, A., M. Grand, M. Liniger, and C. Iversen. 2009. Detection and
frequency of Cronobacter spp. (Enterobacter sakazakii) in different categories of
ready-to-eat foods other than infant formula. Int. J. Food Microbiol. 136:189-192.
4. Beumer, R. R. 1997. Listeria monocytogenes: detection and behavior in food and
in the environment. In: PhD thesis, Laboratory of Food Microbiology Wageningen
University, Wageningen, The Netherlands.
5. Biering, G., S. Karlsson, N. C. Clark, K. E. Jonsdottir, P. Ludvigsson, and O.
Steingrimsson. 1989. Three cases of neonatal meningitis caused by Enterobacter
sakazakii in powdered milk. J. Clin. Microbiol. 27:2054-2056.
6. Block, C., O. Peleg, N. Minster, A. Simhon, I. Arad, and M. Shapiro. 2002.
Cluster of neonatal infections in Jerusalem due to unusual biochemical variant of
Enterobacter sakazakii. Eur. J. Clin. Microbiol. Infect. Dis. 21:613-616.
7. Bovill, R., J. Bew, and J. Baranyi. 2001. Measurements and predictions of growth
for Listeria monocytogenes and Salmonella during fluctuating temperature II.
Rapidly changing temperatures. Int. J. Food Microbiol. 67:131-137.
8. Bowen, A. B., and C. R. Braden. 2006. Invasive Enterobacter sakazakii disease in
infants. Emerg. Infect. Dis. 12:1185-1189.
9. Breeuwer, P., A. Lardeau, M. Peterz, and H. Joosten. 2003. Desiccation and heat
tolerance of Enterobacter sakazakii. J. Appl. Microbiol. 95:967-973.
Chapter 1
36
10. Buchanan, R. L., and M. L. Cygnarowicz. 1990. A mathematical approach toward
defining and calculating the duration of the lag phase. Food Microbiol. 7:237-240.
11. CAC. 1999. Principles and guidelines for the conduct of microbiological risk
assessment. CAC/GL 30-1999.
12. CAC. 2008. Code of hygienic practice for powdered formulae for infants and
young children. CAC/RCP 66-2008.
13. Caubilla-Barron, J., and S. J. Forsythe. 2007. Dry stress and survival time of
Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered
infant formula. J. Food Prot. 70:2111-2117.
14. Cawthorn, D. M., S. Botha, and R. C. Witthuhn. 2008. Evaluation of different
methods for the detection and identification of Enterobacter sakazakii isolated
from South African infant formula milks and the processing environment. Int. J.
Food Microbiol. 127:129-138.
15. Clark, N. C., B. C. Hill, C. M. O'Hara, O. Steingrimsson, and R. C. Cooksey.
1990. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial
outbreaks. Diagn. Microbiol. Infect. Dis. 13:467-472.
16. Cordier, J. L. 2008. Production of powdered infant formulae and microbiological
control measures. p. 145-185. In: J.M. Farber and S.J. Forsythe (eds.),
Enterobacter sakazakii (emerging issues in food safety) ASM Press, Washington,
D.C., New Brunswick, New Jersey.
17. Den Aantrekker, E. D., R. M. Boom, M. H. Zwietering, and M. Van Schothorst.
2003. Quantifying recontamination through factory environments- a review. Int. J.
Food Microbiol. 80:117-130.
18. Drudy, D., M. O'Rourke, M. Murphy, N. R. Mullane, R. O'Mahony, L. Kelly, M.
Fischer, S. Sanjaq, P. Shannon, P. Wall, M. O'Mahony, P. Whyte, and S. Fanning.
2006. Characterization of a collection of Enterobacter sakazakii isolates from
environmental and food sources. Int. J. Food Microbiol. 110:127-134.
19. Edelson-Mammel, S. G., and R. L. Buchanan. 2004. Thermal inactivation of
Enterobacter sakazakii in rehydrated infant formula. J. Food Prot. 67:60-63.
20. Edelson-Mammel, S. G., and R. L. Buchanan. 2005. Survival of Enterobacter
sakazakii in a dehydrated powdered infant formula. J. Food Prot. 68:1900-1902.
21. EFSA. 2004. Opinion of the scientific panel on biological hazards on the request
from the commission related to the microbiological risks in infant formulae and
follow-on formulae. The EFSA Journal. 113:1-35.
22. EFSA. 2007. Opinion of the scientific panel on biological hazards on the request
from the commission for review of the opinion on microbiological risks in infant
formulae and follow-on formulae with regard to Enterobacteriaceae as indicators.
The EFSA Journal. 444:1-14.
Introduction

37
23. Estuningsih, S., and N. Abdullah Sani. 2008. Powdered infant formula in
developing and other countries - Issues and prospects. p. 221-234. In: J.M. Farber
and S.J. Forsythe (eds.), Enterobacter sakazakii (emerging issues in food safety)
ASM Press, Washington D.C., New Brunswick, New Jersey.
24. European Commission (EC). 2007. Commission Regulation (EC) No. 1441/2007
of 5 December 2007 amending Regulation (EC) No 2073/2005 on microbiological
criteria for foodstuffs Official Journal of the European Union. L 322/12.
25. Fanning, S., and S. J. Forsythe. 2008. Isolation and identification of Enterobacter
sakazakii. p. 27-59. In: J.M. Farber and S.J. Forsythe (eds.), Enterobacter
sakazakii (emerging issues in food safety) ASM Press, Washington, D.C., New
Brunswick, New Jersey.
26. FAO/WHO. 2004. Enterobacter sakazakii and Other Micro-organisms in
Powdered Infant Formula: meeting report. Microbiological Risk Assessment
Series No. 6, Rome, Italy.
27. FAO/WHO. 2006. Enterobacter sakazakii and Salmonella in powdered infant
formula: meeting report. Microbiological Risk Assessment Series No. 10 Rome,
Italy.
28. FAO/WHO. 2008. Enterobacter sakazakii (Cronobacter spp.) in powdered follow-
up formulae: meeting report. Microbiological Risk Assessment Series No. 15
Rome, Italy.
29. Farmer III, J. J., M. A. Asbury, F. W. Hickman, D. J. Brenner, and the
Enterobacteriaceae study group. 1980. Enterobacter sakazakii: A new species of
"Enterobacteriaceae" isolated from clinical specimens. Int. J. Syst. Bacteriol.
30:569-584.
30. Friedemann, M. 2007. Enterobacter sakazakii in food and beverages (other than
infant formula and milk powder). Int. J. Food Microbiol. 116:1-10.
31. Friedemann, M. 2009. Epidemiology of invasive neonatal Cronobacter
(Enterobacter sakazakii) infections. Eur. J. Clin. Microbiol. Infect. Dis. 28:1297-
1304.
32. Gurtler, J. B., and L. R. Beuchat. 2007. Survival of Enterobacter sakazakii in
powdered infant formula as affected by composition, water activity, and
temperature. J. Food Prot. 70:1579-1586.
33. Gurtler, J. B., J. L. Kornacki, and L. R. Beuchat. 2005. Enterobacter sakazakii: A
coliform of increased concern to infant health. Int. J. Food Microbiol. 104:1-34.
34. Hawkins, R. E., C. R. Lissner, and J. P. Sanford. 1991. Enterobacter sakazakii
bacteremia in an adult. South. Med. J. 84:793-795.
35. Heuvelink, A. E., M. Ahmed, F. D. Kodde, J. T. M. Zwartkruis-Nahuis, and E. de
Boer. 2002. Enterobacter sakazakii in melkpoeder. De Ware(n)chemicus. 32:17-
30.
Chapter 1
38
36. Himelright, I., E. Harris, V. Lorch, and M. Anderson. 2002. Enterobacter
sakazakii infections associated with the use of powdered infant formula-
Tennessee, 2001. JAMA. 287:2204-2205.
37. ICMSF. 2002. Micro-organisms in Foods 7: Microbiological Testing in Food
Safety Management. Springer Publishers, New York.
38. Iversen, C., and S. J. Forsythe. 2003. Risk profile of Enterobacter sakazakii, an
emerging pathogen associated with infant formula. Trends Food Sci. Technol.
14:443-454.
39. Iversen, C., and S. J. Forsythe. 2004. Isolation of Enterobacter sakazakii and other
Enterobacteriaceae from powdered infant formula milk and related products.
Food Microbiol. 21:771-777.
40. Iversen, C., M. Lane, and S. J. Forsythe. 2004. The growth profile,
thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant
formula milk. Lett. Appl. Microbiol. 38:378-382.
41. Iversen, C., A. Lehner, N. Mullane, E. Bidlas, I. Cleenwerck, J. Marugg, S.
Fanning, R. Stephan, and H. Joosten. 2007. The taxonomy of Enterobacter
sakazakii: proposal of a new genus Cronobacter gen. nov. and descriptions of
Cronobacter sakazakii comb. nov. Cronobacter sakazakii subsp sakazakii, comb.
nov., Cronobacter sakazakii subsp malonaticus subsp nov., Cronobacter turicensis
sp nov., Cronobacter muytjensii sp nov., Cronobacter dublinensis sp nov and
Cronobacter genomospecies I. BMC Evol. Biol. 7:64.
42. Iversen, C., N. Mullane, B. McCardel, B. D. Tal, A. Lehner, S. Fanning, R.
Stephan, and H. Joosten. 2008. Cronobacter gen. nov., a new genus to
accommodate the biogroups of Enterobacter sakazakii, and proposal of
Cronobacter sakazakii gen. nov., comb. nov., Cronobacter malonaticus sp. nov.,
Cronobacter turicensis sp. nov., Cronobacter muytjensii sp. nov., Cronobacter
dublinensis sp. nov., Cronobacter genomospecies 1, and of three subspecies,
Cronobacter dublinensis subsp. Dublinensis subsp. nov., Cronobacter dublinensis
subsp. lausannensis subsp. nov. and Cronobacter dublinensis subsp. lactaridi
subsp. nov. . Int. J. Syst. Evol. Microbiol. 58:1442-1447.
43. Joosten, H. M., and A. Lardeau. 2004. Enhanced microbiological safety of
acidified infant formulas tested in vitro. SAJCN. 17:87-92.
44. Jung, M. K., and J. H. Park. 2006. Prevalence and thermal stability of
Enterobacter sakazakii from unprocessed ready-to-eat agricultural products and
powdered infant formulas. Food Sci. Biotechnol. 15:152-157.
45. Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004. Occurrence of Enterobacter sakazakii in the food production
environments and households. The Lancet. 363:39-40.
46. Kandhai, M. C., M. W. Reij, K. Van Puyvelde, O. Guillaume-Gentil, R. R.
Beumer, and M. Van Schothorst. 2004. A new protocol for the detection of
Introduction

39
Enterobacter sakazakii applied to environmental samples. J. Food Prot. 67:1267-
1270.
47. Kusumaningrum, H. D. 2003. Behaviour and cross-contamination of pathogenic
bacteria in household kitchens - relevance to exposure assessment. In: PhD Thesis,
Laboratory of Food Microbiology Wageningen University, Wageningen, The
Netherlands.
48. Lai, K. K. 2001. Enterobacter sakazakii infections among neonates, infants,
children, and adults - Case reports and a review of the literature. Medicine.
80:113-122.
49. Lammerding, A. M. 1997. An overview of microbial food safety risk assessment.
J. Food Prot. 60:1420-1425.
50. Lammerding, A. M. 2007. Using microbiological risk assessment (MRA) in food
safety management. In: K. Yates (ed.), ILSI Europe Report Series ILSI, Brussel.
51. Lammerding, A. M., and A. Fazil. 2000. Hazard identification and exposure
assessment for microbial food safety risk assessment Int. J. Food Microbiol.
58:147-157.
52. Lehner, A., K. Riedel, T. Rattei, A. Ruepp, D. Frishman, P. Breeuwer, B. Diep, L.
Eberl, and R. Stephan. 2006. Molecular characterization of the alpha-glucosidase
activity in Enterobacter sakazakii reveals the presence of a putative gene cluster
for palatinose metabolism. Syst. Appl. Microbiol. 29:609-625.
53. Liu, Y., X. N. Cai, X. Zhang, Q. L. Gao, X. C. Yang, Z. J. Zheng, M. H. Luo, and
X. T. Huang. 2006. Real time PCR using TaqMan and SYBR Green for detection
of Enterobacter sakazakii in infant formula. J. Microbiol. Methods. 65:21-31.
54. Mullane, N., C. Iversen, B. Healy, C. Walsh, P. Whyte, P. G. Wall, T. Quinn, and
S. Fanning. 2007. Enterobacter sakazakii: an emerging bacterial pathogen with
implications for infant health. Minerva Pediatr. 59:137-148.
55. Mullane, N. R., P. Whyte, P. G. Wall, T. Quinn, and S. Fanning. 2007.
Application of pulsed-field gel electrophoresis to characterise and trace the
prevalence of Enterobacter sakazakii in an infant formula processing facility. Int.
J. Food Microbiol. 116:73-81.
56. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
57. Muytjens, H. L., J. Van der Ros-van de Repe, and H. A. Van Druten. 1984.
Enzymatic profiles of Enterobacter sakazakii and related species with special
reference to the alpha-glucosidase reaction and reproducibility of the test system.
J. Clin. Microbiol. 20:684-686.
Chapter 1
40
58. Muytjens, H. L., H. C. Zanen, H. J. Sonderkamp, L. A. Kollee, I. K. Wachsmuth,
and J. J. Farmer, 3rd. 1983. Analysis of eight cases of neonatal meningitis and
sepsis due to Enterobacter sakazakii. J. Clin. Microbiol. 18:115-120.
59. Nair, M. K. M., and K. S. Venkitanarayanan. 2006. Cloning and Sequencing of the
ompA gene of Enterobacter sakazakii and development of an ompA-targeted PCR
for rapid detection of Enterobacter sakazakii in infant formula. Appl. Environ.
Microbiol. 72:2539-2546.
60. Nazarowec-White, M., and J. M. Farber. 1997. Incidence, survival, and growth of
Enterobacter sakazakii in infant formula. J. Food Prot. 60:226-230.
61. Nazarowec-White, M., and J. M. Farber. 1997. Thermal resistance of Enterobacter
sakazakii in reconstituted dried-infant formula. Lett. Appl. Microbiol. 24:9-13.
62. Nazarowec-White, M., and J. M. Farber. 1999. Phenotypic and genotypic typing
of food and clinical isolates of Enterobacter sakazakii. J. Med. Microbiol. 48:559-
567.
63. Pagotto, F. J., J. M. Farber, and R. F. Lenati. 2008. Pathogenicity of Enterobacter
sakazakii. p. 127-144. In: J.M. Farber, and S.J. Forsythe (eds.), Enterobacter
sakazakii (emerging issues in food safety) ASM Press, Washington, D.C., New
Brunswick, New Jersey.
64. Pagotto, F. J., M. Nazarowec White, A. Bidawid, and J. M. Farber. 2003.
Enterobacter sakazakii: infectivity and enterotoxin production in vitro and in vivo.
J. Food Prot. 66:370-375.
65. Proudy, I., D. Bougle, E. Coton, M. Coton, R. Leclercq, and M. Vergnaud. 2008.
Genotypic characterization of Enterobacter sakazakii isolates by PFGE, BOX-
PCR and sequencing of the fliC gene. J. Appl. Microbiol. 104:26-34.
66. Reij, M. W., and M. H. Zwietering. 2008. Integrating concepts: a case study using
Enterobacter sakazakii in infant formula. p. 177 - 204. In: D.W. Schaffner (ed.),
Microbial Risk Analysis of Foods ASM Press, Washington DC, New Brunswick,
New Jersey.
67. Richards, G. M., J. B. Gurtler, and L. R. Beuchat. 2005. Survival and growth of
Enterobacter sakazakii in infant rice cereal reconstituted with water, milk, liquid
infant formula, or apple juice. J. Appl. Microbiol. 99:844-850.
68. Seo, K. H., and R. E. Brackett. 2005. Rapid, specific detection of Enterobacter
sakazakii in infant formula using a real-time PCR assay. J. Food Prot. 68:59-63.
69. Simmons, B. P., M. S. Gelfand, M. Haas, L. Metts, and J. Ferguson. 1989.
Enterobacter sakazakii infections in neonates associated with intrinsic
contamination of a powdered infant formula. Infect. Control Hosp. Epidemiol.
10:398-401.
Introduction

41
70. Singh, A., R. V. Goering, S. Simjee, S. L. Foley, and M. J. Zervos. 2006.
Application of molecular techniques to the study of hospital infection. Clin.
Microbiol. Rev. 19:512-530.
71. Smeets, L. C., A.Voss, H. L.Muytjens, J. F. G. M. Meis, W. J. G Melchers. 1998.
Genetisch karakterisatie van Enterobacter sakazakii isolaten van Nederlandse
patiënten met neonatale meningitis. Ned. Tijdschr. Med. Microbiol. 6:113-115.
72. Stoll, B. J., N. Hansen, A. A. Fanaroff, and J. A. Lemons. 2004. Enterobacter
sakazakii is a rare cause of neonatal septicemia or meningitis in VLBW infants J.
Pediatr. 144:821-823.
73. U.S. Food Drug and Administration (FDA). 2002. Isolation and enumeration of
Enterobacter sakazakii from dehydrated powdered infant formula. Accessed 10
February, 2009.
74. Urmenyi, A. M. C., and M. D. Berne. 1961. Neonatal death from pigmented
coliform infection. The Lancet. 1:313-315.
75. Van Acker, J., F. de Smet, G. Muyldermans, A. Bougatef, A. Naessens, and S.
Lauwers. 2001. Outbreak of necrotizing enterocolitis associated with Enterobacter
sakazakii in powdered milk formula. J. Clin. Microbiol. 39:293-297.
76. Weir, E. 2002. Powdered infant formula and fatal infecton with Enterobacter
sakazakii. Can. Med. Assoc. J. 166:1570.
77. Williams, T. L., S. R. Monday, S. Edelson-Mammel, R. Buchanan, and S. M.
Musser. 2005. A top-down proteomics approach for differentiating thermal
resistant strains of Enterobacter sakazakii. Proteomics. 5:4161-4169.
78. World Health Organisation (WHO). 1981. International code of marketing of
breast-milk substitutes. Available at:
http://www.who.int/nutrition/publications/code_english.pdf. Accessed June 2009.


Chapter 1
42
Protocol for the detection of Cronobacter spp.

43
Chapter 2
44
Abstr act
Cronobacter spp. is a motile, peritrichous, Gram-negative rod that was previously known
as a “yellow pigmented Enterobacter cloacae”. It is documented as a rare cause of
outbreaks and sporadic cases of life-threatening neonatal meningitis, necrotizing
enterocolitis, and sepsis. Cronobacter spp. have been isolated from milk powder–based
formulas, and there is thus a need to investigate whether and where Cronobacter spp.
occurs in these manufacturing environments. For this purpose, a simple detection method
was developed based on two features of Cronobacter spp.: its yellow pigmented colonies
when grown on tryptone soy agar and its constitutive o-glucosidase, which is detected in a
4-h colorimetric assay. Using this screening method, Cronobacter spp. strains were isolated
from three individual factories from 18 of 152 environmental samples, such as scrapings
from dust, vacuum cleaner bags, and spilled product near equipment. The method is useful
for routine screening of environmental samples for the presence of Cronobacter spp..





This Chapter has been published as
“A new pr otocol for the detection of Enterobacter sakazakii* applied to envir onmental
samples”

Kandhai, M.C., M.W. Reij, K. Van Puyvelde, O. Guillaume-Gentil, R.R. Beumer, and
M. Van Schothorst.
Journal of Food Protection, 2004; 67:6; 1267–1270.

* Throughout this thesis the designation Cronobacter spp. has been used to indicate the
taxonomic change.
Protocol for the detection of Cronobacter spp.

45
Intr oduction
Cronobacter spp. is a member of the family Enterobacteriaceae. Until 1980, this organism
was referred to as yellow pigmented Enterobacter cloacae. It was then reclassified as a
unique species based on differences from E. cloacae in DNA relatedness, the specific
yellow pigment production, biochemical reactions, and antibiotic susceptibility (4). Several
outbreaks or sporadic cases of either severe neonatal meningitis in premature infants or
necrotizing enterocolitis have been attributed to Cronobacter spp. (7). In some of these,
contaminated dry infant formulas have been identified as the source of Cronobacter spp. (3,
13). Since heat treatments such as pasteurization readily kill the micro-organism (11),
contamination must have occurred after processing. It is suspected that this micro-organism
is present in the environment of the processing equipment (10, 11).
Most Cronobacter spp. strains described in the literature were isolated from cerebrospinal
fluid of the patients (7, 13). Isolates were commonly obtained by streaking onto blood or
chocolate agar, using the API 20
E
system for identification. The isolation of Cronobacter
spp. in milk powder is usually performed by enrichment in Enterobacteriaceae enrichment
broth, followed by plating on violet red bile glucose agar and sub-culturing on sheep blood
agar and eosin methylene blue agar (8) or on tryptone soy agar (TSA) (12). Often, isolates
were screened for the presence of DNase activity on toluidine blue agar after 2 and 7 days
of incubation at 36 °C and for the formation of yellow pigmented colonies on TSA at 25 °C
for 48 h (4, 9). An additional characteristic feature is the production of Tween 80–esterase
in 3 to 8 days (1). In all, this full confirmation protocol takes more than a week. A faster
and elective protocol is preferred for the detection of Cronobacter spp. in environmental
samples in food production, in which a variety of related coliforms may be present. This
study describes a simple method for the detection of Cronobacter spp. and its application to
environmental samples collected from three milk powder production plants.
Chapter 2
46
Mater ials and methods
Detection
A total of 152 dry samples were obtained from the environments of three milk powder
production plants. They were collected from floor sweepings, spilled dry products,
scrapings, or vacuum cleaner bags. Approximately 65% of the samples were analyzed for
the presence of Cronobacter spp. without enrichment, whereas for 35% an enrichment step
was used. In case of enrichment, 10 g of dry sample was homogenized in 90 ml of buffered
peptone water (Oxoid Ltd., Basingstoke, Hampshire, UK), incubated for 18 to 20 h at 37
°C, streaked onto violet red bile agar (VRBL; Oxoid), and incubated for 20 to 24 h at 37
°C, as shown in Figure 1. Other samples were streaked directly onto VRBL after
homogenizing. From each VRBL plate, 10 colonies typical for coliforms were purified on
TSA (Oxoid). On many plates, however, fewer than 10 colonies were present, and in that
case all colonies were purified on plates. The TSA plates were incubated for 48 h in
daylight at room temperature. The yellow pigmentation on TSA is a characteristic feature
of Cronobacter spp..

Preliminary identification
Isolates were tested for oxidase, using Oxidase DrySlide OXIDASE (Becton Dickinson and
Company, Sparks, Md.). Oxidase-negative isolates were further identified using the API
20
E
test system (bioMérieux SA, Marcy l’Etoile, France) and the corresponding
identification software (API Lab Plus version 3.3.3), using Cronobacter sakazakii DSM
4485, corresponding to ATCC 29544, as the type strain.

α-Glucosidase activity
For the assessment of α-glucosidase activity, paranitrophenyl-α-D-glucopyranoside (Fluka
Chemie Gmbh, Buchs, Switzerland) was dissolved in distilled water at 50 °C and added to
0.3 M (pH 7.0) phosphate buffer (Merck, KgaA, Darmstadt, Germany) in a final
Protocol for the detection of Cronobacter spp.

47
concentration of 4 g/L of buffer. Individual colonies grown on TSA were suspended in 2 ml
of physiological salt solution, 0.85% NaCl (Merck) in distilled water, whereupon 2 ml of
the paranitrophenyl-α-D-glucopyranoside solution was added. Care was taken to transfer a
whole colony to the test tube. The use of a standardized inoculum level of at least number 1
on the McFarland scale is recommended. The mixture was then incubated in a water bath at
37 °C. The formation of the yellow colored paranitrophenyl (PNP) hydrolysate was
measured after 0, 4, and 24 h using a Novaspec II spectrophotometer (Pharmacia Biotech,
Cambridge, UK) at 405 nm. A minimal absorption of 0.3 at 405 nm after 4 h, equivalent to
16 µM PNP, was considered positive. Positive and negative controls were included as
described below.
For some isolates, the α-glucosidase activity assay was also performed with
paranitrophenyl-α-D-glucopyranoside tablets as follows. Five yellow, freshly grown
colonies on TSA were individually tested for α-glucosidase activity. Each colony was
suspended in glass tubes (10 by 100 mm; Emergo, Landsmeer, The Netherlands) containing
0.25 ml of a 0.85% NaCl solution and a ready-to-use diagnostic tablet (Diatab 50421,
Rosco, Taastrup, Denmark). Negative controls were included each time consisting of a
suspended tablet without bacteria and another one with an E. cloacae 218 strain that is α-
glucosidase negative. A positive control was also included each time consisting of a
suspended tablet with Cronobacter sakazakii DSM 4485. After vigorous vortexing for a
few seconds, tubes were incubated in a water bath at 37 °C for 4 ± 0.1 h. A yellow color in
the supernatant, caused by the release of PNP, indicated the presence of α-glucosidase.

Ribotyping
An automated ribotyping system, the RiboPrinter (DuPont Qualicon,Wilmington, Del.),
was used to confirm the identity of presumptive Cronobacter spp. isolates (2). In this
system DNA was extracted from a colony and then digested with EcoRI into discrete-sized
fragments. The DNA was then transferred to a membrane and probed with a region of the
rRNA operon to reveal the pattern of rRNA genes. The pattern was recorded, digitized, and
stored in a database containing 6 DuPont Cronobacter spp. reference fingerprints combined
Chapter 2
48
with the 2,000 in-house Cronobacter spp. fingerprints (5). Comparison of patterns in the
database allows for the assessment of relatedness between Cronobacter spp. strains and
provides further evidence of the identity of the isolates in addition to the biochemical
identification.


Figur e 1. Schematic diagram for the enrichment and identification of Cronobacter spp..
Protocol for the detection of Cronobacter spp.

49
Results and discussion
From the line environment of three milk powder plants, 152 dry samples were collected.
Samples were analyzed with one of the two procedures, namely, a pre-enrichment step in
buffered peptone water followed by streaking on VRBL agar as described in Figure 1 or a
direct plating of diluted samples on VRBL agar. Coliforms were detected with both
procedures. Some of these coliforms were subsequently identified as presumptive
Cronobacter spp. on the basis of oxidase test and biochemical identification with API 20
E
.
A total of 100 samples were not pre-enriched and yielded in total 30 coliforms, 7 of which
were presumptive Cronobacter spp. strains. Fifty-two samples were subjected to pre-
enrichment, coliforms were isolated from 33 samples, and nine of the coliforms were
presumptively Cronobacter spp.. From the data presented here, it appears that Cronobacter
spp. can be isolated from environmental samples with and without pre-enrichment. In
another study, 27 samples were analyzed using both methods. Presumptive Cronobacter
spp. strains were isolated from 13 of 27 samples without enrichment and from 16 of the
same 27 samples with enrichment (data not published).
From VRBL agar, all colonies were streaked on TSA plates and were identified with API
20
E
. The API 20
E
system was not able to identify isolates obtained from 38 samples. From
samples positive for Cronobacter spp., numerous Cronobacter spp. isolates were obtained
in several cases that could be differentiated by their API 20
E
profiles and their ribotyping
fingerprints. As shown in Table 1, many different species and strains of coliforms were
found. Among the Enterobacter species, the presence of α-glucosidase is known to be
specific for Cronobacter spp. (9). The combination of the yellow colonies on TSA and
positive α-glucosidase activity was therefore considered further evidence for the identity of
Cronobacter spp.. All 32 presumptive Cronobacter spp. isolates and 18 other coliforms,
which had been isolated from environmental samples and had been identified using the API
20
E
system, were tested for their α-glucosidase activity.
Chapter 2
50
Table 1. Presumptive identification of coliforms isolated from environmental samples taken in three
milk powder factories
a

Species Factor y A Factor y B Factor y C
No. of samples 56 49 47
Chryseomonas luteola 2
Citrobacter diversus 1 3
Enterobacter amnigenus 1
Enterobacter cloacae 5 8 9
Cronobacter spp.
b
8 (16)
c
4 (6)
c
6 (10)
c

Erwinia nigrifluens 2
Escherichia coli 1 2
Escherichia hermanii 1 1
Escherichia vulneris 1 3
Klebsiella ornithinolytica 2
Klebsiella oxytoca 1 1
Klebsiella pneumoniae 2 1
Kluyvera spp. 1
Leclercia adecarboxylata 1 1
Pantoea spp. 2 5 6
Rahnella aquatilis 1 1
Serratia ficaria 3 2
Unidentified 7 10 21
a
Identification was based on oxidase test and API 20
E
. The Table indicate the number of samples
from which a certain species was isolated.
b
Please note that 3 of the 32 presumptive isolates that were identified as Cronobacter spp. by the API
20
E
system did not show α-glucosidase activity and were not confirmed to be Cronobacter spp. by
their ribotype fingerprint.
c
Implies that 16, 6, and 10 different isolates were obtained from the samples based on ribotyping
results.

As shown in Table 2, 29 Cronobacter spp. isolates hydrolyzed paranitrophenyl-α-D-
glucopyranoside within 4 h. None of the other coliforms were positive in the 4-h assay.
After 24 h, the difference between Cronobacter spp. and the other coliforms was less clear.
One Pantoea isolate and one strain of Erwinia showed substantial α-glucosidase activity to
be rated positive in the 24-h assay. This indicates that the specificity of the α-glucosidase
Protocol for the detection of Cronobacter spp.

51
activity of Cronobacter spp. only applies for a short-term assay. No false-positive results
were observed with the 4-h α-glucosidase assay.

Table 2. Yellow-pigment formation and the o-glucosidase assay after 4 and 24 h incubation by
strains grown on TSA
Alpha-glucosidase
a

Pr esumptive identification
(no. of str ains)
Yellow-pigment
on TSA 4 h 24 h
Citrobacter diversus (3) + - -
Cronobacter spp. (29) + + +
Cronobacter spp. (3)
b
+ - -
Enterobacter cloacae (4) - - -
Erwinia nigrifluens (1) + - -
Erwinia nigrifluens (1) + - +
Escherichia coli (1) - - -
Escherichia hermanii (1) + - -
Escherichia vulneris (1) - - -
Klebsiella ornithinolytica (1) - - -
Klebsiella oxytoca (1) - - -
Klebsiella pneumoniae (1) + - -
Pantoea spp. (1) + - +
Pantoea spp. (1) + - -
Rahnella aquatilis (1) + - -
a
+ : paranitrophenol concentration (PNP) greater than 16 µM, as measured by OD at 405 nm.
b
Presumptively identified as Cronobacter spp. by the API 20E system but not confirmed by.
ribotyping or by α-glucosidase presence.

Three presumptive Cronobacter spp. isolates were α-glucosidase negative after 4 h. These
three isolates were clearly α-glucosidase negative after retesting, although they were
identified as Cronobacter spp. by the API 20
E
system with a low percentage of match
(80%) with the expected biochemical profile of Cronobacter spp..
The presumptive Cronobacter spp. strains were also characterized using the ribotyping
technique. Ribotyping is a method that generates a highly reproducible and precise
fingerprint of bacterial rRNA genes. Using a database of reference fingerprints, it can be
Chapter 2
52
used to classify and identify bacteria. The technology can be used here as a further tool,
next to yellow pigmentation, oxidase/API 20
E
, and the α-glucosidase assay, to confirm the
identity of Cronobacter spp. and also investigate relatedness among Cronobacter spp.
isolates.
The results showed that 27 presumptive Cronobacter spp. isolates were confirmed to be
Cronobacter spp. by ribotyping. Two other isolates gave unclear fingerprints due most
likely to the lack of DNA digestion by the EcoRI enzyme. The remaining three strains were
the three α-glucosidase-negative isolates and indeed had fingerprints that were significantly
different from the six reference Cronobacter spp. strains of the Riboprinter database. Based
on the absence of α-glucosidase and the ribotyping results, it can be concluded that the
three isolates identified by the API 20
E
system as Cronobacter spp. were not Cronobacter
spp.. It should be noted that at least a 90% of biochemical reaction of the API 20
E
is needed
for a reliable identification. The results of the ribotyping also indicated that in none of the
three factories was a specific strain predominant (results not shown).
In summary, it can be concluded that the combination of yellow pigmented colonies on
TSA and the presence of α-glucosidase activity detected in a 4-h assay are a good method
to differentiate Cronobacter spp. from other coliforms. This method is useful for routine
screening of environmental samples taken from milk powder factories and most likely also
for their dry milk powder–based formulas (6). To ensure the detection of low numbers of
Cronobacter spp. in environmental samples, work is currently ongoing to optimize the
enrichment step.

Acknowledgments
Nadine Braendlin is kindly acknowledged for her expert advice and her help in obtaining
samples. The authors thank Wilma Hazeleger, Prof. dr. Leon Gorris, and Dr. Matz Peterz
for critically reading and reviewing the manuscript.

Protocol for the detection of Cronobacter spp.

53
Refer ences
1. Aldova, E., O. Hausner, and R. Postupa. 1983. Tween-Esterase activity in
Enterobacter sakazakii. Zbl. Bakt. Hyg. A 256:103-108.
2. Bruce, J. 1996. Automated system rapidly identifies and characterizes micro-
organisms in food. Food Technol. 50:77-81.
3. Clark, N. C., B. C. Hill, C. M. O'Hara, O. Steingrimsson, and R. C. Cooksey.
1990. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial
outbreaks. Diagn. Microbiol. Infect. Dis. 13:467-472.
4. Farmer III, J. J., M. A. Asbury, F. W. Hickman, D. J. Brenner, and the
Enterobacteriaceae study group. 1980. Enterobacter sakazakii: A new species of
"Enterobacteriaceae" isolated from clinical specimens. Int. J. Syst. Bacteriol.
30:569-584.
5. Guillaume-Gentil, O., P. Scheldeman, J. Marugg, L. Herman, H. Joosten, and M.
Heyndrickx. 2002. Genetic heterogeneity in Bacilllus sporothermodurans as
demonstrated by ribotyping and repetitive extragenic palindromic-PCR
fingerprinting. Appl. Environ. Microbiol. 68:4216-4224.
6. Heuvelink, A. E., M. Ahmed, F. D. Kodde, J. T. M. Zwartkruis-Nahuis, and E. De
Boer. 2002. Enterobacter sakazakii in melkpoeder. De Ware(n)chemicus. 32:17-
30.
7. Lai, K. K. 2001. Enterobacter sakazakii infections among neonates, infants,
children, and adults - Case reports and a review of the literature. Medicine.
80:113-122.
8. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
9. Muytjens, H. L., J. Van der Ros-Van de Repe, and H. A. Van Druten. 1984.
Enzymatic profiles of Enterobacter sakazakii and related species with special
reference to the alpha-glucosidase reaction and reproducibility of the test system.
J. Clin. Microbiol. 20:684-686.
10. Nazarowec-White, M., and J. M. Farber. 1997. Thermal resistance of Enterobacter
sakazakii in reconstituted dried-infant formula. Lett. Appl. Microbiol. 24:9-13.
11. Skladal, P., M. Mascini, C. Salvadori, and G. Zannoni. 1993. Detection of
bacterial contamination in sterile UHT milk using an L-lactate biosensor. Enzyme
Microb. Technol. 15:508-512.
12. U.S. Food and Drug Administration. 2002. Isolation and enumeration of
Enterobacter sakazakii from dehydrated powdered infant formula. Accessed 10
February, 2009.
Chapter 2
54
13. Van Acker, J., F. de Smet, G. Muyldermans, A. Bougatef, A. Naessens, and S.
Lauwers. 2001. Outbreak of necrotizing enterocolitis associated with Enterobacter
sakazakii in powdered milk formula. J. Clin. Microbiol. 39:293-297.
Enrichment procedure for Cronobacter spp.

55
Chapter 3
56
Abstr act
A method was developed to detect and identify Cronobacter spp. in environmental samples.
The method is based on selective enrichment at 45 ± 0.5 °C in lauryl sulfate tryptose broth
supplemented with 0.5 M NaCl and 10 mg/L vancomycin (mLST) for 22 to 24 h followed by
streaking on tryptone soy agar with bile salts. When exposed to light during incubation at
37 °C, Cronobacter spp. produces yellow colonies within 24 h; identification was confirmed
by testing for a-glucosidase activity and by using API 20
E
strips. All of the Cronobacter spp.
strains tested (n=99) were able to grow in mLST at 45 ± 0.5 °C, whereas 35 of 39 strains of
potential competitors, all belonging to the Enterobacteriaceae, were suppressed. A survey
was carried out with 192 environmental samples from four different milk powder factories.
Using this new protocol, Crononacter spp. was isolated from almost 40% of the samples,
whereas the reference procedure (enrichment in buffered peptone water, isolation on violet
red bile glucose agar, and biochemical identification of randomly chosen colonies) only
yielded 26% positive results. This selective method can be very useful for the rapid and
reliable detection of Cronobacter spp. in environmental samples.






This Chapter has been published as
“A simple and r apid cultur al method for detection of Enterobacter sakazakii* in
envir onmental samples”

Guillaume-Gentil, O., V. Sonnard, M.C. Kandhai, J.D. Marugg, and H. Joosten.
Journal of Food Protection, 2005; 68:1, 64–69.

*Throughout this thesis the designation Cronobacter spp. has been used to indicate the
taxonomic change.
Enrichment procedure for Cronobacter spp.

57
Intr oduction
Cronobacter spp., a yellow member of the family Enterobacteriaceae, has been associated
with various cases of severe neonatal meningitis, bactereamia, and necrotizing enterocolitis,
particularly in premature babies (2–4, 8, 9, 15, 19, 20, 23, 25–27). In some of these cases,
contaminated powdered milk infant formula was identified as the source of Cronobacter
spp. (2, 4, 25, 26). In surveys of dry infant formulas obtained from various countries,
Cronobacter spp. was indeed isolated (21, 24), albeit at very low concentrations of less than
1 CFU/g. The dose-response relationship for this opportunistic pathogen is not known.
However, several cases have been associated with an exposure of prepared infant formula
to improper holding temperatures, which suggests that multiplication of Cronobacter spp.
is often a decisive factor in causing the outbreaks (4, 26). Nevertheless, it remains very
important to reduce the prevalence of the micro-organism in infant formula. Management
of factory environment hygiene is critical to the control of the organism, because the
presence of Cronobacter spp. in infant formula is often attributed to contamination
originating from dry factory environments, where it appears to survive better than other
members of the Enterobacteriaceae (5).
To evaluate the effect of different cleaning strategies on the prevalence of Cronobacter spp.
in the factory environment, sensitive and specific detection methods are required. The only
methods described thus far are based on general methods for the enrichment and isolation
of Enterobacteriaceae (e.g., pre-enrichment in buffered peptone water, sub-culturing in
Enterobacteriaceae enrichment broth, and isolation on violet red bile glucose agar),
followed by biochemical identification of typical colonies (1, 3, 21, 22). This approach is
laborious and requires up to 7 days obtaining a confirmed positive result. Recently, an
improved identification method was proposed (13), which is based on the detection of o-
glucosidase and production of a yellow pigment. With this method, confirmed positive
results can be obtained after 5 days. However, this procedure is not selective for
Cronobacter spp. and may thus be prone to false-negative results in the presence of other
Enterobacteriaceae.
Here, we describe the development of a selective enrichment procedure for Cronobacter
spp. based on lauryl sulphate tryptose, a commonly used enrichment broth for coliforms.
Selectivity was enhanced by the addition of 10 mg/L vancomycin and 0.5 M NaCl and by
the application of a relatively high incubation temperature (45 ± 0.5 °C). We also introduce
a new procedure that produces yellow colonies on the isolation agar within 24 h by
exposing the plates to visible light during incubation at 37 ºC.
Chapter 3
58
Mater ials and methods
Bacterial strains
Ninety-nine strains of Cronobacter spp. were used in this study (Table 1), 70 of which had
been isolated from the environment of milk powder factories in Europe and Asia using a
non-selective enrichment and isolation method. These strains were selected to represent the
major ribotypes of this species. The remaining strains were isolated from private
households (14), or were kindly provided by Dr. S. Forsythe (Nottingham Trent University,
Nottingham, UK) and by Dr. H. L. Muytjens (Radboud University, Nijmegen, The
Netherlands). The identification of the Cronobacter spp. strains was based upon the
combination of yellow pigmentation and a positive -glucosidase reaction, as described by
Kandhai et al., (13). The identity of Cronobacter spp. was further confirmed with API 20
E

strips (bioMe´rieux SA, Marcy l’Etoile, France) and by ribotyping with the RiboPrinter
(DuPont Qualicon, Inc., Wilmington, Del.) (6). The ribopattern of each presumptive
Cronobacter spp. strain was recorded, digitized, and stored in a database containing 11
DuPont reference Cronobacter spp. fingerprints combined with 2,000 in-house
Cronobacter spp. fingerprints (unpublished results).
Thirty-nine non–Cronobacter spp. strains were selected to represent the major species of
other Enterobacteriaceae found in factory environments. The strains either came from the
Nestlé Food Safety Microbiology culture collection or had been isolated from milk powder
or from the environment of milk powder factories. The selection focused on yellow non–
Cronobacter spp. strains (Citrobacter spp., Escherichia hermanii, Escherichia vulneris,
Pantoea agglomerans, Pantoea spp., and Serratia ficaria) that can easily be confused with
yellow Cronobacter spp.. All strains were different from Cronobacter spp. by API 20
E
and
automated ribotyping (data not shown).
All strains were maintained at -30 ºC in bacterial strain storage medium (Bio-Rad
Laboratories, Inc., Hercules, Calif.). Working cultures were prepared by inoculation of
brain heart infusion broth (BHI; Oxoid Ltd., Basingstoke, Hampshire, UK) and incubation
for 20 h at 30 ºC.
Enrichment procedure for Cronobacter spp.

59
Table 1. Cronobacter spp. and other Enterobacteriaceae strains used in this study
Species Str ain code Sour ce
a
Gr owth in
b


LST
30 ºC
LST
45 ºC
mLST
45 ºC
Yellow
pigment
on TSBA
Cronobacter spp. 48; 49; 145; 151; 155; 156; 158; 164; 166; 167;
262; 263; 265; 266; 270; 271; 272; 273; 274;
275; 276; 277; 278; 279; 280; 281; 282; 283;
284; 285; 286; 287; 288; 290; 291; 292; 293;
294; 295; 297; 298; 299; 300; 301; 302; 303;
304; 305; 306; 307; 308; 312; 314; 315; 316;
317; 318; 319; 320; 321; 323; 354; 356; 359;
366; 383; 384; 390; 391
E ++ ++ ++ +
Cronobacter spp. 365 E ++ ++ + +
Cronobacter spp. 309; 327; 368; 369 MP
d
++ ++ ++ +
Cronobacter spp. 310; 328; 393 MP ++ ++ ++ +
Cronobacter spp. 313; 322; 376
e
(=ATCC 29544); 468; 469;
470; 471; 472
C
d
++ ++ ++ +
Cronobacter spp. 473 C
d
++ ++ + +
Cronobacter spp. 375; 377; 378 H ++ ++ ++ +
Cronobacter spp. 454; 455; 457; 458; 459; 460; 461; 462; 463 Not known
f
++ ++ + +
Cronobacter spp. 456 Not known
f
++ ++ ++ +
Citrobacter spp. 31 NFSM
g
++ + - -
Citrobacter spp. 189 NFSM ++ ++ - +
Citrobacter spp. 411 MP ++ + - +
Enterobacter cloacae 20; 62 NFSM ++ ++ - -
Enterobacter cloacae 84 NFSM ++ + - -
Enterobacter cloacae 86 MP ++ ++ - -
Enterobacter cloacae 418 MP ++ ++ ++ -
E aerogenes 149 MP ++ ++ - -
Escherchia coli 186 NFSM ++ + - -
Escherchia hermanii 40 NFSM ++ ++ ++ -
Escherchia hermanii 73 NFSM ++ + - -
Escherchia hermanii 75 NFSM ++ ++ - +
Escherchia vulneris 417 MP ++ + - -
Escherchia vulneris 422; 428; 429; 435; 442 E ++ ++ - +
Hafnia.spp. 42; 108 NFSM ++ ++ - -
Hafnia spp. 64 NFSM ++ + - -
Klebsiella oxytoca 104 NFSM ++ ++ ++ -
Klebsiella pneumoniae 254 NFSM ++ ++ - -
Klebsiella pneumoniae 255 NFSM ++ ++ ++ -
Leclercia 80 NFSM - - - -
Pantoea agglomerans 143 NFSM ++ + - -
Pantoea agglomerans 188 NFSM - - - +
Pantoea agglomerans 246 NFSM ++ ++ - -
Pantoea spp. 168 MP ++ ++ - +
Pantoea spp. 415 MP + + - +
Pantoea spp. 416 MP ++ ++ - +
Pantoea spp. 440; 441; 443; 450 E ++ ++ - +
Providencia spp. 65 NFSM ++ ++ - -
Serratia ficaria 142 NFSM ++ ++ - +
Serratia ficaria 427 E ++ ++ - +
a
E: Isolates from the environment of milk powder factories; MP: isolates from milk powder; C:
clinical isolates; H: households.
b
Growth was assessed by measuring the optical density at 620 nm
after incubation. “-“: OD620 < 0.010; “+” : OD620 = 0.010 – 0.100; “++” OD620 < 0.010.
c
LST (Lauryl Sulphate Tryptose broth) supplemented with 0.5 M NaCL and 10 mg/L vancomycin
(mLST).
d
Radbout University, Nijmegen (The Netherlands).
e
Type strain f Nottingham Trent
University (UK).
g
NFSM: Nestlé Food Safety Microbiology Culture Collection.
Chapter 3
60
Development of a selective enrichment broth for Cr onobacter spp.
The conditions for selective enrichment culture of Cronobacter spp. were established in
microtiter plate growth assays using 15 Cronobacter spp. strains (Table 2). Aliquots (5 µl)
of freshly grown BHI cultures were inoculated into wells containing 200 µl lauryl sulfate
tryptose broth (LST; Oxoid) supplemented with different concentrations (0, 0.25, 0.5, 0.75,
and 1.0 M) of NaCl (Merck, Whitehouse Station, N.J.). The microtiter plates were
incubated at 44, 45, 46, and 47 °C. Growth was assessed by determining the optical density
at 620 nm after 0, 6, 24, and 48 h using a microtiter plate reader (Multiskan/MCC 340,
Labsystems, Helsinki, Finland). Based on the results of these experiments, LST broth
supplemented with 0.5 M NaCl and 10 mg/L vancomycin (Fluka Chemie, Buchs,
Switzerland) (mLST) and incubated at 45 ± 0.5 °C was selected for further evaluation with
99 Cronobacter spp. strains and 39 strains of other Enterobacteriaceae. For this part of the
study, 0.1 ml aliquots of freshly grown BHI cultures diluted 1:100 in physiological saline
were inoculated into 10 ml volumes of standard LST and mLST to give an initial count of
ca. 10
4
CFU/ml and then incubated at 30 or 45 °C for 24 h. The optical density was then
measured at 620 nm.

Table 2. Growth of 15 Cronobacter strains in LST supplemented with various NaCl-concentrations
a
Growth was assessed by measuring the optical density at 620 nm after 24 h incubation.
“-”: OD
620
< 0.010; “+” : OD
620
0.010 - 0.100; “++”:OD
620
> 0.010.
b
Nestlé Food Safety Microbiology (NFSM) 262, 263, 266, 270, 280, 302, 310, 328, 359, 362, 369,
383, 384, 391, 393.
Temper atur e
37 ºC 44 ºC 45 ºC 46 ºC 47 ºC
Gr owth
a
of 15 Cronobacter spp.
str ains
b


Standard LST ++ ++ ++ ++ +
LST supplemented with 0.25 M NaCL ++ ++ ++ ++ +
LST supplemented with 0.50 M NaCL ++ ++ ++ + -
LST supplemented with 0.75 M NaCL ++ ++ + - -
LST supplemented with 1.00 M NaCL + - - - -
Enrichment procedure for Cronobacter spp.

61
Samples
Dry environmental samples were obtained from different areas of four milk powder
factories and included floor sweepings, scrapings of processing areas, spilled dry products,
contents of vacuum cleaner bags, etc. From the same factories, moist environmental sponge
samples (8 by 4 by 1 cm) were obtained.
Non-selective enrichment and isolation
For non-selective enrichment (BPW-VRBL) samples (10 g each) were suspended in 90 ml
of buffered peptone water (BPW; Oxoid). When quantities smaller than 10 g were
analyzed, the volume of BPW was adapted proportionally (1:10). Sponges were soaked
with ca. 100 ml of BPW. The enrichment broth was incubated for 16 to 20 h at 37 °C. A
loopful of pre-enrichment broth (ca. 10 µl) was streaked onto violet red bile lactose agar
(VRBL; Oxoid) and incubated at 37 °C for 24 h. From each plate, five colonies
demonstrating typical coliform characteristics were transferred to tryptone soy agar (TSA;
Oxoid) and examined after 24 h of incubation at 37 °C for yellow pigment production,
oxidase, and α-glucosidase activity as described by Kandhai et al., (13) and identified with
API 20
E
strips.
Two-stage selective enrichment
For the two-stage selective enrichment (BPW-mLST-TSBA), samples were first pre-
enriched in BPW as described above. A loopful of pre-enrichment broth (ca.10 µl) was
transferred to 10 ml of mLST and incubated for 22 to 24 h at 45 ± 0.5 °C followed by
streaking of a loopful onto a 14-cm petri dish containing TSA supplemented with 1.5 g/L
bile salts 3 (TSBA; Oxoid). During incubation (24 h at 37 ºC), plates were exposed to
artificial white light (ca. 600 lx) to induce production of a yellow pigment by Cronobacter
spp.. From each plate, five yellow colonies were selected for confirmation as described
above.

Chapter 3
62
Figur e 1. Schematic diagram for the enrichment and identification of Cronobacter spp. in
environmental samples with mLST-TSBA, BPW-mLST-TSBA and BPW-VRBL.

Single-stage selective enrichment
For the single-stage selective enrichment (mLST-TSBA), 10 g of sample was suspended in
90 ml of mLST and incubated for 22 to 24 h at 45 ± 0.5 °C. When quantities smaller than
Enrichment procedure for Cronobacter spp.

63
10 g were analyzed, the volume of mLST was adapted proportionally (1:10). Sponges were
soaked with ca. 100 ml of mLST. After incubation, suspensions were transferred to TSBA
for evaluation.

Results and discussion
To enhance the selectivity of Cronobacter spp., LST broth, widely used as semi-selective
enrichment broth for coliforms (11, 12), was modified by adding extra NaCl and the
antibiotic vancomycin. The NaCl was added because Breeuwer et al., (5) demonstrated that
Cronobacter spp. has a higher tolerance for osmotic stress than do other members of the
Enterobacteriaceae such as Salmonella spp., Escherichia coli, Klebsiella pneumoniae and
Citrobacter freundii. The vancomycin was added as an additional safeguard against the
growth of competing gram-positive micro-organisms, in particular Bacillus spp. and lactic
acid bacteria (10, 17). Preliminary tests with LST + 0.5 M NaCl had indicated that these
gram-positive bacteria were often found in environmental samples and were apparently not
sufficiently inhibited by the presence of lauryl sulfate (results not shown).
A relatively high incubation temperature was chosen on the basis of results of different
studies (5, 7, 16, 24) indicating that Cronobacter spp. is able to grow at temperatures up to
46 or 47 °C. All 15 Cronobacter spp. strains evaluated grew well in standard LST up to 46
°C and were able to grow in LST at 37 °C at increasing osmolarities up to 1 M NaCl
(Table 2). The combination of LST with 0.5 M NaCl and an incubation temperature of 45 ±
0.5 °C was selected for further evaluation.
The extended growth study with pure cultures (Table 1) revealed that many of the selected
strains of Enterobacteriaceae were good competitors in LST at 45 ± 0.5 °C and that the
addition of NaCl was essential for selective enrichment of Cronobacter spp. all of the tested
Cronobacter spp. strains were able to grow, whereas 35 out of the 39 non–Cronobacter
spp. Enterobacteriaceae strains were inhibited. The remaining competitors belonged to a
variety of species: Escherichia hermanii, Enterobacter cloacae, Klebsiella oxytoca, and
Klebsiella pneumoniae. However, none of these strains produces a yellow pigment and
could be easily distinguished from Cronobacter spp. on the isolation medium TSBA.
Chapter 3
64
The performance of mLST broth was assessed in a comparative study involving 192
environmental samples collected from four milk powder factories. Both the one and two-
stage enrichment procedures (mLST-TSBA and BPW-mLST-TSBA, respectively) were
used, and the non-selective enrichment method (BPW-VRBL) (14) was used as a reference.
Similar numbers of Cronobacter spp. positive samples were obtained with BPW-mLST-
TSBA and mLST-TSBA (73 and 75, respectively), whereas the reference method yielded
only 51 positive results (Table 3). The difference in efficiency between the mLST-based
procedures and the reference procedure is probably due to both the selectivity of mLST and
the utilization of TSBA as isolation agar rather than randomly picking coliform colonies
from VRBL plates.
Nevertheless, not all the yellow colonies growing on TSBA could be confirmed as
Cronobacter spp.. Some strains were identified as Pantoea spp., Pantoea agglomerans, or
Escherichia vulneris. Thus, additional tests such α-glucosidase activity and identification
with API 20
E
strips are necessary to eliminate false-positive results.

Table 3. Detection of Cronobacter spp. in environmental samples from four milk powder factories,
using three different protocols
No. (%) samples
Confir med positive with:
Source Tested Confirmed
positive
mLST-
TSBA
BPW-mLST-
TSBA
BPW-VRBL
Factory A 48 20 18 18 9
Factory B 46 10 9 10 5
Factory C 44 24 23 19 19
Factory D 54 34 25 26 18
Total 192 (100) 88 (46) 75 (39) 73 (38) 51 (26.5)

The total number of positive samples found in this comparative study was still considerably
higher (88) than that obtained with either of the two mLST-based methods (Table 3). This
difference is likely due to an uneven distribution of Cronobacter spp. in the samples, which
had been divided to allow comparison of the different procedures as outlined in Figure 1.
Enrichment procedure for Cronobacter spp.

65
However, both methods may be prone to false-negative results, and further improvements
of the sensitivity may be necessary.
Although the direct mLST-TSBA method is 1 day faster than the BPW-mLST-TSBA route,
the latter method may be preferred for routine purposes because it can be combined with
Salmonella testing; the first stage of this method also consists of a non-selective pre-
enrichment. The direct mLST-TSBA method has been used in the development of a
Cronobacter spp. specific PCR assay on the BAX system (Dupont Qualicon) that provides
results within 30 to 32 h (18).
In this study, we also used an accelerated procedure to obtain yellow colonies on the
isolation agar medium (TSBA) by exposing the plates to artificial white light during their
incubation at 37 °C. Prior to implementing this method, we tested more than 100 different
Cronobacter spp. strains from our collection, and in all cases the presence of yellow
colonies could already be observed within 24 h (data not shown). This modification
represents a clear advantage to the methods used previously, which are based on incubation
at 25 °C for 48 to 72 h (3, 7, 21).
The procedures outlined in this study will enable efficient and rapid isolation of
Cronobacter spp. from environmental samples. Effective environmental analysis can
contribute to a better understanding of the route(s) by which Cronobacter spp. enters
factories producing powdered infant formulas and will help managers identify in-factory
ecological niches where this organism can persist. Such information is essential for
developing control and prevention strategies for this emerging opportunistic pathogen.

Acknowledgments
The authors thank Mrs. E. Bidlas for determination of ribopatterns and Dr. T. Jackson for
his critical review of the manuscript.

Chapter 3
66
Refer ences
1. Aldova, E., O. Hausner, and R. Postuva. 1983. Tween-esterase activity in
Enterobacter sakazakii. Zbl. Bakt. Hyg. A256: 103-108.
2. Anonymous. 2002. Center for Disease Control and Prevention. Enterobacter
sakazakii infections associated with the use of powdered infant formula -
Tennessee, 2001. MMWR Morb. Mortal. Wkly Rep. 51 (14):297-300.
3. Anonymous. 10 March 2003. United States Food and Drug Administration. 2002.
Isolation and enumeration of Enterobacter sakazakii from dehydrated powdered
infant formula. Available at: http://www.cfsan.fda.gov/~comm/mmesakaz.html.
4. Biering, G., S. Karlsson, N. C. Clark, K. E. Jonsdottir, P. Ludvigsson, and O.
Steingrimsson. 1989. Three cases of neonatal meningitis caused by Enterobacter
sakazakii in powdered milk. J. Clin. Microbiol. 27:2054-2056.
5. Breeuwer, P., A. Lardeau, M. Peterz, and H. Joosten. 2003. Dessication and heat
tolerance of Enterobacter sakazakii. J. Appl. Microbiol. 95:967-973.
6. Bruce, J. 1996. Automated system rapidly identifies and characterizes micro-
organisms in food. Food Technol. 50:77-81.
7. Farmer, J. J., M. A. Asbury, F. W. Hickman, D. J. Brenner and the
Enterobacteriaceae study group. 1980. Enterobacter sakazakii: a new species of
"Enterobacteriaceae" isolated from clinical specimens. Int. J. Syst. Bacteriol.
30:569-584.
8. Food and Agriculture Organization. 2003. Codex Alimentarius: risk profile of
Enterobacter sakazakii in powdered infant formula. CX/FH 03/13. Food and
Agricultural Organization of the United Nations, Rome, Italy.
9. Gallagher, P. G., and W. S. Ball. 1991. Cerebral infarctions due to CNS infection
with Enterobacter sakazakii. Pediatr. Radiol. 21:135-136.
10. Giraffa, G., and F. Sisto. 1997. Susceptibility to vancomycin of Enterococci
isolated from dairy products. L. Appl. Microbiol. 25:335-338.
11. International Dairy Federation 1998. IDF 73B:1998: Enumeration of coliforms -
colony count technique and most probable number technique at 30 °C.
International Dairy Federation, Geneva, Switzerland.
12. International Organization for Standardization. 1991. ISO 4831:1991:
Microbiology - General guidance for the enumeration of coliforms - most probable
number technique. International Organization for Standardization, Geneva,
Switzerland.
13. Kandhai, M. C., M. W. Reij, K. Van Puyvelde, O. Guillaume-Gentil, R. R.
Beumer, and M. Van Schothorst. 2004a. A new protocol for the detection of
Enterobacter sakazakii applied to environmental samples. J. Food Prot. 67: 1267-
1270.
Enrichment procedure for Cronobacter spp.

67
14. Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004b. Occurrence of Enterobacter sakazakii in food production
environments and households. The Lancet 363:39-40.
15. Lai, K. K. 2001. Enterobacter sakazakii infections among neonates, infants,
children, and adults. Medicine 80:113-122.
16. Leclercq, A., C. Wanegue, and P. Baylac. 2002. Comparison of fecal coliform
agar and violet red bile lactose agar for fecal coliform enumeration in foods. Appl.
Environ. Microbiol. 68: 1631-1638.
17. Manie, T., S. Khan, V. S. Brözel, W. J. Veith, and P. A. Gouws. 1998.
Antimicrobial resistance of bacteria isolated from slaughtered and retail chickens
in South Africa. L. Appl. Microbiol. 26: 253-258.
18. Marugg, J., T. Jackson, O. Guillaume-Gentil, H. Joosten, M. Barbour, and P.
Mrozinski. Development of a PCR based method for the detection of E. sakazakii
in milk powders and environmental samples (manuscript in preparation).
19. Monroe, P. W., and W.L. Tift. 1979. Bacteremia associated with Enterobacter
sakazakii (yellow-pigmented Enterobacter cloacae). J. Clin. Microbiol. 10:850-
851.
20. Muytjens, H. L., and L. A. Kollee. 1990. Enterobacter sakazakii meningitis in
neonates: causative role of formula? Pediatr. Infect. Dis. J. 9:372-373.
21. Muytjens, H. L., H. Roelofs-Willemse, and G. H. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26: 743-746.
22. Muytjens, H.L., J. Van der Ros-Van de Ripe, and H. A. M. Van Druten. 1984.
Enzymatic profiles of Enterobacter sakazakii and related species with special
reference to the alpha-glucosidase reaction and reproducibility of the test system.
J. Clin. Microbiol. 20:684-686.
23. Muytjens, H. L., H. C. Zanen, H. J. Sonderkamp, L. A. Kollee, I. K. Wachsmuth,
and J.J. Farmer 3
rd
. 1983. Analysis of eight cases of neonatal meningitis and sepsis
due to Enterobacter sakazakii. J. Clin Microbiol. 18: 115-120.
24. Nazarowec-White, M., and J. M. Farber. 1997. Incidence, survival, and growth of
Enterobacter sakazakii in infant formula. J. Food Prot. 60:226-230.
25. Simmons, B. P., M. S. Gelfand, M. Haas, L. Metts, and J. Ferguson. 1989.
Enterobacter sakazakii infections in neonates associated with intrinsic
contamination of a powdered infant formula. Infect. Control Hosp. Epidemiol.
10:398-401.
26. Van Acker, J., F. de Smet, G. Muyldermans, A. Bougatef, A. Naessens, and S.
Lauwers. 2001. Outbreak of necrotizing enterocolitis associated with Enterobacter
sakazakii in powdered milk formula. J. Clin. Microbiol. 39:293-297.
27. Willis, J., and J. E. Robinson. 1988. Enterobacter sakazakii meningitis in
neonates. Pediatr. Infect. Dis. J. 7:196-199.
Chapter 3
68
Occurrence of Cronobacter spp.

69
Chapter 4
70
Abstr act
Cronobacter spp. occasionally causes illness in premature babies and neonates.
Contamination of infant formulae during factory production or bottle preparation is
implicated. Advice to health-care professionals focuses on bottle preparation, but the
effectiveness of prevention depends on the degree of contamination and contamination
sites, which are generally unknown. To keep contamination to a minimum in the finished
product depends on knowledge of the occurrence of Cronobacter spp.. We used a refined
isolation and detection method to investigate the presence of this micro-organism in
various food factories and households. Environmental samples from eight of nine food
factories and from five of 16 households contained Cronobacter spp.. The widespread
nature of this micro-organism needs to be taken into account when designing preventive
control measures.









This Chapter has been published as
“Occur r ence of Enterobacter sakazakii* in food pr oduction envir onments and
households”

Kandhai, M.C., M.W. Reij, L.G.M. Gorris, O. Guillaume-Gentil, and M. Van Schothorst.
The Lancet, 2004; 363:1, 39–40.

* Throughout this thesis the designation Cronobacter spp. has been used to indicate the
taxonomic change.
Occurrence of Cronobacter spp.

71
Intr oduction
Several species within the genus Enterobacter have been recognised as important causative
agents of hospital acquired infections. Cronobacter spp. in particular has caused several
outbreaks or sporadic cases of severe neonatal meningitis in premature babies, or
necrotising enterocolitis. Cronobacter spp. has been isolated from various clinical
materials, such as cerebrospinal fluid, blood, skin, wounds, respiratory tract (sputum,
throat, and nose), digestive tract, and urine (1). Cronobacter spp. was isolated from
powdered substitutes for breast milk (4). Infant formulae are pasteurised during
manufacturing, and Cronobacter spp. does not survive such heat treatment (5). The micro-
organism’s presence in the finished product, therefore, probably originates from the factory
environment, from heat-sensitive micronutrients added after pasteurisation or from bottle
preparation.
In several investigations into outbreaks or cases of Cronobacter spp. infection in premature
babies and neonates, the micro-organism was isolated from blenders, bottle cleaning
brushes, etc. (3). Muytjens and Kollee investigated the occurrence of Cronobacter spp.
more widely, but could not isolate the micro-organism from any environment they
examined, which included surface water, soil, mud, rotting wood, grain, bird dung, rodents,
domestic environments, cattle, and untreated cow’s milk (3). Cronobacter spp. has,
however, been isolated from cheese, minced beef, sausage meat, and vegetables (2), but we
are unaware of any investigation into the natural habitat of Cronobacter spp.. Why
Cronobacter spp. infection occurs only in neonates and premature babies, and why it is
associated only with the consumption of infant formulae, is unclear.
The US Food and Drug Administration issued an alert in April, 2002, to health-care
professionals about the risks associated with Cronobacter spp. infections among neonates
fed with milk-based infant formulae, and recommended certain bottle-preparation practices.
The alert stated that a major contribution to the prevention of Cronobacter spp. infections
in premature babies and neonates is the prevention of contamination of infant formulae
during production and bottle preparation. However, knowledge of the aetiological and
ecological characteristics of Cronobacter spp. is sparse, and its occurrence in factories that
Chapter 4
72
produce infant formulae, hospital kitchens, and households has hitherto not been studied in
depth. We have started a systematic investigation of the habitat of Cronobacter spp..

Mater ials and methods
Environmental samples were obtained from nine factories and 16 households (Table1). In
factories, samples were obtained by scraping or sweeping surfaces in the production-line
environment or by sampling vacuum cleaner bags. Samples from households were taken
from vacuum cleaner bags. From each sample, we analysed 10 g for the presence of
Cronobacter spp. by selective enrichment in lauryl sulphate tryptose broth, supplemented
with 0.5 mol/L salt, incubated at 45 ºC for 22 to 24 h. A loopful was streaked on violet red
bile lactose agar to select for coliforms during the subsequent incubation at 37 ºC for 24 h.
All coliform colonies on violet red bile lactose agar were streaked on to tryptone soy agar
plates and incubated for 48 h in daylight at room temperature. For the identification of
Cronobacter spp., we tested all yellow-coloured oxidase-negative colonies for the presence
of o-glucosidase. Presumptive Cronobacter spp. isolates were identified with the API 20
E

test system (bioMerieux, Marcy L’Etoile, France) and were also characterised as
Cronobacter spp. with the ribotyping technique.

Results and discussion
Cronobacter spp. was isolated with varying frequency from nearly all environments
examined. Although the proportion of positive samples differed between environments, the
differences were not significantly different (Table 1). The presence of Cronobacter spp. in
factories producing milk powder, cereals, chocolate, potato flour, and pasta, as well as in
domestic environments, strongly indicates that it is a widespread micro-organism. Since
measures to prevent neonatal cases of Cronobacter spp. infections are being discussed in
various countries by industries, public-health workers, and consumers, our findings may
contribute to the current knowledge. Most importantly, the widespread nature of
Cronobacter spp. should be taken into account in the design of effective control measures.
Occurrence of Cronobacter spp.

73
Table 1. Environmental samples of various origins tested for the presence of Cronobacter spp.
Or igin Number of
samples
analysed
Number of samples
positive for
Cronobacter spp. (%)
95% confidential
inter val
Milk powder factory 23 2 (9) 0.01 - 0.27
Milk powder factory 26 9 (35) 0.19 - 0.53
Milk powder factory 11 1 (9) 0.002 - 0.404
Milk powder factory 8 2 (25) 0.03 - 0.64
Chocolate factory 8 2 (25) 0.03 - 0.64
Cereal factory 9 4 (44) 0.15 - 0.78
Potato flour factory 15 4 (27) 0.08 - 0.53
Pasta factory 26 6 (23) 0.10 - 0.42
Spice factory 5 0 0 - 0.52
Households 16 5 (31) 0.12 – 0.57

Acknowledgments
This research is part of PhD research project of Chantal Kandhai, which is supported by the
Nestlé Research Centre, Lausanne, Switzerland. The sponsor had no role in study design,
data collection, data analysis, and data interpretation, or in writing the report.

Refer ences
1. Lai, K. K. 2001. Enterobacter sakazakii infections among neonates, infants,
children, and adults - Case reports and a review of the literature. Medicine.
80:113-122.
2. Leclercq, A., C. Wanegue, and P. Baylac. 2002. Comparison of fecal coliform
agar and violet red bile lactose agar for fecal coliform enumeration in foods. Appl.
Environ. Microbiol. 68:1631-1638.
3. Muytjens, H. L., and L. A. Kollee. 1990. Enterobacter sakazakii meningitis in
neonates: causative role of formula? Pediatr. Infect. Dis. J. 9:372-373.
4. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
5. Nazarowec-White, M., and J. M. Farber. 1997. Thermal resistance of Enterobacter
sakazakii in reconstituted dried infant formula. Lett. Appl. Microbiol. 24:9-13.
Chapter 4
74
Appendix: Cor r espondence

Cronobacter spp. in factor ies and households
The isolation of Cronobacter spp. from various factories, including infant formula factories
(1), and the related contamination of infant formulae with Cronobacter spp. are causes for
concern. I was surprised to see only one recommendation from the authors at the end of the
article: “The widespread nature of this micro-organism needs to be taken into account when
designing preventive control measures.” Two other recommendations quickly come to
mind: (1) enhancement of the promotion of and support for breastfeeding, and (2) inclusion
of a warning on infant formulae and other breast-milk substitutes that the product might be
contaminated by Cronobacter spp. and other micro-organisms. I was wondering why the
authors did not make such recommendations, when my eyes fell on the last paragraph of the
article: “This research is part of a PhD research project . . . which is supported by the Nestlé
Research Centre.” A few lines higher I read that no conflicts of interest were declared.
Nestlé produces infant formula and other breast-milk substitutes. If this is not a conflict of
interest, then what is?



Maaike Arts
UNICEF, 72 Ly Thuong Kiet Street, Hanoi,
Vietnam.

Refer ences
1. Kandhai M. C., M. W. Reij, L. G. M Gorris, O. Guillaume-Gentil, M. Van
Schothorst. 2004. Occurrence of Enterobacter sakazakii in food production
environments and households. The Lancet. 363: 39–40.
Occurrence of Cronobacter spp.

75
Author s’ r eply
Maaike Arts asks why we did not include in our Research letter other recommendations
about Cronobacter spp. in infant formulae. We believe that our conclusion “The
widespread nature of this micro-organism needs to be taken into account when designing
preventive control measures” is the only one justified on the basis of the results we reported
in the paper. The purpose of our study was to gain a better understanding of the ecology of
Cronobacter spp. as a starting point for designing effective control strategies. On the basis
of the research we reported, we cannot substantiate the recommendations suggested by
Arts. However, we would like to draw readers’ attentions to the joint FAO/WHO Expert
Workshop on E. sakazakii and other micro-organisms in powdered infant formula (held in
Geneva, Switzerland, Feb 2–5, 2004)

(1) as a potential source of information.
With respect to the suggested conflict of interest, we are convinced that we clearly
disclosed our affiliations and financial support as required by the journal. As stated in the
original publication, “The sponsor had no role in the study design, data collection, data
analysis, and data interpretation, or in writing the report.”



Kandhai, M. Chantal, Martine W. Reij, Leon G. M. Gorris, Olivier Guillaume-Gentil, and
Mike van Schothorst
Wageningen University, Laboratory of Food Microbiology, PO Box 8129, 6700 EV
Wageningen, Netherlands (MCK, MWR, LGMG, MvS); and Nestlé Research Centre,
Quality and Safety Assurance Department, Lausanne,Switzerland (OGG)

Refer ences
1. FAO/WHO. 2004. Enterobacter sakazakii and other micro-organisms in powdered
infant formula: meeting report. Microbiological Risk Assessment Series No. 6,
Rome, Italy. Available at: http://who.int/foodsafety/micro/ meetings/feb2004.

Chapter 4
76

Prevalence of Cronobacter spp.

77
Chapter 5
78
Abstr act
Cronobacter spp. is an opportunistic pathogen, possibly occurring in many different foods
and environments causing illnesses predominantly through powdered infant formula. To
add to the existing knowledge on the occurrence of Cronobacter spp. and to help identify
habitats that could be potential sources of this organism, a broad survey of foods
manufactured or marketed in The Netherlands, including relevant non-food environments,
was conducted over a 5-years period (2001 – 2005). Using a specifically designed real-
time polymerase chain reaction method for definite confirmation, Cronobacter spp. was
isolated from milk powders (7/175), powdered formulae for consumers < 1 year (8 /395),
follow-up formula for consumers > 1 year (1/5), other powdered instant products (1/182),
dry cereals (6/123), fresh raw minced meats (7/222), vegetables (2/47), spices (1/28),
samples of fresh human faeces (1/98), and human skin (1/116) samples. None of the
samples taken from fresh raw bovine milk (0/51), fresh human milk (0/7), baby bottles
(0/86) and bottle-nipples (0/95) were found to be positive for Cronobacter spp.. This study
confirms that Cronobacter spp. can occur in a wide range of dry and wet foods, not limited
to infant formulae or milk products, as well as on and in humans. This widespread
occurrence is important to consider when manufacturing foods for consumers at potential
risk for Cronobacter spp. infection.




Part of this Chapter has been accepted for publication as
“A study into the occur ence of Cronobacter spp. in The Nether lands between 2001 and
2005”

Kandhai, M.C., A.E. Heuvelink, M.W. Reij, R.R. Beumer, R. Dijk, J.J.H.C. van Tilburg,
M. van Schothorst, and L.G.M. Gorris.
Food Control, 2010, doi:10.1016/j.foodcont.2010.01.007.
Prevalence of Cronobacter spp.

79
Intr oduction
Cronobacter spp. (Enterobacter sakazakii) are Gram-negative, motile, non-spore forming,
facultative anaerobic bacteria. The genus was originally defined as a Enterobacter
sakazakii species in 1980 (21), but very recently it has been reclassified (39, 40) as six
species in a new genus, Cronobacter gen. nov., within the family of Enterobacteriaceae.
Five of the new species are Cronobacter sakazakii, Cronobacter turicensis, Cronobacter
malonaticus, Cronobacter muytjensii and Cronobacter dublinensis. The sixth species is
preliminarily indicated as genomospecies I. To indicate the taxonomic change, the
designation “Cronobacter spp.” is used consistently in the subsequent part of this paper,
where the micro-organism is referred to in general terms or where historic information that
is not specific to the newly recognized particular species is cited.
Although other Enterobacter species have been involved in nosocomial diseases, the
importance of Cronobacter spp. is particularly due to the fact that it has been recognised as
an opportunistic pathogen for infants, especially neonates (1, 11, 29). The micro-organism
has been linked with serious infections in infants, causing bacteraemia and meningitis, and
has also been isolated from infants in association with necrotizing enterocolitis (18, 19, 55).
The International Commission on Microbiological Specifications for Foods (ICMSF)
placed Cronobacter spp. in the category of severe hazards for restricted populations, being
rare in the frequency of involvement in food-borne disease but with a lethality rate up to
70% among neonates (30). The Scientific Panel on Biological Hazards of the European
Food Safety Authority (EFSA) noted that although Cronobacter spp. has caused illness in
infants up to 1 year, neonates up to ca 4 to 6 weeks of age, pre-term or low birth-weight
infants and those immuno-compromised are at greatest risk (16). Outbreaks of infant
meningitis, necrotizing enterocolitis and or bacteraemia have been reported (2, 50, 61, 63,
72). All Cronobacter species have been linked retrospectively to clinical cases of infection
in either infants or adults, and therefore all species should be considered pathogenic (20).
Cases of Cronobacter spp. disease have been detected most commonly among newborn and
very young infants in hospital nurseries and neonatal intensive care units. The first cases
attributed to this organism occurred in 1958 in England (73). Since then and up to July
2008, around 156 documented cases of Cronobacter spp. infection and at least 29 deaths
Chapter 5
80
from all parts of the world appeared in the published literature and in reports submitted by
public health organizations and laboratories (20). Considering the recent review of
outbreaks caused by Cronobacter, an average lethality rate of 19% was calculated. More
recently, the overall lethality of 67 invasive cases was estimated to be 26.7%. The lethality
of Cronobacter meningitis, bacteraemia and necrotizing enterocolitis (NEC) was estimated
to be 41.9%, < 10% and 19%, respectively (23).
Specific international attention has been given to the safety of food for infants and young
children as it relates to the possible presence of Cronobacter spp. in powdered formulae
intended for this consumer group. The Food and Agriculture Organization of the United
Nations (FAO) and the World Health Organization (WHO) have established two
international risk assessments on the topic (18, 19) that have provided a scientific basis for
the Codex Alimentarius Committee on Food Hygiene (CCFH), for drafting an new Code of
Hygienic Practice (6) aimed at providing risk management guidance at the international
level. In the Code of Hygienic Practice products for infants up to 12 months of age are
considered, although several expert meetings have established that those younger than 2
months and infants of HIV positive mothers are at greatest risk for infection caused by
Cronobacter spp. (18-20).
The occurrence of Cronobacter spp. at low concentrations in powdered infant formulae has
been frequently reported (8, 28, 36, 59, 62). Notably, Cronobacter spp. has been isolated
from a variety of environmental samples, including milk powder factories and other
factories (43), as well as blenders and cleaning brushes used to prepare infant feeds in
hospital kitchens (4, 61). The micro-organism has been isolated from several food products
(i.e., beef, sausage meat, vegetables, cheese, powdered fruit flakes) and other
environmental sources, including water (22), dust from households (43), insects (27, 49, 53)
and rats (24). However, only powdered infant formulae have been linked to outbreaks of
infections as an unequivocal source (1, 9, 23).
In order to design appropriate and effective control measures for powdered infant formula,
the various possible sources of Cronobacter spp. need to be well understood. This would
include identification of its natural habitats and/or specific niches, as well as its occurrence
Prevalence of Cronobacter spp.

81
in factory environments and as a contaminant of food ingredients and finished food (or
feed) products. To increase our knowledge on sources and occurrence of the micro-
organism, several investigations were conducted in The Netherlands over a five years
period, namely between 2001 and 2005. These investigations considered food products
marketed in the country in retail and wholesale, imported food and feed products, as well as
food products, manufactured in The Netherlands. For part of the investigation period, in
addition to analysis for Cronobacter spp., the level of Enterobacteriaceae was also
assessed as an indicator of product hygiene, possibly related to the occurrence of the
pathogen. A number of different methods were used in these investigations, reflecting the
rapid evolution of detection and identification methodologies during the time of the
investigation. Though methods differed, in all cases, a specifically designed real-time
Polymerase Chain Reaction (PCR) protocol was used for definite confirmation of
presumptive Cronobacter spp. isolates.

Mater ial and methods
Samples
The occurrence of Cronobacter spp. was determined in milk powders (n=175), powdered
formulae (PF) for young children below 1 year of age (n=395), follow-up formulae (FUF)
for consumers > 1 year of age (n=15), other powdered instant foods (n=182), dry cereals
(n=123), fresh raw minced meats (n=222), fresh raw cow’s milk (n=51), dried vegetables
(n=47), dry spices (n=28), fresh human breast milk (n=7), fresh human faeces (n=98),
human skin (n=116), baby bottles (n=86), and bottle-nipples (n=95).
Of the 175 milk powders tested, 42 were produced in The Netherlands while the remainder
(133) were taken from cargos and bulk goods as part of systematic examinations of food
imports into The Netherlands. The milk powders originated from Middle- and Eastern-
European countries: Ukraine (43), Poland (53), Estonia (18), the Czech Republic (8),
Lithuania (5), Latvia (4) and Slovakia (2). All milk powders were intended for human
consumption, with the exception of 27 samples imported from the Ukraine that were
intended for animal feed.
Chapter 5
82
A total of 295 powdered formulae obtained from retailers across the country were sampled,
including 283 powdered formulae intended for use for infants up to the age of 12 months,
and 12 follow-up formulae, intended for young children older than 12 months. An
additional 115 powdered formulae were obtained directly from five manufacturers in The
Netherlands, including 112 powdered formulae for consumers below 1 year of age and 3
follow-up formulae.
Several other food products were obtained from retail, comprising of dry powdered foods,
including ice cream mixtures (n=62), instant pudding powders (n=25), flavoured instant
drinks (n=33), dips (n=19), coffee creamers (n=4), and dry cereals (n=123), as well as raw
meats, including beef (n=141), mixed beef and pork (n=50), pork (n=16), calf (n=10), and
poultry (n=5), and vegetables (n=47) and spices (n=28). Additionally, 27 samples of ice
cream mixtures were obtained from wholesale, whereas 12 instant pudding powders had
been collected directly in factories in The Netherlands.
Fresh cow’s milk (n=51) was collected from bulk tanks from 26 farms at two time points,
two months apart. Human breast milk samples (n=7) were obtained from nurturing mothers.
Samples of human faeces were collected from volunteers, enjoying a general healthy status
and belonging to different age categories: 41 samples from children younger than 5 years of
age, 9 from children between 9 and 14 years, and 48 from persons older than 14 years.
Samples of human skin particles were obtained by using scrub gloves (Parfumerie Douglas
Nederland B.V., Nijmegen, The Netherlands). Volunteers (n=116), belonging to the age
categoried 0-20 years (n=8), 20-40 years (n=96) and 40-60 years (n=12), were asked to
scrub off as much skin area as possible while taking a shower. Most volunteers scrubbed
only their arms and legs. Used baby bottles and bottle-nipples were collected at day care
facilities and households on the day of use.

Enrichment and isolation of Cr onobacter spp.
During the period of investigation, isolation and identification methodology for this
particular micro-organism experienced rapid improvement and innovation. For this reason,
Prevalence of Cronobacter spp.

83
different methods were employed in the course of the study. Table 1 gives an overview of
the methods, which is further described in the following, with reference to the various
abbreviated identification codes used to identify the specific methods. Irrespective of the
method employed, all presumptive Cronobacter spp. colonies were further identified using
the API 20
E
system (bioMérieux, Marcy l’Etoile, France) and subsequently confirmed by a
specific PCR assay (see below for PCR procedure details).
From the powdered food products collected in 2001-2003, 25 g or 50 g portions were
carefully sprinkled onto a nine-fold volume of buffered peptone water (BPW) at a
temperature of 45 °C. After soaking for 30 to 60 min on the work bench, the powders were
dissolved by swirling slowly. Following overnight incubation at 37 °C, the pre-enrichments
were streaked onto violet red bile lactose (VRBL; Oxoid Ltd., Basingstoke, Hampshire,
England) agar (Method A1) and 10 ml portions were mixed with 90 ml of
Enterobacteriaceae enrichment (EE; Oxoid Ltd.) broth. After overnight incubation at
37 °C, the EE cultures were streaked onto VRBL agar plates and incubated overnight
at 37 °C (Method A2). Typical coliform-colonies were picked and tested for the production
of yellow-pigmented colonies on Tryptone Soya agar (TSA; Oxoid Ltd.) and a positive α-
glucosidase reaction within 4 h, as described previously (45). Positive isolates were further
identified and confirmed as mentioned above.
From the BPW pre-enrichments of the dry powdered products and dry cereals collected in
2004 and 2005, 0.1 ml was added to 10 ml of lauryl sulphate tryptose broth (Oxoid Ltd.)
supplemented with 0.5 M sodium chloride (NaCl) and 1 mg/ml vancomycin (Oxoid Ltd.),
further referred to as mLST (25). After overnight incubation at 44 ± 0.5 ºC, the selective
enrichments were directly streaked onto TSA plates (Method B). Colonies that produced
yellow pigment after overnight incubation at 37 °C were tested with the E. sakazakii
specific 4-h α-glucosidase assay (45), whereupon positive isolates were further identified
and confirmed.
The samples of raw meats were examined essentially according to procedure B, except that
“Enterobacter sakazakii isolation agar” (ESIA; AES Chemunex, Bruz cedex, France) was
used instead of TSA (Method C1). The ESIA plates were incubated overnight at 44 ± 0.5 ºC
Chapter 5
84
and turquoise-coloured colonies were picked off and streaked onto TSA for further
identification and confirmation as described above.
The raw milk samples (10 ml) and the entire scrub gloves containing human skin particles
were directly selectively enriched, in mLST (90 and 100 ml, respectively) without pre-
enrichment in BPW. The mLST enrichments were streaked onto ESIA plates (Method C2).
Turquoise coloured-colonies were streaked onto TSA for further identification and
confirmation.
Samples of human faeces, used baby bottles and bottle-nipples, were enriched in
“Enterobacter sakazakii selective broth” (ESSB; AES Chemunex). Ten grams of the human
faeces samples were added to 90 ml ESSB. Samples of used baby bottles were obtained by
rinsing the inside of the bottles with 10 ml sterile peptone physiological salt solution
containing 0.1% Tween 80 (Van Merck-Schuchardt, Hohenbrunn, Germany) and adding
the rinsing solution to 90 ml ESSB. Bottle-nipples were individually placed directly into
100 ml ESSB. Following 24 h incubation at 37 °C, the ESSB enrichments were streaked
onto ESIA plates (Method D). Presumptive Cronobacter spp. colonies were further
confirmed and identified as described above.

Table 1. Methods for the isolation of Cronobacter spp. used in this study
Method Pr e-enr ichment Selective enr ichment Isolation
A1 BPW
a
n.d.
b
VRBL
c
A2 BPW EE
d
VRBL
B BPW mLST
e
TSA
f
C1 BPW mLST ESIA
g
C2 n.d. mLST ESIA
D n.d. ESSB
h
ESIA
E n.d. n.d. ESIA+TSA+VRBL+DFI
i

a
BPW, buffered peptone water.
f
TSA, tryptone soy agar.
b
n.d., not included in this method.
g
ESIA, “Enterobacter sakazakii isolation agar”.
c
VRBL, violet red bile lactose agar.
h
ESSB, “Enterobacter sakazakii selective broth”.
d
EE, Enterobacteriaceae enrichment broth.
i
DFI, Druggan-Forsythe-Iversen agar.
e
mLST, modified lauryl sulphate tryptose broth.

Prevalence of Cronobacter spp.

85
The vegetables and spices were homogenized in peptone physiological salt solution (1:10
(w:v)) and dilutions were not enriched but directly streaked onto the following four media:
ESIA, TSA, VRBL and Druggan-Forsythe-Iversen (DFI; Oxoid Ltd.) agar plates (Method
E). Typical colonies were confirmed with the following biochemical identification
methods: API 20
E
(bioMérieux), Microbact (Oxoid Ltd.), BBL Crystal (Becton Dickinson
and Company, Sparks Maryland, USA), Enterotube II (Becton Dickinson an Company) and
the presence of α-glucosidase (45) and subsequently confirmed with the PCR assay.

Enumeration of Enter obacter iaceae
In order to assess the general hygienic status of all products, all samples collected in 2005,
except for two samples of flavored instant drinks and one sample of coffee creamer, were
examined for the number of Enterobacteriaceae (CFU/g) by using the pour plating
technique, in violet red bile glucose (VRBG; Oxoid Ltd.) agar (32). Colonies of
presumptive Enterobacteriaceae were, in accordance to the ISO 21528-1 method,
confirmed by means of tests for fermentation of glucose and the presence of oxidase (31).

PCR confirmation of Cr onobacter spp.
Presumptive Cronobacter spp. colonies were confirmed using a newly developed PCR
assay, which is specific for all Cronobacter species tested. The new real-time PCR method
targets a specific sequence within the gene encoding the major outer membrane protein
(OmpA) of Cronobacter spp. (GenBank accession number DQ000206; National Centre for
Biotechnology Information, Washington, D.C., USA). The set of oligonucleotides was
designed using Primer Express software (Applied Biosystems, Foster City, USA) (Table 2).
The specificity of the sequences was tested by a blast search in GenBank (BLASTn version
2.2.10, National Centre for Biotechnology Information). The oligonucleotides were
purchased from Eurogentec (Liege, Belgium). The Cronobacter spp. probe was labelled at
the 5’ end with the reporter dye 6-carboxy-fluorescein (FAM) and at the 3’ end with the
black-hole quencher (BHQ).
Chapter 5
86
Table 2. Primers and probe used for the real-time PCR assay specific for Cronobacter spp.
Designation Sequence (5’ → 3’) Position
a

Esak ompA-F ggt gaa gga ttt aac cgt gaa ctt 319-342
Esak ompA-R gcg cct cgt tat cat cca aa 388-369
Esak ompA-probe ccc gga aaa gcg cat ggc c 348-366
a
Positions correspond to GenBank accession no. DQ000206

For selectivity tests, 35 isolates of Cronobacter spp. and 44 isolates representing a range of
other genera and species were grown overnight at 37 °C on TSA. One loop-full of bacterial
culture was suspended in 200 μl sterile distilled water, whereupon 100 μl was used for
DNA extraction. DNA was extracted using the automated MagNA Pure Compact® (Roche
Diagnostics, Mannheim, Germany) system and the MagNA Pure Compact® Nucleic Acid
Isolation kit I (Roche Diagnostics). Amplification and detection were carried out in a
LightCycler Instrument (version 2.0, Roche Diagnostics) in a 20 μl PCR mixture containing
2.5 μl of the sample DNA, 500 nM each of the primers, 250 nM of the probe, and 4 μl
LightCycler TaqMan Master mix (Roche Diagnostics). Samples were amplified with an
initial denaturation step at 95 °C for 10 min to activate the FastStart Taq DNA polymerase
and 40 cycles of denaturation at 95 °C for 10 s and annealing and extension at 60 °C for 15
s. The temperature transition rate was 20 °C/s. Samples positive for the target gene were
identified by the instrument at the cycle number where the fluorescence attributable to the
target sequences exceeded that measured for background. Those scored as positive by the
instrument were confirmed by visual inspection of the graphical plot (cycle number versus
fluorescence value) generated by the instrument.

Subtyping of Cr onobacter spp.
The Pulsed Field Gel Electrophoresis (PFGE) technique of contour-clamped homogeneous
electric fields (CHEF) was used for genomic typing of isolates confirmed as Cronobacter spp.
by PCR. Genomic DNA was digested in agarose plugs with XbaI (10 U) (Roche Diagnostics).
The resulting fragments were resolved by CHEF-PFGE with a CHEF DR-III apparatus (Bio-
Prevalence of Cronobacter spp.

87
Rad Laboratories, Richmond, California, USA) at a constant voltage of 200 V for 24 h at 13 °C
and a linearly ramped pulse time of 5 to 60 s. The fingerprints generated were processed using
BioNumerics software (Applied Maths, Kortrijk, Belgium). Isolates that had more than
95% fragments in common were designated “closely related”. Isolates were considered
“indistinguishable” if 100% of fragments were identical.

Results
Selectivity of the real-time PCR
As shown in Table 3, all 35 Cronobacter spp. strains tested, which included 13 official
reference strains and 22 isolates from various sources obtained in previous investigations at
the Laboratory of Food Microbiology of theWageningen University (43), gave a positive
signal in the real-time PCR assay. All of the following 44 strains gave a fluorescence end-
point signal below the threshold line (n=1 unless indicated otherwise): Aeromonas
hydrophila, Citrobacter spp., Citrobacter freundii (n=2), Enterobacter amnigenus,
Enterobacter aerogenes, Enterobacter cloacae (n=6), Enterococcus faecium, Escherichia
coli, Escherichia coli O157, Hafnia alveii, Klebsiella pneumoniae (n=3), Klebsiella spp.,
Klebsiella oxytoca (n=2), Listeria innocua, Listeria monocytogenes (n=2), Pseudomonas
aeruginosa (n=3), Pseudomonas putida, Proteus mirabilis (n=2), Proteus vulgaris,
Salmonella Enteritidis, Salmonella Typhimurium (n=2), Serratia marcescens, Shigella
flexneri (n=2), Shigella sonnei (n=3), and Yersinia enterocolitica (n=3).

Isolation of Cr onobacter spp.
Table 4 summarizes the results of the isolation of Cronobacter spp. from the different
samples investigated in this study. The micro-organism was isolated from the following dry
powdered food sources tested: 8 out of a total of 395 powdered formulae, 1 out of 15
follow-up formulae, 7 out of 175 milk powders, 1 out of 182 of the other powdered instant
products (including ice cream mixtures, instant pudding powders, flavoured instant drinks,
Chapter 5
88
dips and coffee creamers). Other food products yielding Cronobacter spp. isolates were:
dry cereals (6 out of 123), raw minced meats (7 out of 222), vegetables (2 out of 47) and
spices (1 out of 28). In addition, the micro-organism was isolated twice from human
sources, including faeces (1 out of 98 fresh human faeces) and skin scrubbings (1 out of
116 human skin scrub samples). Cronobacter spp. was not isolated from the fresh raw
cow’s milk samples nor from the samples of fresh human breast milk, used baby bottles or
bottle-nipples. The presumptive Cronobacter spp. isolates obtained throughout the study
period, with the exception of the human faecal isolate, were all confirmed using a newly
developed real-time PCR assay, which proved to be highly specific for Cronobacter spp.,
as shown in Table 3 and shown below.

Table 3. Real-time PCR results of 79 test strains, using the probes as described in Table 2
Str ain Str ain code/Or igin Results r eal-
time PCR
a

Cronobacter sakazakii
b
Enterobacter sakazakii NCTC 8155 +
Cronobacter sakazakii Enterobacter sakazakii NCTC 9238 +
Cronobacter sakazakii Enterobacter sakazakii NCTC 11467 +
Cronobacter sakazakii Enterobacter sakazakii DSM 4485 +
Cronobacter dublinensis Enterobacter sakazakii NCTC 9844
Cronobacter dublinensis Enterobacter sakazakii NCTC 9846 +
Cronobacter genomospecies 1 Enterobacter sakazakii NCTC 9529 +
Cronobacter dublinensis subsp.
dublinensis
Enterobacter sakazakii DSM 18705 +
Cronobacter muytjensii Enterobacter sakazakii ATCC 51329 +
Cronobacter turicensis Enterobacter sakazakii DSM 18703 +
Cronobacter malonaticus Enterobacter sakazakii DSM 18702T +
Cronobacter dublinensis subsp.
lausannensis
Enterobacter sakazakii DSM 18706T +
Cronobacter dublinensis subsp.
lactardi
Enterobacter sakazakii DSM 18707T +
Cronobacter spp. (n=22)
c
Enterobacter sakazakii Isolates from
Wageningen University culture collection
+
Aeromonas hydrophila M800 -
Citrobacter spp. PHLS 073-177 -
Citrobacter freundii ATCC 8090 -
Citrobacter freundii NCTC 6272 -

Prevalence of Cronobacter spp.

89
Table 3. Continued
Str ain Str ain code/Or igin Results r eal-
time PCR
a

Enterobacter amnigenus Isolate VWA -
Enterobacter aerogenes Isolate VWA -
Enterobacter cloacae ATCC 23355 -
Enterobacter cloacae (n=5) Isolates VWA -
Enterococcus faecium ATCC 29212 -
Escherichia coli ATCC 25922 -
Escherichia coli O157 NCTC 12900 -
Hafnia alveii NCTC 8105 -
Klebsiella spp. Isolate VWA -
Klebsiella oxytoca ATCC 49131 -
Klebsiella oxytoca Isolate VWA -
Klebsiella pneumoniae ATCC 33495 -
Klebsiella pneumoniae ATCC 13883 -
Klebsiella pneumoniae Isolate VWA -
Listeria innocua NCTC 11288 -
Listeria monocytogenes SLCC 2479 -
Listeria monocytogenes NCTC 5348 -
Pseudomonas aeruginosa ALM32 -
Pseudomonas aeruginosa ATCC 27853 -
Pseudomonas aeruginosa NCTC 10662 -
Pseudomonas putida Isolate VWA -
Proteus mirabilis ALM29 -
Proteus mirabilis NCTC 11938 -
Proteus vulgaris ATCC 13315 -
Salmonella Enteritidis ALM36 -
Salmonella Typhimurium ALM40 -
Salmonella Typhimurium ATCC 13311 -
Serratia marcescens ATCC 13880 -
Shigella flexneri ALM60 -
Shigella flexneri Isolate VWA -
Shigella sonnei ATCC 11060 -
Shigella sonnei Isolate GG&GD Haarlem -
Shigella sonnei Isolate VWA -
Yersinia enterocolitica ALM5 -
Yersinia enterocolitica O3 Isolate UMC St. Radboud, Nijmegen -
Yersinia enterocolitica O9 CCUG8239A -
a
+, positive; -, negative
b
The Cronobacter species identification is according to Iversen et al., 2007 (39).
c
Presumptively identified by the α-glucosidase assay (45) and the API 20
E
system (bioMérieux).
Chapter 5
90
Table 4. Prevalence of Cronobacter spp. in samples of various origins, using methods as described
in Table 1
Sample Pr evalence of
Cronobacter spp.
# positive/total (%)
95%
Confidence
inter val (%)
a

Year of
sampling
Sample
size
Method
b

Powder ed foods
Milk powder 7
c
/ 171
d
(4.1)
1.8 – 7.6
2001 25 g A
e

0 / 4
0 – 60.2
2005 50 g B
PF < 1 yr 1 / 40 (2.5) 0.1 – 12.8 2001 25 g A
2 / 101 (2.0)
0.3 – 6.6
2002 25 g A
0 / 101
0 – 3.6
2003 50 g A
5 / 139 (3.6)
1.3 – 7.6
2004 50 g B
0 / 14
0 – 23.1
2005 50 g B
1 / 11 (9.1) 0.2 – 40.4 2004 50 g B

Follow-up
formulae, >1 yr
0 / 4
0 – 60.2
2005 50 g B
0 / 87 0 – 4.1 2003 50 g A

Ice cream
mixtures
0 / 2
0 – 84.1
2005 50 g B
Flavored instant
drink
0 / 33 0 – 10.6 2005 50 g B
0 / 12 0 – 26.5 2003 50 g A

Instant pudding
powder
1 / 25 (4.0)
0.1 – 19.8
2005 50 g B
Dip 0 / 19
0 – 17.7
2005 50 g B
Coffee creamer 0 / 4
0 – 60.2
2005 50 g B
Dr y cer eals 6 / 123 (4.9) 2.0 – 9.6
2005
50 g B
Raw minced
meats



Beef 2 / 141 (1.4)
0.2 – 9.6
2005 50 g C1
Mixed beef /
pork
5 / 50 (10.0) 3.6 – 20.5 2005 50 g C1
Pork 0 / 16
0 – 20.6
2005 50 g C1
Calf 0 / 10
0 – 30.9
2005 50 g C1
Poultry 0 / 5 0 – 52.2 2005 50 g C1
Dr y vegetables 2 / 47 (4.2) 0.5 – 13.9
2005
25 g E
Spices 1 / 28 (3.6)
0.1 – 17.8
2005 10 g E
Raw cow’s milk 0 / 51 0 – 7.0
2005
10 ml C2
Human milk 0 / 7 0 – 41.0
2005
10 ml C2
Human faeces 1 / 98 (1.0) 0 – 5.4
2005
10 g D
Human skin 1 / 116 (0.9) 0 – 4.6 2005 Scrub
glove
C2

Prevalence of Cronobacter spp.

91
Table 4. Continued
Sample Pr evalence of
Cronobacter spp.
# positive / total (%)
95%
Confidence
inter val (%)
a

Year of
sampling
Sample
size
Method
Baby bottles 0 / 86 0 – 4.2 2005 10 ml rinse D
Bottle-nipples 0 / 95 0 – 3.8 2005 each D

a
The 95% confidence interval was calculated, assuming the drawing process is a series of
independent draws from a binomial distribution with the probability p.
b
The methods used are summarized in Table 1.
c
3 samples were produced in The Netherlands, 2 in Ukraine, and 2 in Poland.
d
Samples collected from cargos and bulk goods and included 27 products intended for animal feed.
e
A, includes use of either method A1 or method A2.

Subtyping of Cr onobacter spp.
In total, 35 isolates were obtained of which 34 were positively confirmed by PCR, the
remaining isolate (from human faeces) had not been kept for molecular confirmation and
PFGE subtyping. Of the total of 35 Cronobacter spp. isolates, 34 were subjected to PFGE
subtyping. Because of degradation of genomic DNA, four of the 34 isolates subjected to the
subtyping procedure (one of the six isolates from dry cereals, the isolate from the pudding
powder and one of the two beef isolates and the human skin isolate) did not yield a result.
Of the remaining 30 isolates that had been successfully subtyped, 28 isolates generated a
unique XbaI restriction pattern each, whereas two isolates generated patterns that could not
be distinguished (Figure 1). These two isolates were cultured from two samples of minced
mixed beef and pork, which had been purchased from two different outlets of the same
supermarket chain, visited one month after each other.

Chapter 5
92

Figur e 1. Cluster analysis of PFGE fingerprints of the 30 Cronobacter spp. isolates.
Note that the PFGE fingerprints of the faeces isolate is not included in this Figure, as it was not
analysed. * indicates the two generated patterns that could not be distinguished.
Prevalence of Cronobacter spp.

93
Enter obacter iaceae counts
Enterobacteriaceae counts were investigated in products sampled in 2005, including dry
powders and cereals as well as raw minced meats, (Table 5). However, for three samples
Enterobacteriaceae counts were not done.
A total of 223 dry powder samples had Enterobacteriaceae counts below the detection limit
(≤ 10 CFU/g), and these included 5 of the 6 dry cereals samples that had been found
positive for Cronobacter spp.. The sixth dry cereal sample (a type of muesli) that had been
found positive for Cronobacter spp., contained Enterobacteriaceae at 15 CFU/g. The
number of Enterobacteriaceae in the instant pudding powder that yielded Cronobacter spp.
was below the detection limit. A wheat bran sample containing over 10,000 CFU/g
Enterbacteriaceae, was negative for Cronobacter spp..
Enterobacteriaceae counts were found to be quite variable in the various minced meats.
Out of 222 samples, 28 (13%) had counts below detection, 47 (21%) between 10 and 100
CFU/g, 80 (36%) between 100 and 1,000 CFU/g, 43 (19%) between 1,000 and 10,000
CFU/g, and 24 (11%) over 10,000 CFU/g. For the seven meat samples positive for
Cronobacter spp., the Enterobacteriaceae counts were: s 10 CFU/g, 55 CFU/g, 91 CFU/g,
110 CFU/g, 250 CFU/g, 680 CFU/g, and 2,000 CFU/g.

Table 5. Enumeration of Enterobacteriaceae in products collected in 2005
Enter obacter iaceae (CFU/g) Dr y pr oducts
a
Raw minced meats
10
b
223 28
10 - 100 1 47
100 – 1,000 0 80
1,000 – 10,000 0 43
> 10,000 1
c
24
a
All dry samples collected in 2005, except for two samples of flavored instant drinks and one sample
of coffee creamer.
b
Detection limit.
c
Wheat bran.

Chapter 5
94
Discussion
Isolation and confirmation methods
An investigation of the occurrence of Cronobacter spp. and its prevalence in a variety of
samples of different origin was carried out in The Netherlands between the years 2001 and
2005. During this study, various traditional and innovative isolation and (molecular)
detection/confirmation methods were used, following in a timely way the ongoing
developments and improvements in this field. Several isolation and (molecular) detection
methods have been developed over the past years aiming to detect and type Cronobacter
spp., in particular in powdered infant formulae (13, 25, 33, 34, 45, 46, 51, 52, 65, 70).
Regarding detection, several chromogenic and fluorogenic agar media have been described
for Cronobacter spp. (34, 52, 68), mainly based on its unique feature, being the presence of
α-glucosidase, as compared to other Enterobacter species (60). The results obtained with
the specific real-time PCR method, which we developed to further improve detection and
definite confirmation, indicated that it represents a reliable confirmation test (Table 3)
which is able to differentiate all Cronobacter spp. from many other Enterobacteriaceae.
The PFGE subtyping method also showed to be a reliable method to subtype Cronobacter
spp. isolates. It may proof useful for tracing infections to the source in sporadic cases or
outbreak investigations, or for the identification of contamination sources in food
manufacturing operations.
While most of the recent studies focused on the development of molecular
detection/confirmation of Cronobacter spp., further research should focus on improving the
enrichment step, ideally providing a specific and sensitive enrichment procedure, as this so
far remains an important bottle-neck (13, 35, 37, 41, 64). This is a challenge considering
the relative low levels of this micro-organism that are found in foods and the fact that it is
present among other micro-organisms. The ISO procedure for milk and milk products
published in 2006 (33) uses a selective enrichment procedure part of which was applied in
the current study. In the selective enrichment step, mLST broth is used (25) in combination
with a relatively high incubation temperature to suppress potential competitors, as in our
Methods B, C1, and C2 (Table 2). However, it has been demonstrated that some
Prevalence of Cronobacter spp.

95
Cronobacter spp. isolates do not grow well in mLST (37); therefore recently, a new
enrichment broth for Cronobacter spp., Cronobacter screening broth (CSB), has been
developed (35).

Occurrence and prevalence of Cr onobacter spp.
The results obtained (Table 4) show that also in The Netherlands, this opportunistic
pathogen may be present in a variety of foods and feeds, either being produced within the
country or reaching the national market via import from other countries, and substantiates
humans as potential sources for contamination.
The prevalence of Cronobacter spp. on human skin and in fecal samples was rather low
(~1%), with the micro-organism being isolated once from a faecal sample of a female adult,
as well as once from skin particles collected from another female adult. Notably, the micro-
organism was not detected in the seven human breast milk samples investigated. Farmer et
al., (1980) previously reported isolating the micro-organism from various clinical sources,
including a child’s stool (21). More recent studies also reported that Cronobacter spp. could
be isolated from the stool of a child with diarrhea (66) and from stool samples of
hospitalised infants (0.5%) and patients in the age group 61-70 years (6.5%) (48). While the
occurrence of Cronobacter spp. in human faeces in the current investigation is not a new
observation, its presence in skin scrapings is. These observations highlight that the micro-
organism can be present on or in humans, making them a potential source of contamination
both during product manufacture as well as food preparation.
Our study of Cronobacter spp. in a variety of food and feed products, with prevalences
ranging from 1.4 to 10.0% of samples where positives were observed, confirms and further
extend the results of previous studies on the occurrence of the micro-organism. Though
prevalences have not been systematically reported, previously, Cronobacter spp. has been
isolated from powdered infant formulae and milk powders (8, 10, 18-20, 36, 37, 57, 59) as
well as from a variety of other foods (3, 22, 26), food factory environments (43, 58, 71),
consumer homes (45), and environmental samples, including soil (15) and insects (27, 49,
Chapter 5
96
53). In a recent review (22), the widespread occurrence of Cronobacter spp. in food and
beverages other than infant formula was documented, noting that the presence of the micro-
organism had been confirmed in several foods and food ingredients from plant origin, such
as cereals, fruits and vegetables, legume products, herbs and spices, as well as in foods of
animal origin such as milk, meat and fish and products made from these foods. The
spectrum of food contaminated with Cronobacter spp. seems to cover both raw/fresh and
processed food (being dried, frozen, ready-to-eat, fermented and cooked food products), as
well as water suitable for food preparation.
Overseeing the variety of sources of Cronobacter spp. identified to-date, it is not apparent
that there is one single, most important natural habitat for this micro-organism. In the
course of previous research into the occurrence of Cronobacter spp. in dry environments, a
range of other Enterobacteriaceae were isolated, such as Klebsiella spp., Citrobacter spp.,
Erwinia spp., Enterobacter cloacae, Escherichia coli, Enterobacter agglomerans (Pantoea
agglomerans) and Serratia spp. (45). Based on the habitat profile of these possibly co-
inhabiting micro-organisms and their physiological features, it is likely that these micro-
organisms (and may be thus also Cronobacter spp.) are associated with the phytic flora (56,
69). Recent research indeed showed that Cronobacter spp. could be isolated from vegetable
products (22, 42, 48, 69) and the vegetable drawer of domestic refrigerators (2.2% out of
137 refrigerators) (47). Our results extend these observations, as Cronobacter spp. was
isolated from dried vegetables (4.2%) and spices (3.6%) (Table 4).
So far only powdered infant formulae have been definitely attributed to cases of
Cronobacter spp. infection, and most of the investigative attention has been focused on
these products. Nevertheless, actual data on prevalence and concentration are scarce.
Concerning prevalence in powdered formulae, some reported prevalence figures in positive
batches are about 1% and 3% (data quoted in FAO/WHO, 2008) (20), 6.7% (74) and less
recently 14% (59). Relating to concentration, operations with good practices in place
should be able to achieve levels of Cronobacter spp. of 10
-4
CFU/g or better (6), while
levels in contaminated products may be of the order of 10
-3
CFU/g (12, 19). The current
study did not determine the actual levels of Cronobacter spp. in various sources; it only
investigated its prevalence. Following prevalence data were obtained (with prevalence
Prevalence of Cronobacter spp.

97
expressed per specific investigation, as indicated in Table 4): 4.1% for milk powders, 2.0-
3.6% for powdered formulae intended for consumers < 1 year, 9.1% for follow up formula
intended for children > 1 year, 4.0% for instant pudding powders, 4.9% for dry cereals, 1.4
to 10.0% for raw minced meats, 4.2% for dried vegetables and 3.6% for spices. Notably, in
our specific survey some samples from milk powders, powdered formulae, instant pudding
powders, and raw minced meats tested negative for Cronobacter spp. as did dry powder
products such as ice-cream mixtures, instant drinks, dips and coffee creamers. It is apparent
that prevalence rates of Cronobacter spp. in powdered formulae and other foods/feeds are
in the similar low range as reported for PIF as mentioned above and concur with prevalence
estimates of Jung and Park of 5 to 10% for agricultural food components such as brown rice
and tomatoes (42). The relatively low frequency that Cronobacter spp. were detected in our
study and others may be related to limitations of (selective) enrichment and isolation
methods used, but may as well reflect that that the level at which Cronobacter spp. actually
is present in various products/samples indeed is very low.
In powdered products and cereal samples taken in 2005 and raw minced meat samples, the
Enterobacteriaceae levels were determined to possibly link product hygiene to the
occurrence of Cronobacter spp.. Table 5 clearly shows that most of the dry products
(including milk powders) contained low levels of Enterobacteriaceae (< 10 CFU/g). The
apparently sound hygiene status of the products investigated is important because milk
powders represent a basic ingredient in the majority of dry powdered foods, including dry-
mixed powdered infant formulae and follow-up formulae. All these products (being
powdered infant formula, follow-up formula, and special diet formula) thus are key to the
health and well-being of infants and young children and our finding that the prevalence of
Cronobacter spp. in these products is relatively infrequent or absent and that levels of
Enterobacteriaceae in general were found to be below the detection limit, are positive
indicators of the product safety and hygiene.

Of the two cereal samples showing levels of Enterobacteriaceae above detection (Table 5),
the one with the lowest level of contamination (a muesli product) was positive for
Chapter 5
98
Cronobacter spp., while the sample with the highest level of contamination had been found
negative. Indeed, it has been reported before that although Enterobacteriaceae may be a
good general hygiene indicator and are promoted as such by Codex Alimentarius for
powdered infant formulae (6), they do not necessarily provide any insights in the
occurrence of Cronobacter spp. (17).

Considering Cr onobacter spp. as a hazard in food manufacture
This publication is one of a series of studies over the last several years by many
investigators all dedicated to the elucidation of the occurrence and prevalence of
Cronobacter spp.. Overall, a picture has emerged that the pathogen is widespread in nature,
factory environments and people’s homes, occurring in foods and food ingredients, soil,
insects, in/on humans and utensils, etc. Thus, there may be many sources and vectors that
can play a role in the spread and ultimate contamination of foods by the opportunistic
pathogen. To consider all such sources and vectors may not be practical nor necessary in
assuring food safety for the relevant consumers. There may be relatively fewer relevant
sources and vectors to control, considering the actual consumer risk the micro-organism
poses when present in a food. Consumer risk is determined by a combination of factors,
including sensitivity of the consumer and the actual exposure of consumers to the pathogen
through consumption.
Risk assessments or comparable studies developed under the auspices of the Food and
Agricultural Organization of the United Nations (FAO) and the World Health Organization
(WHO) (18, 19) and of the European Food Safety Authority (EFSA) (16) have shown that
Cronobacter spp. poses different levels of risk of infection for different consumer age
groups. Among infants, those at greatest risk of infection are neonates (< 28 days),
particularly pre-term, low-birth weight (< 2500 g), and immunocompromised infants and
those less than two months of age and also infants of HIV-positive mothers (20). Cases
occurring in this age-group are relatively well confirmed and documented. Also other
infants below the age of 1 year are at particular risk, though solid case reports are fewer.
Above that age limit, confirmed and documented cases are extremely sparse at the moment.
To limit consumer risk, tightest control is needed for products intended for the consumers at
Prevalence of Cronobacter spp.

99
highest risk, i.e. for powdered formulae intended for very young infants. Recent guidance
developed by Codex Alimentarius (6) therefore contains microbiological criteria for
product safety related to Cronobacter spp. in powdered infant formula, formula for special
medical purposes and human milk fortifiers.
Considering consumer exposure and risk, there are various routes by which Cronobacter
spp. can end up in powdered formulae (6), two important ones relate to food manufacture:
1) through ingredients added in dry mixing operations during the manufacturing of
powdered formula,
2) through contamination of the formula from the processing environment in the
steps during or following the drying.
Whilst control of contamination of ingredients to some extent is possible by applying some
form of thermal treatment during their production and assuring prevention of post treatment
re-contamination, adherence to strict microbiological requirements and thorough selection
and purchase procedures (12), current technology cannot completely eliminate this micro-
organism from the manufacturing environment and prevent it gaining access to the
processing line and final food product. Therefore, Codex provided guidance for the
establishment of monitoring programs for Cronobacter spp. in high hygiene processing
areas and in powdered formula preparation units (6).
A recent Microbiological Risk Assessment conducted under the auspices of the Food and
Agricultural Organization of the United Nation and the World Health Organization (19) has
shown that factory hygiene may be a factor contributing to the risk posed by Cronobacter
spp. in powdered infant formulae as it impacts on the ultimate consumer exposure. The
current study once more highlighted the possibility for humans as vectors for the micro-
organism. These results suggest that also the hygiene awareness of personnel need to be
taken into consideration to prevent or reduce the prevalence of Cronobacter spp. in the
factory environments. Improvements in processing conditions such as very strict
management of water in high hygiene zones, may help further reduce the level
Enterobacteriaceae, including Cronobacter spp.. Cronobacter spp. are present in the milk
powder processing environment (43, 54, 58), probably due to their resistance to desiccation
Chapter 5
100
(5, 7, 14). Presence of water could lead to outgrowth, since strains of Cronobacter spp.
have relatively short lag times and high specific growth rates at temperatures commonly
found in those environments (38, 44). Due to its prevalence in dry environments, such as
dust samples in factories and households (43), total elimination may be difficult, if not
impossible, to achieve. Therefore and because contamination and proliferation of
Cronobacter spp. in prepared formula prepared in hospitals and homes has been recognised
as a key risk factor for sensitive consumers as well (19, 67), proper bottle preparation and
handling practices are critical in order to sufficiently control the risk. It is noteworthy that
used baby bottles and bottle-nipples were found to be negative for Cronobacter spp., which
may indicate an appropriate level of hygiene at the day care facilities and homes sampled
from. Nevertheless, caregivers should be mindful of these risk factors in preparing and
handling infant formulae since Cronobacter spp. cells are able to survive over a long period
in dry powdered infant formulae (7, 14) and can grow rapidly (44), once the powder has
been reconstituted.

Conclusion
Cronobacter spp. is not a pathogen for the majority of the population, but rather a concern
for specific sub-populations in the public. Measures for prevention should consider the
actual target sub-population of a food as well as the consumer exposure due to the presence
of the micro-organism in food. For relevant foods and sub-populations, strict control of
contamination from food ingredients, during manufacture and processing as well as final
preparation is paramount to limit the consumer risk. One should consider using sterile
liquid formula, for those groups at greatest risk.
Our results do not point to a specific niche of Cronobacter spp., but rather extend already
available evidence that Cronobacter spp. can be found in many different food products,
food environments as well as on and in humans, making all these potential contamination
sources of the micro-organism for foods during manufacture and preparation. The
concentration and prevalence of Cronobacter spp. is not equal in all the potential sources
and that not all sources are linked to the production and preparation of food for the
Prevalence of Cronobacter spp.

101
consumers that are at most risk. These foods are milk powder and infant formulae and the
consumers at most risk are neonates (< 28 days), particular pre-term, low-birth weight (<
2500 g), and immuno-compromised infants, and those less than two months of age and also
infants of HIV-positive mothers. Thus, it is a key that for these relevant foods and for these
particular consumers, adequate control measures are being implemented that take into
account the widespread occurrence of Cronobacter spp. as well as the contamination routes
and growth opportunities that are still presented to the micro-organism in some of current
practice. Over the past years several improvements have been made to the (pre)-enrichment
step and microbiological culture methods for the recovery of Cronobacter spp. and this has
helped to better investigate the prevalence in various locations.

Acknowledgements
This research has been financially supported by Nestlé Research Center, Lausanne,
Switzerland. The authors would like to thank Ellen Brinkman and Ornella Vázquez for their
contribution in collecting and analyzing samples, and also the volunteers who donated
faecal and skin particle samples. Marcel Zwietering is gratefully acknowledged for
critically reviewing the manuscript.

Refer ences
1. Acker Van J., F de Smet, G. Muyldermans, A. Bougatef, A. Naessens, and S.
Lauwers. 2001. Outbreak of necrotizing enterocolitis associated with Enterobacter
sakazakii in powdered milk formula. J. Clin. Microbiol. 39:293-297.
2. Bar Oz, B., A. Preminger, O. Peleg, C. Block, and I. Arad. 2001. Enterobacter
sakazakii infection in the newborn. Acta Paediatr. 90:356-358.
3. Baumgartner, A., M. Grand, M. Liniger, and C. Iversen. 2009. Detection and
frequency of Cronobacter spp. (Enterobacter sakazakii) in different categories of
ready-to-eat foods other than infant formula. Int. J. Food Microbiol. 136:189-192.
4. Biering, G., S. Karlsson, N. C. Clark, K. E. Jonsdottir, P. Ludvigsson, and O.
Steingrimsson. 1989. Three cases of neonatal meningitis caused by Enterobacter
sakazakii in powdered milk. J. Clin. Microbiol. 27:2054-2056.
Chapter 5
102
5. Breeuwer, P., A. Lardeau, M. Peterz, and H. Joosten. 2003. Desiccation and heat
tolerance of Enterobacter sakazakii. J. Appl. Microbiol. 95:967-973.
6. CAC. 2008. Code of hygienic practice for powdered formulae for infants and
young children. CAC/RCP 66-2008.
7. Caubilla-Barron, J., and S. J. Forsythe. 2007. Dry stress and survival time of
Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered
infant formula. J. Food Prot. 70:2111-2117.
8. Cawthorn, D. M., S. Botha, and R. C. Witthuhn. 2008. Evaluation of different
methods for the detection and identification of Enterobacter sakazakii isolated
from South African infant formula milks and the processing environment. Int. J.
Food Microbiol. 127:129-138.
9. CDC. 2002. Enterobacter sakazakii infections associated with the use of powdered
infant formula - Tennessee, 2001. MMWR. 51:297-300.
10. Chap, J., P. Jackson, R. Siqueira, N. Gaspar, C. Quintas, J. Park, T. Osaili, R.
Shaker, Z. Jaradat, S. H. P. Hartantyo, N. A. Sani, S. Estuningsih, and S. J.
Forsythe. 2009. International survey of Cronobacter sakazakii and other
Cronobacter spp. in follow up formulas and infant foods. Int. J. Food Microbiol.
136:185-188.
11. Clark, N. C., B. C. Hill, C. M. O'Hara, O. Steingrimsson, and R. C. Cooksey.
1990. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial
outbreaks. Diagn. Microbiol. Infect. Dis. 13:467-472.
12. Cordier, J. L. 2008. Production of powdered infant formulae and microbiological
control measures. p. 145-185. In: J. M. Farber, and S. J. Forsythe (eds.),
Enterobacter sakazakii (emerging issues in food safety) ASM Press, Washington,
D.C., New Brunswick, New Jersey.
13. Druggan, P., and C. Iversen. 2009. Culture media for the isolation of Cronobacter
spp. Int. J. Food Microbiol. 136:169-178.
14. Edelson-Mammel, S. G., and R. L. Buchanan. 2005. Survival of Enterobacter
sakazakii in a dehydrated powdered infant formula. J. Food Prot. 68:1900-1902.
15. Edrington, T. S., W. E. Fox, T. R. Callaway, R. C. Anderson, D. W. Hoffman, and
D. J. Nisbet. 2009. Pathogen prevalence and influence of composted dairy manure
application on antimicrobial resistance profiles of commensal soil bacteria. Foodb.
Path. Dis. 6:217-224.
16. EFSA. 2004. Opinion of the scientific panel on biological hazards on the request
from the commission related to the microbiological risks in infant formulae and
follow-on formulae. The EFSA Journal. 113:1-35.
17. EFSA. 2007. Opinion of the scientific panel on biological hazards on the request
from the commission for review of the opinion on microbiological risks in infant
Prevalence of Cronobacter spp.

103
formulae and follow-on formulae with regard to Enterobacteriaceae as indicators.
The EFSA Journal. 444:1-14.
18. FAO/WHO. 2004. Enterobacter sakazakii and other Micro-organisms in
powdered infant formula: meeting report. Microbiological Risk Assessment Series
No. 6, Rome, Italy.
19. FAO/WHO. 2006. Enterobacter sakazakii and Salmonella in powdered infant
formula: meeting report. Microbiological Risk Assessment Series No. 10 Rome,
Italy.
20. FAO/WHO. 2008. Enterobacter sakazakii (Cronobacter spp.) in powdered follow-
up formulae: meeting report. Microbiological Risk Assessment Series No. 15
Rome, Italy.
21. Farmer III, J. J., M. A. Asbury, F. W. Hickman, D. J. Brenner, and the
Enterobacteriaceae study group. 1980. Enterobacter sakazakii: A new species of
"Enterobacteriaceae" isolated from clinical specimens. Int. J. Syst. Bacteriol.
30:569-584.
22. Friedemann, M. 2007. Enterobacter sakazakii in food and beverages (other than
infant formula and milk powder). Int. J. Food Microbiol. 116:1-10.
23. Friedemann, M. 2009. Epidemiology of invasive neonatal Cronobacter
(Enterobacter sakazakii) infections. Eur. J. Clin. Microbiol. Infect. Dis. 28:1297–
1304.
24. Gakuya, F. M., M. N. Kyule, P. B. Gathura, and S. Kariuki. 2001. Antimicrobial
resistance of bacterial organisms isolated from rats. East Afr. Med. J. 78:646-649.
25. Guillaume-Gentil, O., V. Sonnard, M. C. Kandhai, J. D. Marugg, and H. Joosten.
2005. A simple an rapid cultural method for detection of Enterobacter sakazakii in
environmental samples. J. Food Prot. 68:64-69.
26. Gurtler, J. B., J. L. Kornacki, and L. R. Beuchat. 2005. Enterobacter sakazakii: A
coliform of increased concern to infant health. Int. J. Food Microbiol. 104:1-34.
27. Hamilton, J. V., M. J. Lehane, and H. R. Braig. 2003. Isolation of Enterobacter
sakazakii from midgut of Stomoxys calcitrans. Emerg. Infect. Dis. 9:1355-1356.
28. Heuvelink, A. E., M. Ahmed, F. D. Kodde, J. T. M. Zwartkruis-Nahuis, and E. de
Boer. 2002. Enterobacter sakazakii in melkpoeder. De Ware(n)chemicus. 32:17-
30.
29. Himelright, I., E. Harris, V. Lorch, and M. Anderson. 2002. Enterobacter
sakazakii infections associated with the use of powdered infant formula-
Tennessee, 2001. JAMA. 287:2204-2205.
30. ICMSF. 2002. Micro-organisms in Foods 7: Microbiological Testing in Food
Safety Management. Springer Publishers, New York.
Chapter 5
104
31. ISO. 2004. ISO 21528-1 Microbiology of food and animal feeding stuffs -
Horizontal methods for the detection and enumeration of Enterobacteriaceae -
Part 1: Detection and enumeration by MPN technique with pre-enrichment.
32. ISO. 2005. ISO 21528-2 Microbiology of food and animal feeding stuffs -
Horizontal methods for the detection and enumeration of Enterobacteriaceae -
Part 2: Colony-count method.
33. ISO. 2006. ISO/TS 22964 and IDF/RM 210-first edition 2006-02-01: Milk and
milk product- Detection of Enterobacter sakazakii.
34. Iversen, C., P. Druggan, and S. J. Forsythe. 2004. A selective differential medium
for Enterobacter sakazakii, a preliminary study. Int. J. Food Microbiol. 96:133-
139.
35. Iversen, C., P. Druggan, S. Schumacher, A. Lehner, C. Feer, K. Gschwend, H.
Joosten, and R. Stephan. 2008. Development of a novel screening method for the
isolation of "Cronobacter" spp. (Enterobacter sakazakii). Appl. Environ.
Microbiol. 74:2550-2553.
36. Iversen, C., and S. J. Forsythe. 2004. Isolation of Enterobacter sakazakii and other
Enterobacteriaceae from powdered infant formula milk and related products. Food
Microbiol. 21:771-777.
37. Iversen, C., and S. J. Forsythe. 2007. Comparison of media for the isolation of
Enterobacter sakazakii. Appl. Environ. Microbiol. 73:48-52.
38. Iversen, C., M. Lane, and S. J. Forsythe. 2004. The growth profile,
thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant
formula milk. Lett. Appl. Microbiol. 38:378-382.
39. Iversen, C., A. Lehner, N. Mullane, E. Bidlas, I. Cleenwerck, J. Marugg, S.
Fanning, R. Stephan, and H. Joosten. 2007. The taxonomy of Enterobacter
sakazakii: proposal of a new genus Cronobacter gen. nov and descriptions of
Cronobacter sakazakii comb. nov Cronobacter sakazakii subsp sakazakii, comb.
nov., Cronobacter sakazakii subsp malonaticus subsp nov., Cronobacter turicensis
sp nov., Cronobacter muytjensii sp nov., Cronobacter dublinensis sp nov and
Cronobacter genomospecies I. BMC Evol. Biol. 7:64.
40. Iversen, C., N. Mullane, B. McCardel, B. D. Tal, A. Lehner, S. Fanning, R.
Stephan, and H. Joosten. 2008. Cronobacter gen. nov., a new genus to
accommodate the biogroups of Enterobacter sakazakii, and proposal of
Cronobacter sakazakii gen. nov., comb. nov., Cronobacter malonaticus sp. nov.,
Cronobacter turicensis sp. nov., Cronobacter muytjensii sp. nov., Cronobacter
dublinensis sp. nov., Cronobacter genomospecies 1, and of three subspecies,
Cronobacter dublinensis subsp. Dublinensis subsp. nov., Cronobacter dublinensis
subsp. lausannensis subsp. nov. and Cronobacter dublinensis subsp. lactaridi
subsp. nov. . Int. J. Syst. Evol. Microbiol. 58:1442-1447.
Prevalence of Cronobacter spp.

105
41. Joosten, H., J. Marugg, R. Stephan, A. Klijn, T. Jackson, and C. Iversen. 2008. A
rapid and reliable alternative to ISO 21528-1 : 2004 for detection of
Enterobacteriaceae. Int. J. Food Microbiol. 125:344-346.
42. Jung, M. K., and J. H. Park. 2006. Prevalence and thermal stability of
Enterobacter sakazakii from unprocessed ready-to-eat agricultural products and
powdered infant formulas. Food Sci. Biotechnol. 15:152-157.
43. Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004. Occurrence of Enterobacter sakazakii in the food production
environments and households. The Lancet. 363:39-40.
44. Kandhai, M. C., M. W. Reij, C. Grognou, M. Van Schothorst, L. G. M. Gorris, and
M. H. Zwietering. 2006. Effects of preculturing conditions on lag time and
specific growth rate of Enterobacter sakazakii in reconstituted powdered infant
formula. Appl. Environ. Microbiol. 72:2721-2729.
45. Kandhai, M. C., M. W. Reij, K. Van Puyvelde, O. Guillaume-Gentil, R. R.
Beumer, and M. Van Schothorst. 2004. A new protocol for the detection of
Enterobacter sakazakii applied to environmental samples. J. Food Prot. 67:1267-
1270.
46. Kang, E. Sil, Y. S. Nam, and K. W. Hong. 2007. Rapid detection of Enterobacter
sakazakii using TaqMan real-time PCR assay. J. Microbiol. Biotechnol. 17:516-
519.
47. Kilonzo-Nthenge, A., F. C. Chen, and S. L. Godwin. 2008. Occurrence of Listeria
and Enterobacteriaceae in domestic refrigerators. J. Food Prot. 71:608-612.
48. Kim, K., S. S. Jang, S. K. Kim, J. H. Park, S. Heu, and S. Ryu. 2008. Prevalence
and genetic diversity of Enterobacter sakazakii in ingredients of infant foods. Int.
J. Food Microbiol. 122:196-203.
49. Kuzina, L. V., J. J. Peloquin, D. C. Vacek, and T. A. Miler. 2001. Isolation and
identification of bacteria associated with adult laboratory Mexican fruit flies,
Anastrepha ludens (Diptera : Tephritidae). Curr. Microbiol. 42:290-294.
50. Lai, K. K. 2001. Enterobacter sakazakii infections among neonates, infants,
children, and adults - Case reports and a review of the literature. Medicine.
80:113-122.
51. Lehner, A., S. Nitzsche, P. Breeuwer, B. Diep, K. Thelen, and R. Stephan. 2006.
Comparison of two chromogenic media and evaluation of two molecular based
identification systems for Enterobacter sakazakii detection. BMC Microbiol. 6:15.
52. Leuschner, R. G. K., F. Baird, B. Donald, and L. J. Cox. 2006. A medium for the
presumptive detection of Enterobacter sakazakii in infant formula (vol 21, pg 527,
2004). Food Microbiol. 23:409-409.
Chapter 5
106
53. Mramba, F., A. Broce, and L. Zurek. 2006. Isolation of Enterobacter sakazakii
from stable flies, Stomoxys calcitrans L. (Diptera : Muscidae). J. Food Prot.
69:671-673.
54. Mullane, N., B. Healy, J. Meade, P. Whyte, P. G. Wall, and S. Fanning. 2008.
Dissemination of Cronobacter spp. (Enterobacter sakazakii) in a powdered milk
protein manufacturing facility. Appl. Environ. Microbiol. 74:5913-5917.
55. Mullane, N., C. Iversen, B. Healy, C. Walsh, P. Whyte, P. G. Wall, T. Quinn, and
S. Fanning. 2007. Enterobacter sakazakii: an emerging bacterial pathogen with
implications for infant health. Minerva Pediatr. 59:137-148.
56. Mullane, N. R., D. Drudy, P. Whyte, M. O'Mahony, A. G. M. Scannell, P. G.
Wall, and S. Fanning. 2006. Enterobacter sakazakii: biological properties and
significance in dried infant milk formula (IMF) powder. Int. J. Dairy Technol.
59:102-111.
57. Mullane, N. R., J. Murray, D. Drudy, N. Prentice, P. Whyte, P. G. Wall, A. Parton,
and S. Fanning. 2006. Detection of Enterobacter sakazakii in dried infant milk
formula by cationic-magnetic-bead capture. Appl. Environ. Microbiol. 72:6325-
6330.
58. Mullane, N. R., P. Whyte, P. G. Wall, T. Quinn, and S. Fanning. 2007.
Application of pulsed-field gel electrophoresis to characterise and trace the
prevalence of Enterobacter sakazakii in an infant formula processing facility. Int.
J. Food Microbiol. 116:73-81.
59. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
60. Muytjens, H. L., J. Van der Ros-van de Repe, and H. A. Van Druten. 1984.
Enzymatic profiles of Enterobacter sakazakii and related species with special
reference to the alpha-glucosidase reaction and reproducibility of the test system.
J. Clin. Microbiol. 20:684-686.
61. Muytjens, H. L., H. C. Zanen, H. J. Sonderkamp, L. A. Kollee, I. K. Wachsmuth,
and J. J. Farmer, 3
rd
. 1983. Analysis of eight cases of neonatal meningitis and
sepsis due to Enterobacter sakazakii. J. Clin. Microbiol. 18:115-120.
62. Nazarowec-White, M., and J. M. Farber. 1997. Incidence, survival, and growth of
Enterobacter sakazakii in infant formula. J. Food Prot. 60:226-230.
63. Noriega, F. R., Kotloff, K.L., Martin, M.A., Schwalbe, R.S. 1990. Nosocomial
bacteremia caused by Enterobacter sakazakii and Leuconostoc mesenteroides
resulting from extrinsic contamination of infant formula. Pediatr. Infect. Dis. J.
9:447-449.
64. O'Brien, S., B. Healy, C. Negredo, S. Fanning, and C. Iversen. 2009. Evaluation of
a new one step enrichment in conjunction with a chromogenic medium for the
Prevalence of Cronobacter spp.

107
detection of Cronobacter spp. (Enterobacter sakazakii) in powdered infant
formula. J. Food Prot. 72:1472 -1475.
65. Oh, S. W., H. J. Chung, and D. H. Kang. 2006. Comparison of enrichment broths
for detection and isolation of Enterobacter sakazakii from infant formula milk. J.
Rapid Meth. Autom. Microbiol. 14:325-336.
66. Raghav, M., and P. K. Aggarwal. 2007. Isolation and characterization of
Enterobacter sakazakii from milk foods and environment. Milchw. Milk Sci. Int.
62:266-269.
67. Reij, M. W., and M. H. Zwietering. 2008. Integrating concepts: a case study using
Enterobacter sakazakii in infant formula. p. 177 - 204. In: D.W. Schaffner (ed.),
Microbial Risk Analysis of Foods ASM Press, Washington DC, New Brunswick,
New Jersey.
68. Restaino, L., E. W. Frampton, W. C. Lionberg, and R. J. Becker. 2006. A
chromogenic plating medium for the isolation and identification of Enterobacter
sakazakii from foods, food ingredients, and environmental sources. J. Food Prot.
69:315-322.
69. Schmid, M., C. Iversen, I. Gontia, R. Stephan, A. Hofmann, A. Hartmann, B. Jha,
L. Eberl, K. Riedel, and A. Lehner. 2009. Evidence for a plant-associated natural
habitat for Cronobacter spp.. Res. Microbiol. 160:608-614.
70. Seo, K. H., and R. E. Brackett. 2005. Rapid, specific detection of Enterobacter
sakazakii in infant formula using a real-time PCR assay. J. Food Prot. 68:59-63.
71. Shaker, R., T. Osaili, W. Al-Omary, Z. Jaradat, and M. Al-Zuby. 2007. Isolation
of Enterobacter sakazakii and other Enterobacter spp. from food and food
production environments. Food Control. 18:1241-1245.
72. Simmons, B. P., M. S. Gelfand, M. Haas, L. Metts, and J. Ferguson. 1989.
Enterobacter sakazakii infections in neonates associated with intrinsic
contamination of a powdered infant formula. Infect. Control Hosp. Epidemiol.
10:398-401.
73. Urmenyi, A. M. C., and M. D. Berne. 1961. Neonatal death from pigmented
coliform infection. The Lancet. 1:313-315.
74. Yoo, M. K., S. S. Kim, and S. Oh. 2005. Isolation and genotyping of Enterobacter
sakazakii from powdered infant formula manufactured in Korea. J. Food Sci.
Biotechnol. 14:875-877.


Chapter 5
108
Lag time and specific growth rate

109
Chapter 6
110
Abstr act
Cronobacter spp. can be present, although in low levels, in dry powdered infant formulae
and it has been linked to cases of meningitis in neonates, especially those born
prematurely. In order to prevent illness, product contamination at manufacture, during
preparation, and also growth after reconstitution has to be minimised using appropriate
control measures. In this publication several determinants of the growth of Cronobacter
spp. in reconstituted infant formula are reported. The following key growth parameters
were determined: lag time, specific growth rate and maximum population density. Cells
were harvested at different phases of growth, and spiked into powdered infant formula.
After reconstitution in sterile water, Cronobacter spp. was able to grow at temperatures
between 8 and 47 ºC. The estimated optimal growth temperature was 39.4 ºC, whereas the
optimal specific growth rate was 2.31 h
-1
. The effect of temperature on the specific growth
rate was described using two secondary growth models. The resulting minimum – and
maximum temperature estimated with the Secondary Rosso equation were 3.6 ºC and 47.6
ºC, respectively. The estimated lag time varied from 83.3 ± 18.7 h at 10 ºC to 1.73 ± 0.43 h
at 37 ºC and could be described with the hyperbolic model and reciprocal square root
relation. Cells harvested at different phases of growth did not exhibit any significant
difference in either specific growth rate or lag time. Strains did not have different lag times
and lag times were short taking into account that the cells had spent several (3-10) days in
dry powdered infant formula. The growth rates and lag times at various temperatures
obtained in this study may help in calculation of the period for which reconstituted infant
formula can be stored at a specific temperature without detrimental impact on health.



This Chapter has been published as
“Effect of pr e-cultur ing conditions on lag time and specific gr owth r ate of
Enterobacter sakazakii
*
in r econstituted powder ed infant for mula”

Kandhai, M.C., M.W. Reij, C. Grognou, M. Van Schothorst, L.G.M. Gorris, and M.H.
Zwietering.
Applied and Environmental Microbiology, 2006; 72:4, 2721-2729.

* Throughout this thesis the designation Cronobacter spp. has been used to indicate the
taxonomic change.
Lag time and specific growth rate

111
Intr oduction
Cronobacter spp. is a motile peritrichous, Gram-negative rod that occasionally causes
neonatal meningitis and sepsis, with mortality rates of 40 to 80% (4). The recovery of
Cronobacter spp. from samples of commercially available dry powdered infant formulae
has been reported (5, 9, 10). Cronobacter spp. in infant formulae have been associated with
outbreaks of meningitis, sepsis and necrotizing enterocolitis in prematures and full term
infants, particularly those with predisposing medical conditions (3). Although the levels of
Cronobacter spp. occurring in dry powdered infant formulae are generally very low,
reconstituted infant formula is a good medium for growth. When present in dry formula
Cronobacter spp. may grow during preparation, cooling, storage and holding of the bottles,
increasing the probability of illness. Occasional contamination of the dried infant formula
during manufacture is a source of the micro-organism occurring in reconstituted product.
However, as Cronobacter spp. has been detected in various other dry environments (8),
contamination may also occur during reconstitution of the dried infant formula in hospitals
or at home.
In order to prevent illness, product contamination at manufacture and/or during preparation
and growth after reconstitution has to be minimised using appropriate control measures.
Mathematical models are useful tools to evaluate the effectiveness of control measures.
Depending on the source and the history of contaminating bacterial cells, which influence
their physiological state, and the suitability of the product to sustain their growth, i.e. the
product’s (intrinsic) conditions and the environmental (extrinsic) conditions, microbial cells
will either start to grow immediately or show a distinct phase of no apparent growth (the
lag phase). In the case of Cronobacter spp. cells, reconstituted infant formulae offer rich
growth environments that allow immediate proliferation provided that the cells are in a
sound physiological state, that the external conditions (mainly temperature) are favourable
and that there is sufficient time for growth. Should lag times be apparent before growth,
then this may be a result of an injured state of the Cronobacter spp. cells from which they
may gradually recover, as is evidenced by the start of cell proliferation (17). Baranyi &
Roberts (1) emphasized that the lag time is a period of adjustments to the new environment
during which only intracellular conditions change. Growth models can simulate growth
Chapter 6
112
after reconstitution and the effect of key intrinsic or extrinsic conditions can be determined.
To develop growth models, insight into parameters describing growth of the micro-
organism, such as the lag time and specific growth rate, is required.
This study describes the effects of a number of pre-culturing conditions on key growth
parameters for Cronobacter spp. growing in reconstituted (with sterile water) powdered
infant formula. Furthermore, the effects of temperature on specific growth rate and lag time
were quantified and compared with literature values. Viable counts were used to construct
growth curves that were used to derive the key growth parameters by curve fitting using the
Modified Gompertz equation as primary growth model (20). During the secondary
modelling step the Square root Ratkowsky model (13) and the Secondary Rosso model (14)
were fitted to the estimates of the specific growth rates at various temperatures. Likewise,
the lag time data were fitted, using the logarithm of the inverse of the Ratkowsky model
and the hyperbolic equation (19). The resulting parameters permit useful predictions of the
growth of Cronobacter spp. in reconstituted infant formula and aid in the design of
effective control measures to reduce exposure of susceptible consumers in both hospital and
household settings.

Mater ials and methods
Organisms
Four Cronobacter isolates were used in this study. Strain MM9 was isolated from milk
powder and strain MC10 was isolated from a patient. Both isolates were obtained from Dr.
H.L. Muytjens, University Medical Centre (UMC) St. Radboud, Nijmegen, The
Netherlands. Strain 94 (S94) was isolated from a sample of a vacuum cleaner bag obtained
from a domestic environment (8). Cronobacter sakazakii ATCC 29544 (4) was the
reference strain. Stock cultures were maintained at –20 ºC in cryo vials (Greiner Bio-one
GmbH, Frickenhausen, Germany), containing 0.7 ml of a stationary-phase culture
suspension in Brain Heart Infusion (BHI; Becton Dickinson and Co., 38800 Le Pont de
Claix, France) broth with 0.3 ml 87% glycerol (Fluka-Chemica, GmbH CH9471 Buchs,
Lag time and specific growth rate

113
Switzerland). Characterization of these isolates together with 70 other Cronobacter isolates,
using the pulsed field electrophoresis technique (15) revealed a large variability in
molecular fingerprints. Specific groups could not be distinguished from the dendrograms
(results not shown).

Preparation bacterial suspension
Strain ATCC 29544 was cultured by transferring 250 µl of the stock culture into 250 ml
BHI followed by incubation at 37 ºC. Cells were incubated for 3, 6, 16, 24 and 72 hours to
obtain respectively the exponential phase, early stationary phase, middle stationary phase,
stationary phase and late stationary phase cells. Strains MC10, MM9 and S94 were
incubated at 37 ºC for 18 hours to obtain middle stationary cells. In order to obtain enough
cells in the lag phase, 1 ml of a stock culture of ATCC 29544 from the –20 ºC was diluted
in 2 ml glycerol, whereupon the suspension was centrifuged (Hermle Z 231M, B. Hermle
GmbH & Co., D-7209 Gosheim, Germany) for 10 minutes at 10,000 x g. Cells harvested
were transferred into 30 ml BHI and incubated for 1.5 hour at 37 ºC.
In order to obtain cells for spiking dry infant formula, these BHI grown cultures were
centrifuged for 5 minutes at 20 °C at 2958 x g (MSE, Mistral 3000i, Leicester, United
Kingdom). Cells were washed twice in 1% physiological salt solution and subsequently
suspended in 30 ml 1% physiological salt solution for the lag phase cells or 250 ml 1%
physiological salt solution, resulting in a cell suspension of about 10
4
CFU/ml for lag phase
grown cells and between 10
5
and 10
7
CFU/ml for the other growth phases.

Spiking of the infant formula
Infant formula was bought in local shops; the numbers of the viable bacteria in the infant
formula were sufficiently low to prevent them from influencing the growth of the spiked
cells (data not shown). The obtained bacterial suspension was sprayed on commercial dry
infant formula (1:50 (w/w)), using a perfume sprayer (Designed by Gérard Brinard, DA
Drogisterij, Leusden, The Netherlands). The final bacterial concentration 3-4 days after
Chapter 6
114
spiking was 10
2
-10
4
CFU/g of dry powder and the infant formula maintained a water
activity of a
w
< 0.22. All growth experiments were performed within 10 days after spiking
the infant formula.

Design
The initial estimates of the lowest temperature supporting growth (T
min
), highest
temperature supporting growth (T
max
) and the specific growth rate (µ
opt
) at optimal growth
temperature (T
opt
), shown in Table 1, were used to predict the specific growth rates at
various temperatures. The lag time at each temperature was initially estimated by the k
value, which is the product of the specific growth rate (µ
m
) and lag time (ì) and which is
known to have a large variability, but can be considered constant over a wide range of
temperatures (18).

Table 1. Parameters for growth of Cronobacter spp. in reconstituted infant formula: initial estimates
based on published data
Par ameter Initial estimate Refer ence (s)
T
min
5.5

ºC (10)
T
max
47 ºC (4, 6)
T
opt
37 ºC (6)
µ
opt
2.5 h
-1
(6)
k 3.75 (7)

In order to determine appropriate sampling times and dilutions for plate counting, the
number of micro-organisms in each sample and at every sampling time was roughly
predicted using the exponential growth function (N
(t)
= N
(0)
e

m
*t)
and the secondary Rosso
equation (Equation 3), using the design values as shown in Table 1. Here N
t
is the number
of micro-organisms [CFU/ml] at time t [h], N
0
is the number of micro-organisms at time of
inoculation, µ
m
is the specific growth rate [h
-1
], t is the time [h].
Lag time and specific growth rate

115
Growth experiments
To prepare samples for growth experiments, ten-gram portions of spiked infant formula
were reconstituted in 100 ml sterilised tap water. In the experiments to determine the effect
of various growth phases, bottles with strain ATCC 29544 were incubated as follows (the
number of bottles incubated is given in parentheses: 10 ºC (n=9); 21 ºC (n=11); 29 ºC
(n=14); 37 ºC (n=14). Middle stationary phase grown cells of strain ATCC 29544 were
additionally used to assess (in duplicate) growth in reconstituted infant formula at the
following temperatures (ºC): 8, 14, 38, 39, 41, 43, 45, 46, 47, 48, 49, 50. Middle stationary
phase grown cells of strains MM9, MC10 and S94 were incubated only at 29 and 37 ºC.
Growth of Cronobacter spp. was measured at various time intervals, depending on the
temperature of incubation, appropriate dilutions were made in Peptone Saline Solution
(NaCl 8.5 g/L and neutralised Bacteriological Peptone 1g/L from Oxoid, Basingstoke,
England). Samples were surface plated onto tryptone soy agar (TSA; Ltd., Basingstoke,
Hampshire, England) with a spiral plater (Eddy Jet, IUL instruments, I.K.S. bv., Leerdam,
The Netherlands). Inoculated plates were incubated for 20 to 24 h at 37 ºC before manual
counting.

Data analysis
For describing the evolution of the microbial count with time, a primary model was used.
The three kinetic parameters, namely lag time, specific growth rate and maximum
population density were estimated by fitting with the Modified Gompertz equation
(Equation 1). This resulted in estimates for the lag time (ì; [ h]), specific growth rate (µ
m
; [
h
-1
]) and the asymptotic value (A; [-]) (20) at the tested temperature for each growth curve.

( )
)
`
¹
¹
´
¦

|
|
.
|

\
|
+ ÷ ÷ = 1 exp exp
0
t
Ln t
A
e
m
A
N
N
ì
u
(1)
Chapter 6
116
In order to obtain reliable estimates for the growth parameters, experimental growth curves
had to meet the following requirements:
1) at least 2 data points should fall within the lag-time, unless the lag time was
shorter than 2 hours,
2) for data points within the exponential phase, there should be at least 3 data points
over a range of 3-hours and 3-logs,
3) 3 data points should be in the stationary phase at least 1-hour apart.
Growth curves that did not meet these requirements were excluded.

A Bélehrádek type model (Equation 2), also known as the (expanded) square root model of
Ratkowsky (13), was used to describe the relation between the specific growth rate and the
temperature. This model contains four parameters of which two are easily interpretable,
T
min
and T
max
.

If
max min
T T T < < (2)
then
( ) ( ) ( ) | | { } ( )
2
max
exp 1
min
T T c T T b T
m
÷ ÷ ÷ = u

and if
max
min
T T
T T
>
s
then 0 =
m
u
where T
min
is the extrapolated minimum temperature [ºC] at which the specific growth rate (µ
m
;[h
-1
])
is zero, T
max
is the extrapolated maximum temperature at which µ
m
= 0, and b [ºC
-1
h
-0.5
], and c [ºC
-1
]
are so-called Ratkowsky parameters

(13).

The secondary growth model of Rosso et al., (14) (Equation 3), was used as well to
describe the effect of temperature on growth rate. This model contains all four interpretable
parameters µ
opt,
T
min,
T
max
and T
opt
.
Lag time and specific growth rate

117
If
max min
T T T < < (3)
then
( )
( )( )
( )( ) ( )( ) | |( )
opt
min opt
2
min opt max opt opt min opt
2
min max
u u
|
|
.
|

\
|
÷ ÷ + ÷ ÷ ÷ ÷
÷ ÷
=
T T T T T T T T T T T
T T T T
T
m


and if
max
min
T T
T T
>
s
then 0 =
m
u
where T
min
and T
max
are defined in equation 2, T
opt
[ºC] is the temperature at which the specific growth
rate µ
m
[h
-1
] is optimal, and µ
opt
is the µ
m
at optimal temperature.

The logarithm of the inverse of the secondary Ratkowsky model (Equation 4) and the
hyperbolic equation (Equation 5) were used (19) to describe the lag time temperature
relation.

( ) ( ) ( ) ( ) | | { } ( ) | |
2
max
exp 1
min
ln ln
÷
÷ ÷ = T T c T-T b T ì (4)

where b [ºC
-1
h
-0.5
], and c [ºC
-1
] are the so-called Ratkowsky parameters. The T
min
and T
max
values
were assumed to be equal to the T
min
and T
max
of equation 3 (Secondary Rosso growth model),
describing the specific growth rate.

( )
q T
p
÷
= ì ln (5)

where p is a measure for the decrease of the lag time when the temperature is increased and q is the
temperature at which the lag time in infinite (no growth). The q value is comparable to T
min
.
Chapter 6
118
Statistical analysis
In order to determine whether pre-culturing conditions have a significant effect on lag times
and/ or specific growth rates, the data obtained with strain ATCC 29544 at 10, 21, 29 and
37 °C were subjected to univariate analysis of variance. Lag time data were log transformed
and specific growth rate data were square root transformed in order to obtain homogeneity
of variance. A significance level of 5% was used. All data analyses were performed using
SPSS (SPSS release 11.5 for Microsoft Windows 95/98/NT/2000; SPSS Inc., Chicago,
U.S.A). Fitting was done by minimizing the residual sum of squares (RSS) with both the
solver function in Excel and Table curve 2D, windows 2.03, for verification. Standard
deviations were calculated with Excel and are reported as plus or minus the mean.

Results
Determination of growth parameters, using pre-cultured middle stationary
ATCC 29544 cells
In order to design the experiments optimally and to obtain growth curves meeting the
specified requirements, the course of each individual growth curve of Cronobacter spp. at
10, 21, 29, and 37 °C was predicted based on published growth parameters (see Table 1, for
initial estimates). Examples of the design growth curve at 21 ºC and 37 ºC are shown in
Figures 1A and 1B, respectively.
These figures show, furthermore, the resulting experimental count data of growth in
reconstituted, contaminated infant formula. The modified Gompertz equation (Equation 1)
was fitted to the observed number of micro-organisms in time, resulting in estimates for the
lag time, specific growth rate, asymptote and the initial number of organisms (19). For the
growth curves at 10 ºC and 29 ºC comparable results were obtained (data not shown).
Variations in the initial number of micro-organisms at time zero may have been due to a
gradual decline of the cell number in the dry infant formula.
Lag time and specific growth rate

119






Figur e 1. Predicted and measured growth curves at 21 ºC (A) and 37 ºC (B) of Cronobacter
sakazakii ATCC 29544, pre-cultured to middle stationary state. The different symbols indicate
different replicate experiments. The straight line represents the design growth curve at 21 ºC, using
the cardinal values as shown in Table 1 under “initial estimate”; Dotted lines represent fits of the
Modified Gompertz equation to each single experiment.
Chapter 6
120
Effects of the physiological gr owth phase on µ
m
and ì
Six different pre-cultures of the reference strain, ATCC 29544 were prepared to yield a
variety of physiological growth phases at the moment of spiking of the powdered infant
formula. After reconstitution with sterile tap water, all six types were incubated with
various replications at 10, 21, 29 and 37 ºC to represent a temperature range relevant for
reconstituted infant formulae. Growth was observed in all cases and at every temperature.
Temperature had a marked effect on both the specific growth rate (μ
m
) and the lag phase (λ)
(Figures 2A and 2B). With increasing temperature, μ
m
strongly increased. Values for
specific growth rates varied from 0.12 ± 0.04 h
-1
at 10 ºC to 2.29 ± 0.45 h
-1
at 37 ºC, and no
apparent effect of the physiological state was found. In Figure 2B it is shown that the lag
time decreased with increasing temperature and that there was also no apparent effect of the
various physiological growth phases on the lag time.


Figur e 2A. Square root of specific growth rate data for various physiological growth states of strain
ATCC 29544 as function of temperature; × ODJ SKDVHJURZQFHOOV exponential phase cells; ^:
early stationary phase cells; -: middle stationary phase cells; o: stationary phase cells; : late
stationary phase cells.
Lag time and specific growth rate

121

Figur e 2B. Log lag time for various physiological growth states of ATCC 29544 as a function of the
temperature; ×: lag phase cells; △: exponential phase cells; ^: early stationary phase cells; -: middle
stationary phase cells; o: stationary phase cells; : late stationary phase cells.

Estimates of the specific growth rates and lag times for the various pre-culturing conditions
(lag phase, exponential phase, early stationary, middle stationary phase and late stationary
phase) of strain ATCC 29544 as shown in Figures 2A and 2B were analysed with the
univariate analysis of variance test. As depicted in Table 2, this resulted in p-values > 0.05,
except for the lag time at 21 ºC (p=0.048). However, this value can be considered as
borderline significance. This test corroborates the visual observation (Figures 2A and 2B)
that the cell history had no significant effect on either the specific growth rate or on the lag
time during subsequent cultivation in reconstituted infant formula. Furthermore, statistical
analysis showed that the k-value (product of λ and μ
m
) was also not significantly influenced
by the cell history (Table 2).



Chapter 6
122
Effects of the strain variability on µ
m
and ì
The effect of strain variability on the growth parameters was studied by growing three other
strains MC10, MM9 and S94, pre-cultured to middle stationary phase, at 29 and 37 °C. It is
shown in Figures 3 and 4 that neither the lag time nor the maximum specific growth rate
was significantly different for strains of different origin.

Table 2. Statistical evaluation (p-values) of the univariate analysis of variance for the effects of the
physiological growth phase of Cronobacter sakazakii ATCC 29544 cells on lag time, specific growth
rate and the product of both (k) at 10, 21, 29 and 37 °C.Bold value indicates p-value < 0.05 meaning a
significant effect of the physiological state on the variable
Value at:
Par ameter

10 °C 21 °C 29 °C 37 °C
Lag time (ì) [h] 0.092 0.048 0.629 0.254
Specific growth
rate (µ
m
) [h
-1
]
0.461 0.295 0.314 0.166
k-value (ì x u
m
)
[-]
0.407 0.408 0.539 0.512

Estimations of growth parameters
Since statistical analysis showed that the physiological growth phases of strain ATCC
29544 at 10, 21 29 and 37 °C did not significantly influence the specific growth rates and
lag times, further experiments were performed over a wide range of temperatures, from 8
up to 50 ºC with middle stationary pre-cultured cells of strain ATCC 29544 only. Although
growth was observed up to 47 ºC, no reliable estimates of the specific growth rate and lag
time could be derived from the models at that temperature as the growth curves did not
meet the requirements as formulated in the materials and method section. All of the
estimated specific growth rates were combined and modelled with equation 2 and equation
3 as shown in Figure 3, where the square root of the specific growth rate (õ
m
) is plotted as
function of the temperature. The secondary growth parameters derived from optimal fits are
shown in Table 3 and 4.
Lag time and specific growth rate

123


Table 3. Parameter values for the effect of the temperature on the specific growth rate (µ
m
) for
Cronobacter spp. in reconstituted infant formula resulting from fits by the Ratkowsky Secondary
growth model (Equation 2)
Par ameter

Estimate

95% Confidence
inter val
Dimensions

T
min
2.19 -0.26 - 4.64
ºC
T
max
48.9 47.9 - 49.9
ºC
b 0.047 0.0407 - 0.0521
ºC
-1
h
-0.5

c 0.239 0.144 - 0.335
ºC
-1

RSS 1.28

Table 4. Parameter values for the effect of the temperature on the specific growth rate (µ
m
) for
Cronobacter spp. in reconstituted infant formula resulting from fits by the Rosso secondary growth
(Equation 3)
Par ameter

Estimate

95% Confidence
inter val
Dimensions

T
min
3.60 1.42 - 5.79
ºC
T
max
47.6 47.1 - 48.2
ºC
T
opt
39.4 38.1- 40.7
ºC
µ
opt
2.31 2.13 - 2.48
h
-1

RSS 1.31

From Figure 3 it is apparent that the differences between the fits of these models are
smaller than the experimental variability. The residual sum of squares (RSS) for the
secondary Rosso model (Equation 3) was 1.31 and for the square root Ratkowsky model
(Equation 2) was 1.28. Although the RSS was slightly higher, the secondary Rosso model
was used for further evaluation since it consists of four parameters that all have a biological
meaning and can be interpreted as such. Transforming the square root μ
m
data from Figure 3
back to μ
m
, it can be seen that the specific growth rate of Cronobacter spp. varied from
0.115 h
-1
(\ μ
m
=0.339 h
-0.5
) at 10
o
C to 1.113 h
-1
(\ μ
m
= 1.063 h
-0.5
) at 46
o
C, and 2.242 h
-1

(\ μ
m
= 1.498 h
-0.5
) at 37 ºC. For comparison the experimental specific growth data
Chapter 6
124
measured by Iversen et al., (6) and Nazarowec-White et al., (10) have been transformed and
included in Figure 3.

Figur e 3. Square root of measured and fitted specific growth rates as function of the temperature. ○:
growth rates of ATCC 29544 pre-cultures to various growth states as estimated by the fit of the
Modified Gompertz model to each individual growth curve. Also for middle stationary phase grown
cells of ● = MM9, ♦= S94 and ▲= MC10. +: growth rates published by Nazarowec-White et al., (9);
◊: growth rates published by Iversen et al., (7); fits by secondary growth model of - - - - Ratkowsky
and

: Rosso (fitted to square root transformed data of the current study only).

The estimated lag times as resulting from the fit by the modified Gompertz model were
plotted against the temperature and presented in Figure 4. The longest lag time observed
was 83.3 ± 18.7 h at temperatures around 10 ºC. The temperature at which the lag time was
minimal was between 37 and 39 ºC, with an estimated minimal lag time of 1.73 ± 0.44 h.
The effect of the temperature on the lag time was described by fitting the reciprocal square
root relation (Equation 4) and the hyperbolic model (Equation 5) to the logarithmic
transformation of the data. It is assumed that the T
min
and T
max
value are equal to the T
min

and T
max
value of Equation 3 describing the specific growth rate. Parameters estimated with
both models are given in Table 5.
Lag time and specific growth rate

125
Table 5. Parameter values for the effects of temperature on the lag time (ì) for Cronobacter spp. in
reconstituted infant formula
Model Par ameter Estimate
95% Confidence
inter val
T
min
3.60 -
a

Inverse Ratkowsky
(Equation 4) T
max
47.6 -
a

b 0.023 0.021 - 0.024
c 0.645 0.275 - 1.176
RSS 2.29
p 25.8 22.3 - 29.3
Hyperbolic
(Equation 5) q 3.75 2.88 - 4.62
RSS 2.11

a
– fixed (from Table 4) by using the values estimated with the secondary Rosso model for growth
(Equation 3).







Figur e 4. Log lag time as function of temperature. ○ = ATCC 29544 pre-cultures to various growth
states. Also for middle stationary phase grown cells of ● = MM9, ♦ = S94 and ▲= MC10. + indicates
the lag times at 10 and 23
o
C as published by Nazarowec-White et al.,(10). Logarithmic transformed
lag time data of current study, modelled with the hyperbola model (----------) and reciprocal
Ratkowsky model (
_________
).
Chapter 6
126
The logarithmic transformed lag times as fitted by the reciprocal square root relation and
the hyperbolic model are represented in Figure 4. From this graph it can be concluded that
both models fits the data reasonably well. The residual sum of squares (RSS) for the
hyperbolic model was 2.11 and for the reciprocal square root relation 2.29. In Figure 5, the
parameter k, the product of the lag time and the specific growth rate (µ
m
x ì), as function of
temperature is shown. The k-value between 8 and 47 °C was 5.08 ± 3.37 [-]. At
temperatures ranging from 20 to 46 ºC values for k were between 0.82 and 11.6 [-], with an
average of 4.05 ± 1.92 [-]. Below 20 ºC it increased to an average value of 10.06 ± 4.49 [-].
Using the k-value of 5.08 to predict the lag time at any temperature and equation 3 to
determine the specific growth rate resulted in a RSS value of 16.7.

Figur e 5. Parameter k (product of ì and u
m
) as function of the temperature for each growth
experiment of strain ATCC 29544 pre-cultured to various growth phases and grown at temperatures
from 8 up to 47 °C; × ODJ SKDVHJURZQFHOOV Hxponential phase grown cells; ^: early stationary
phase grown cells; -: middle stationary phase grown cells; o: stationary phase grown cells; : late
stationary phase grown cells. The dotted line represents the average value for the data from 20 up to
46 ºC.

Based on the RSS values for the models describing the lag time it can be concluded that the
hyperbolic model and reciprocal square root model best convey the experimental lag times.
Lag time and specific growth rate

127
The reciprocal square root model is recommended, since it has the ability of increasing the
lag time at higher temperatures and it contains more interpretable parameters.
The maximum number of cells of strain ATCC 29544 reached at the various incubation
temperatures between 8 and 46 ºC varied between approximately 10
7
and 10
9
CFU/ml, with
an average of 10
8.2
CFU/ml (Figure 6). At temperatures tested above 45 ºC, the maximum
number of cells reached was ±10
6
CFU/ml. The initial inoculum level, which varied
between ± 10
2
and 10
5
CFU/ml did not seem to affect the maximum population density.

Figur e 6. The initial inoculum of strain ATCC 29544, pre-cultured to various physiological states and
its maximum population density at various temperatures. ∆: maximum population density; o: initial
number of cells. The dotted line represents the average value for the maximum population density
data from 8 up to 46 ºC.

Discussion
Cronobacter spp. may contaminate infant formulae either during production or during
bottle preparation. In factory environments Cronobacter spp. may grow in wet spots and
survive in dust containing residues of infant formulae and contaminate the product after the
drying process. In hospital and household kitchens infant formulae may be contaminated
likewise with dry or wet residues by utensils and via environmental vectors. In all situations
Chapter 6
128
the physiological state of the contaminating cells is not known. But it is known that the
duration of the lag time may depend not only on the growth environment but also on the
previous history of cells (12, 16). It was envisaged that the time of pre-culturing, i.e. the
physiological growth phase of the inoculum, could have an effect on the lag time but
probably not on the specific growth rate (16). In our study, based on visual observation of
growth curves and on statistical analysis of growth data, no effect on either parameter was
found. A possible explanation might be that the cells were spiked into dry powdered infant
formula prior to the growth experiments, and their metabolic activity had possibly changed
in all cases to a comparable level. The period (3 to 10 days) the cells were in the powdered
infant formulae before carrying out a growth experiment was found not to influence the
growth parameters. Cells that had been in the powdered infant formula for up to 4 weeks
did not show a longer lag phase either (data not shown). Another explanation for not
measuring differences in resuscitation times might be the rich growth medium
(reconstituted powdered infant formula) used. In this medium Cronobacter spp. cells had a
rather short lag time of 1.71 ± 0.50 h at 37 °C and small differences in lag phase may not
become apparent.
Our specific growth rate data were compared to specific growth rates published by
Nazarowec-White et al., (1997) and Iversen et al., (2004) (6, 10) as shown in Figure 3. In
our study the specific growth rates of Cronobacter spp. spiked into dry infant formula is
comparable to Nazarowec-White et al., (1997) who reported similar specific growth rates.
Specific growth rates reported by Iversen et al., (2004) were consistently over 10% lower.
One cause for the difference with the latter study may be due to differences in growth
conditions: in our experiments spiked dry powdered infant formula samples reconstituted
with sterile tap water were used as growth medium, whereas Iversen et al., measured
growth in other media by inoculating with an overnight TSB culture (6). Another difference
is the use of the rapid automated bacterial impedance technique, which measures growth
only at much higher levels (> 10
6
CFU/ml) than the plate count method used in our study
(16). In the current study, the plate count technique was used and growth was measured in a
more representative range and over multiple logs of bacterial counts, thus the estimates for
the lag time and specific growth rate can be considered more accurate (2).
Lag time and specific growth rate

129
The experimental design, based on initial estimates of specific growth rate and lag time,
allowed prediction of the growth with enough accuracy to determine the appropriate
sampling times and dilutions to measure growth curves, such that our quality requirements
were met for 95% of the experimental growth curves. The modified Gompertz model
(Equation 1), the secondary Rosso model (Equation 3) and the reciprocal square root model
(Equation 4) successfully estimated the lag time and specific growth rate for the whole
growth temperature range (Table 4 and 5). The theoretical minimal-, and maximum growth
temperatures as fitted with both the secondary growth models are 2.19 ºC and 48.9 ºC
(Equation 2) and 3.60 ºC and 47.6 ºC (Equation 3), respectively. In practice, however,
growth of Cronobacter spp. strains has been observed between 5.5 ºC and 47 ºC (4).
The fact that low numbers of Cronobacter spp. cells have occurred in dry infant formulae
(5, 9, 10) in combination with the relatively short lag time and high specific growth rates
found in this study, underscores the need for careful preparation and use. A study by
Pagotto et al., (2003), however, showed that a number of 10
5
CFU/ml of certain
Cronobacter spp. strains could be lethal to suckling mice after ingestion, but not all strains
appeared to be pathogenic (11). As there is a lack of information about virulence factors,
not all the Cronobacter spp. strains necessarily need to be regarded as potential pathogens.
Nevertheless the results described in this paper (Table 4 and 5) allow to predict growth of
Cronobacter spp. in reconstituted infant formulae under conditions that closely mimic the
conditions in the actual food product in both hospital and household settings and will be
useful in designing effective control measures.

Acknowledgements
The authors wish to thank Ingrid Maas for skilful technical assistance. The authors thank
Ingrid Hombergen and Etienne Vidal for carrying out part of the experiments. Dr. H.L.
Muytjens is thanked for providing the Cronobacter spp. strains MM9 and MC10. The
financial support by Nestlé Research Center, Lausanne, Switzerland, is gratefully
acknowledged.
Chapter 6
130
Refer ences
1. Baranyi, J., and T. A. Roberts. 1995. Mathematics of predictive food
microbiology. Int. J. Food Microbiol. 26:199-218.
2. Begot, C., I. Desnier, J. D. Daudin, J. C. Labadie, and A. Lebert. 1996.
Recommendations for calculating growth parameters by optical density
measurements. J. Microbiol. Methods. 25:225-232.
3. FAO/WHO. 2004. Enterobacter sakazakii and Other Micro-organisms in
Powdered Infant Formula: meeting report. Microbiological Risk Assessment
Series No. 6, Rome, Italy.
4. Farmer III, J. J., M. A. Asbury, F. W. Hickman, D. J. Brenner, and the
Enterobacteriaceae study group. 1980. Enterobacter sakazakii: A new species of
"Enterobacteriaceae" isolated from clinical specimens. Int. J. Syst. Bacteriol.
30:569-584.
5. Heuvelink, A. E., M. Ahmed, F. D. Kodde, J. T. M. Zwartkruis-Nahuis, and E.de
Boer. 2002. Enterobacter sakazakii in melkpoeder. De Ware(n)chemicus. 32:17-
30.
6. Iversen, C., M. Lane, and S. J. Forsythe. 2004. The growth profile,
thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant
formula milk. Lett. Appl. Microbiol. 38:378-382.
7. Juneja, V. K., H. M. Marks, and L. Huang. 2003. Growth and heat resistance
kinetic variation among various isolates of Salmonella and its application to risk
assessment. Risk Analysis. 23:199-213.
8. Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004. Occurrence of Enterobacter sakazakii in the food production
environments and households. The Lancet. 363:39-40.
9. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
10. Nazarowec-White, M., and J. M. Farber. 1997. Incidence, survival, and growth of
Enterobacter sakazakii in infant formula. J. Food Prot. 60:226-230.
11. Pagotto, F. J., M. Nazarowec White, A. Bidawid, and J. M. Farber. 2003.
Enterobacter sakazakii: infectivity and enterotoxin production in vitro and in vivo.
J. Food Prot. 66:370-375.
12. Pin, C., G. D. Garcia de Fernando, J. A. Ordonez, and J. Baranyi. 2002. Analysing
the lag-growth rate relationship of Yersinia enterocolitica. Int. J. Food Microbiol.
73:197-201.
13. Ratkowsky, D. A., J. Olley, T. A. McMeekin, and A. Ball. 1982. Relationship
between temperature and growth rate of bacterial cultures. J. Bacteriol. 149:1-5.
Lag time and specific growth rate

131
14. Rosso, L., J. R. Lobry, S. Bajard, and J. P. Flandrois. 1995. Convenient model to
describe the combined effects of temperature and pH on microbial growth. Appl.
Environ. Microbiol. 61:610-616.
15. Smeets, L. C., A. Voss, H. L. Muytjens, J. F. G. M. Meis, and W. J. G. Melchers.
1998. Genetische karakterisatie van Enterobacter sakazakii isolaten van
Nederlandse patiënten met neonatale meningitis. Ned. Tijdschr. Med. Microbiol.
6:113-115.
16. Swinnen, I. A. M., K. Bernaerts, E. J. J. Dens, A. H. Geeraerd, and J. F. Van Impe.
2004. Predictive modelling of microbial lag phase: a review. Int. J. Food
Microbiol. 94:137-159.
17. Van Schothorst, M. 1976. Resuscitation of injured bacteria in foods. p. 317-328. In
Skinner and Hugo (eds.), Inhibition and inactivation of vegatative microbes.
Academic Press, Inc. (London), Ltd., London.
18. Zwietering, M. H., H. G. A. M. Cuppers, J. C. de Wit, and K. Van 't Riet. 1994.
Evalution of data transformations and validation of models for the effect of
temperature on the bacterial growth. Appl. Environ. Microbiol. 60:195-203.
19. Zwietering, M. H., J. T. de Koos, B. E. Hasenack, J. C. de Wit, and K. Van 't Riet.
1991. Modeling of bacterial growth as a function of temperature. Appl. Environ.
Microbiol. 57:1094-1101.
20. Zwietering, M. H., I. Jongenburger, F. M. Rombouts, and K. Van 't Riet. 1990.
Modeling of the bacterial growth curve. Appl. Environ. Microbiol. 56:1875-1881.



Chapter 6
132
Growth during cooling

133
Chapter 7
134
Abstr act
Reconstituted infant formulae are excellent growth media for Cronobacter spp. (formerly
Enterobacter sakazakii) and other micro-organisms that may be present in such products.
Immediate consumption or rapid cooling and storage at low temperature are therefore
recommended as control measures to prevent microbial growth. Placing a container filled
with reconstituted liquid formula in the refrigerator, however, does not mean that the
temperature of the liquid is directly the same as the set-point of the refrigerator. This study
describes the temperature profiles and methods to predict lag time and possible growth of
Cronobacter spp. during the cooling process in three types of containers. The overall heat
transfer coefficients (α) were determined and were shown to have a very large variability in
both household refrigerators and an air-ventilated refrigerator equipped with a fan. A
mathematical model was built to predict the growth of Cronobacter spp. under dynamic
temperature conditions using three models for the lag time. The various estimations for the
lag time had a remarkably strong impact on the predicted growth. The assumption of a
constant k-value (k = lag time x specific growth rate = λ x μ = 2.88) fitted the experimental
data best. Predictions taking into account the large variability in heat transfer showed that
proliferation of Cronobacter spp. during cooling may be prevented by limiting the volume
to be cooled to portion-size only, or by reconstituting at temperatures of 25 °C or lower.
The model may also be used to predict growth in other situations where dynamic
temperature conditions exist.



This Chapter has been published as
“Gr owth of Cronobacter spp. under dynamic temper atur e conditions occur r ing dur ing
cooling of r econstituted powder ed infant for mula”

Kandhai, M.C., P. Breeuwer, L.G.M. Gorris, M.H. Zwietering, and M.W. Reij
Journal of Food Protection, 2009; 72: 2489-2498.
Growth during cooling

135
Intr oduction
Cronobacter spp., until recently known as the species Enterobacter sakazakii, is a group of
opportunistic pathogenic micro-organisms belonging to the family of Enterobacteriaceae
(9, 13). Although only recently brought to wider attention of infant formulae manufacturers
and the public through several outbreaks and public recalls, Cronobacter spp. have been
associated with neonatal deaths as early as 1958 (32). Cases of neonatal meningitis or
necrotizing enterocolitis due to Cronobacter infection have in certain cases been associated
with powdered infant formulae (4,06,029). Several investigations showed that the micro-
organism can be found in a variety of environments and that contamination may occur
during manufacturing and during preparation of the bottles (12, 16, 19, 20, 22). Reported
concentrations in dry powdered formulae ranged from 0.002 to 0.92 CFU/g (7).
In dry powdered infant formula Cronobacter spp. are not able to grow, but after the
addition of water, reconstituted infant formula is a good medium for growth with only the
barriers of short storage and low temperature to prevent bacterial growth. After
reconstitution Cronobacter spp. can multiply at temperatures between 3.6 and 47.6 ºC (12,
17, 23). At room temperature (25 ºC) the generation time was determined to be 0.59 h (17).
A study by Pagotto et al., (2003), suggests that a number of 10
5
CFU/ml of certain
Cronobacter spp. strains could cause illness in suckling mice after ingestion (24).
To prevent multiplication of micro-organisms most manufacturers prescribe consumption
of infant formula directly after preparation. For practical reasons, however, caregivers
sometimes prefer to prepare all bottles needed for one day in advance (26). Specifically in
hospitals it is not uncommon that infant formulae are prepared once per day and placed in a
refrigerator to be used within 24 h (27) or within 30 h with the refrigerator set at 4 °C (1).
However, even if the temperature of the refrigerator is controlled, the temperature of the
reconstituted formula itself is not controlled yet. Many manufacturers instruct that their
powdered infant formulae should be reconstituted with moderately warm (lukewarm) water
to assure that the powder dissolves well. The liquid formula then needs time to cool down
in the refrigerator. The cooling rate of the liquid depends on the temperature, the filling
Chapter 7
136
rate, and the air velocities in the refrigerator, the geometry of the food container, the
thermal properties of food and bottle, and the volume of the container to be cooled.
In household refrigerators, also referred to as static refrigerators, heat transfer at the
container surface is principally due to natural convection by a very limited airflow caused
by variations in air density. These variations are mainly related to differences in
temperature, filling of the fridge, and humidity gradients. In air-ventilated refrigerators
mechanic ventilation forces air convection which improves heat transfer (18). Air-
ventilated refrigerators are often, but not always, used in hospitals and nurseries.
To our knowledge, investigations into the cooling process of baby bottles are scarce. Rosset
et al., (27) studied temperature profiles in 25 neonatal care units and concluded that cold
storage was the operation with the largest impact on potential growth of Cronobacter spp..
In the report of the technical meeting on Enterobacter sakazakii and Salmonella in
powdered infant formula (8) a risk assessment model has been discussed that estimates the
fate of Cronobacter spp. during various scenarios of preparation and storage. The aim of
our study was to make a detailed investigation, both experimentally and by modeling, of the
growth opportunities of Cronobacter spp. during cooling of reconstituted infant formula in
refrigerators and to use the findings to propose control measures that may reduce and
prevent the growth of Cronobacter spp. during the cooling process.

Mater ials and methods
Preparation of reconstituted infant formula for cooling experiments
Powdered infant formula suitable for infants less than 6 months of age, was bought locally
and had a bacterial count of less than 10
2
CFU/g, which did not influence the experiments
(data not shown). In accordance with the instructions on the label, tap water was brought to
a full boil in a covered water boiler, cooled down on the bench to approximately 45 ºC and
mixed with the powdered infant formula to obtain reconstituted formula with a temperature
of 40 ± 3 °C.

Growth during cooling

137
Cooling experiments
A total of 25 cooling experiments were performed using an experimental set up whereby
three different container types with reconstituted infant formulae were placed in the
refrigerator and the temperature during cooling was recorded. The following three
polycarbonate containers were used: a measuring beaker equipped with a lid with a volume
of 1,000 ml and a diameter of 0.10 m, (baby) bottle Type 1 (maximum 260 ml) with a
diameter of 0.05 m, and (baby) bottle Type 2 (maximum 120 ml) with a diameter of 0.04 m
(Table 1).

Table 1. Parameters used in Equations 2 and 3
Par ameter Value Dimension Refer ence
Specific heat capacity of infant formula (c
p
) 3.93 kJ kg
-1
°C
-1
(3)
Density of infant formula (ρ) 1,032 kg m
-3
(30)
Radius of 1-liter measuring beaker (r) 0.05 m -
Radius of bottle Type 1 (r) 0.025 m -
Radius of bottle Type 2 (r) 0.02 m -
Height of 1-liter measuring beaker (ht) 0.14 m -
Height of bottle Type 1 (ht) 0.115 m -
Height of bottle Type 2 (ht) 0.95 m -
Thermal conductivity of infant formula (k
liquid
) 0.52 W m
-1
°C
-1
(30)
Thermal conductivity of bottle wall,
polycarbonate (k
containerwall
)
0.21 W m
-1
°C
-1
(2)

Reconstituted infant formula was divided among the measuring beaker and the two types of
baby bottles. While the bottles had different geometries, both were filled with 120 ml of
formula. The measuring beaker held 1,000 ml. The three containers were placed, in the
center of, either an incubator ventilated with air or in one of the 9 different static air
refrigerators (further referred to as household refrigerators). Eight out of nine household
refrigerators were located in households, and contained different amounts of other food
materials. The ninth household type refrigerator was located in the laboratory, and was
Chapter 7
138
empty except for the three containers. The temperature of the household refrigerators
varied, which was recorded as a given parameter and not set as a controlled experimental
parameter. During cooling the temperature was recorded in the center and near the wall of
each container every 5 minutes for 24 h using thermistor metal oxide sensors (CM-UU) and
a squirrel data logger (both: Eltek Ltd., Cambridge, UK). The temperature measured in the
center of the bottles was slightly higher than the temperature measured near the wall and
the difference was 0.5 ºC at most. The temperature measured in the center of the bottle was
chosen for use in the model predictions, for a more fail safe approach. The internal air
temperature in the refrigerator was measured with a thermistor as well.

Cooling model for reconstituted infant formula
Equation 1 describes the change in temperature of a bottle or beaker (T
C
) assuming that the
temperature in the liquid is homogeneous and that heat is transferred from the liquid via the
wall of the bottle to the air.

( )
R
T
C
T A .
dt
C
dT
p
c . . V ÷ ÷ = o p (1)
where V [m
3
] is the volume of the reconstituted infant formula in the measuring beaker/bottle, ρ [kg
m
-3
] is the density of the reconstituted infant formula; c
p
[J kg
-1
°C
-1
] is the specific heat capacity of
the reconstituted infant formula; t [s] is the time; α[ J m
-2
°C
-1
s
-1
= W m
-2
°C
-1
] is the overall heat
transfer coefficient; A [m
2
] is the surface area of the object; T
C
[ºC] is the temperature of the prepared
liquid formula and T
R
[ºC] is the air temperature in the refrigerator.

Values of the parameters are shown in Table 1. After integration and rewriting this equation
transforms into:
( )
e
t
p
c r
R
T
, C
T
R
T
t , C
T
|
|
.
|

\
|
÷
÷ + =
p
o 2
0
(2)
where T
C,0
[ºC] is the temperature in the center of formula in the measuring beaker or bottle at the
beginning of the cooling process and r [m] is the radius of the measuring beaker or bottle.
Growth during cooling

139
Here, we assume that both the bottles and the measuring beaker can be described as a
cylinder and that only the side surface area of the liquid (A = 2·π·r·h) contributes to heat
transfer, not the top or the bottom of the container. The ratio of surface area and volume
(A/V) can then be described as 2 r
-1
.

As shown in equation 3, the overall heat transfer coefficient α consists of the external heat
transfer coefficient in air, α
external
, the coefficient of the wall, α
containerwall
[the ratio of
thermal conductivity of the container material (k
containerwall
) and the thickness of the wall],
and the internal heat transfer coefficient, α
product
.
product all containerw exernal
o o o o
1 1 1 1
+ + =

(3)

The external heat transfer coefficient largely depends on the movement of air in the
refrigerator. When a fan is present, this coefficient can be expected to be higher than in
household refrigerators that are not equipped with a fan. The internal heat transfer
coefficient depends on the movement of liquid in the bottle. During cooling of the bottles
the liquid may circulate slowly due to free convection. If free convection does not occur,
the internal heat transfer coefficient can be estimated by the ratio of the thermal
conductivity of the reconstituted infant formula (k
liquid
) and the radius of the container (r).

Test organism
Stock cultures of Cronobacter sakazakii (formerly named Enterobacter sakazakii) ATCC
29544 were maintained at -80 ºC in cryo vials (Greiner Bio-one GmbH, Frickenhausen,
Germany), containing a stationary-phase culture suspension in Brain Heart Infusion (BHI;
Difco, Becton Dickson, Maryland, U.S.A) broth with 30% (vol/vol) glycerol (Fluka-
chemica, Buchs, Switzerland).

Chapter 7
140
Growth medium, inoculation and sampling
The inoculum of strain ATCC 29544 was prepared by transferring 100 µl of the stock
culture into 100 ml BHI broth, followed by incubation for 18 to 20 h at 37 ºC. Cells were
harvested, washed, sprayed onto dry infant formula to obtain a final concentration of 10
4
to
10
6
CFU/g and stored for 3 to 10 days as described elsewhere (17). The powder containing
Cronobacter spp. was then mixed with water, similar to the procedure described at
“Preparation of reconstituted infant formula for cooling experiments”, to obtain
reconstituted formula with 10
3
to 10
5
CFU/ml at a temperature of 37 ± 3 °C. This formula
was distributed between the Type 1 bottle and the measuring beaker and placed in the
laboratory refrigerators immediately. No other products were present in the refrigerators.
Samples were taken every 30 min during the first 4 h and subsequently every 60 min, by
inserting a 1-ml disposable pipette with balloon, leaving the bottles in the refrigerator. If
appropriate, samples were first diluted in Peptone Saline Solution (PPS; NaCl 8.5 g/L
supplemented with 1 g/L neutralized Bacteriological Peptone from Oxoid, Basingstoke,
England) and then plated in duplicate onto Tryptone Soy Agar (TSA; Oxoid LTD.,
Basingstoke, Hampshire, England) using a spiral plater device (Eddy Jet, Leerdam, The
Netherlands). Plates were incubated for 20 to 24 h at 37 ºC and counted manually. The
inoculation levels chosen are higher than the very low numbers found in infant formulae (7,
11, 21, 23). A minimum level required to accurately count the organisms plated by the
spiral plater, however, is 6 ×10
2
CFU/ml, and inoculum levels were chosen to be slightly
higher than this minimum.

Predicting growth of Cr onobacter spp. under dynamic temperature
conditions
A growth model was built that uses the lag time and the specific growth rate of
Cronobacter spp. during the constantly changing temperature conditions that occur during
cooling. The specific growth rate (µ
m
) for Cronobacter spp. at each temperature (T
i
) was
Growth during cooling

141
described with the secondary growth model (Equation 4) of Rosso et al., (28) using cardinal
parameters previously established (Table 2) (17).

If
max min
T T T < < (4)
then
( )
( )( )
( )( ) ( )( ) | |( )
opt
min opt
2
min opt max opt opt min opt
2
min max
u u
|
|
.
|

\
|
÷ ÷ + ÷ ÷ ÷ ÷
÷ ÷
=
T T T T T T T T T T T
T T T T
T
m
i i
i
i


and if
max
min
T
i
T
T
i
T
>
s
then ( ) 0 =
i
T
m
u

where T
i
is the temperature [ºC] at each moment in the center of the volume of the infant formula in
the measuring beaker or the bottle; T
min
[C] is the extrapolated minimum temperature, T
max
[C] is the
extrapolated maximum temperature, T
opt
[C] is the temperature at which the specific growth rate is
maximal, and µ
opt
[h
-1
] is the µ
m
[h
-1
] at this optimal temperature (28).

Three models were used to estimate the lag time. First, lag times were estimated using the
inverse of the secondary Ratkowsky model (Equation 5) with parameters previously
determined as shown in Table 2 (17).

( ) ( ) ( ) ( ) | | { } ( ) | |
2
max
exp 1
min
ln ln
÷
÷ ÷ ÷ = T
i
T c . T
i
T b
i
T ì (5)

where b [ºC
-1
h
-0.5
], and c [ºC
-1
] are the so-called Ratkowsky parameters (34). T
i
[ºC] used is the
average temperature of the simulation interval. The T
min
and T
max
parameters were determined from
previous growth experiments at temperatures ranging from 8 up to 47 ºC (17).
Chapter 7
142
Second, the lag times were estimated using a constant k-value, which is the product of the
specific growth rate and the lag time (Equation 6). It is known that the k-value has a large
variability but it can be considered constant over a wide range of temperatures (34).

( ) ( )
i
T .
i
T k u ì = (6)

Third, the logarithmic model was used to estimate lag times (Equation 7) (25).

( ) ( ) ( )
l
b
i
T
l
c
i
T + = ln
10
log ì (7)

where, λ [h] is the lag time, b
l
and c
l
are parameters as used in the microbial risk assessment model
(25) (available at http://www.mramodels.org/).



Table 2. Earlier determined cardinal growth parameters taken from literature, to calculate the specific
growth rate (µ [h
-1
]) and the duration of the lag phase (ì [h]) as a function of temperature
Model Par ameter Value Dimension Equation Refer ence
T
min
3.6

°C 4 (17)
T
max
47.6 °C 4 (17)
T
opt
39.4 °C 4 (17)
Secondary Rosso
growth model
µ
opt
2.31 h
-1
4 (17)
T
min
2.19 5 (17)
T
max
48.9 5 (17)
b 0.023 °C
-1
h
-0.5
5 (17)
Inverse of
secondary
Ratkowsky model
c 0.645 °C
-1
5 (17)
b
l
4.309 - 7 (25) Logaritmic model
c
l
-1.141 - 7 (25)
Growth during cooling

143
It should be noted that all specific growth rate and the lag time models (Equation 4, 5, 7)
are used with earlier estimated parameters (17) and are therefore predictions. To
specifically estimate lag time effects in this study, the k-value (Equation 6) was determined
from experiments taking into account the effect of a dynamically changing temperature on a
lag time value.

To determine the length of the lag time at constantly changing temperature, the fraction of
the lag time elapsed (ìf
t
) in every time step was calculated numerically according to
Equation 8.
( )
1
1
60
÷
+
÷
=
|
|
|
|
.
|

\
|
i
f
i , i
T
t
i
f ì
ì
A
ì

(8)

where At is the time step of the simulation [5 min] and ìf
t
[-] is the fraction of the lag time elapsed at
step i and
) (
, 1 i i
T
÷
ì
[h] is the average of the lag time of the current and the previous time step.

Each fraction ìf
i
was added to the value of the previous time step. Once ìf
i
≥ 1, the lag time
was assumed to be completed and growth was assumed to be starting with a specific growth
rate corresponding to the temperature prevailing in the bottle at that moment.
To include changing temperature also during the growth stage, the dynamic first order
model for bacterial growth was used and was solved numerically. In each time step the
average temperature (and consequently the specific growth rate) was taken as constant,
providing the exponential function as outcome (Equations 9A to C). As a conservative
estimate, growth was predicted assuming ì=0, using Equation 9A only, thus assuming that
Cronobacter spp. started growing instantaneously after addition of water. So, any further
steps can be described by Equation 9A, assuming that during the full time step (Δt) there is
exponential growth. When there was a lag time, the fraction of the lag time elapsed was
Chapter 7
144
calculated using Equation 8. When the fraction of the lag time elapsed was less than 1, the
number of micro-organisms was stable (Equation 9B). To exactly determine the
commencement of growth after the lag phase, in the step that ìf
t
has become greater than 1,
Equation 9C describes growth in the remaining fraction (ìf
i
- 1) of this time step. Equation
9C is valid only in the first iteration after the lag phase has been completed.

After the first 1 >
i
f ì :
( ) t .
i
T
m
e
i
N
i
N
A u
1 ÷
= (9A)

For 1 <
i
f ì :
0
N
i
N = (9B)

For the first value of 1 >
i
f ì :
( ) ( ) t .
i
f .
i
T
m
e N
i
N
A ì u 1
0
÷
= (9C)

where N
i
[CFU/ml] is the number of micro-organisms present at time step i; µ
m
(T
i
) [h
-1
] is the specific
growth rate at the temperature prevailing in the current time step i and ìf
i
[-] is the fraction of the lag
time elapsed at step i.

Data analysis
Temperature profiles during cooling were fitted into Equation 2 using the solver function in
Microsoft Office Excel 2003 by minimizing the residual sum of squares (RSS) to estimate
the overall heat transfer coefficient (α) and the air temperature (T
R
). Experimental growth
curves were fitted to the dynamic growth model to estimate the k-value using the same
software. Standard deviations were calculated with Execl and are reported as plus or minus
the mean. Confidence intervals were calculated in Excel. The increase in numbers of
Cronobacter spp. cells assessed in the duplicate growth experiments and the α estimated for
the different type of refrigerators were analyzed by analysis of variance (ANOVA),
Growth during cooling

145
univariate analyses using SPSS (SPSS release 12.01 for Microsoft Windows; SPPS Inc.,
Chicago, U.S.A.). Significant differences are presented at a 95%-confidence level
(P s 0.05).

Results and discussion
Temperature variability within refrigerator
In our experiments large variations in air temperatures were observed both in time and
between different locations in the same empty household refrigerator (Figure 1A). Starting
one hour after closing the door with the temperature set at 6.3 °C, the temperature near the
cooling element in the back of the upper department varied between -8.4 °C and +10.4 °C.
This variation is due to switching on and off the refrigerator’s engine. The lower, middle
and upper compartments of the refrigerator showed more constant temperatures and were
on average of 5.5 ± 0.8, 6.5 ± 1.3 and 9.9 ± 0.7 °C respectively. These results are in line
with data reported elsewhere. James and Evans (14) surveyed 252 household refrigerators
without a fan and found that the coldest spot was on average 2.9 °C colder than the warmest
spot within the same refrigerator, whereas the greatest range of average temperatures
observed in a refrigerator was 12 ºC. Individual (not averaged) temperature differences
ranged from 4.5 up to 30.5 °C in the same refrigerator. Door openings resulted in an
increase in average temperature and in an even larger spatial variability (15). A variation in
local air temperature does not immediately change the temperature of a bottle placed in that
refrigerator, but varying temperature profiles may change the air flow and thus affect the
heat transfer coefficient (α). Additionally, the speed of cooling as described in Equation 1,
will be affected by variations in air temperature.
In air-ventilated refrigerators equipped with a fan, the air is continuously in motion and the
variation in air temperature is limited (Figure 1B). Near the air inlet the temperature varied
from 5.4 to 7.3 ºC (average 6.3 ± 0.5 °C) and near the door the temperature was 7.7 ± 0.1
°C, but all other locations had temperatures of 6.7 ± 0.1 °C.
Chapter 7
146
As a result of the variations in air temperature it was not possible to establish one value for
the air temperature required to fit the data to equation 2. Therefore, both the heat transfer
coefficient (α) and the air temperature (T
R
) were fitted simultaneously to the experimental
data according to Equation 2. The differences between the measured T
R
and the fitted T
R

were on average less than 1 ºC (data not shown).


Figur e 1. Air temperatures in an empty household set at 6.3 °C (A) and an air-ventilated refrigerator
set at 6.3 °C (B). Measurements started one hour after closing the door. Locations were (grey dashed
line), door of upper shelf; (grey solid line), center of the upper shelf; (x), back wall of upper shelf,
close to the cooling element (A) or the air inlet (B); (black dashed line), center of the lower shelf.
Growth during cooling

147
Overall heat transfer coefficient (α)
Figure 2 shows typical cooling profiles of the 1-liter beaker and both bottles placed in either
a household refrigerator (2A) and an air-ventilated refrigerator (2B). During the cooling
process the temperature fluctuated slightly due to variations in the air temperature. In a
household refrigerator set at 6.3 °C the time required for cooling the prepared formula to 10
ºC was typically 7.5 hours for 1,000 ml formula in the measuring beaker and 3.5 hours for
the 120 ml volume in the two types of bottles. In an air-ventilated refrigerator, cooling
times typically were 2 hours less.
Figur e 2A. Example of temperature profiles of reconstituted powdered infant formula in (black solid
line), 1-liter measuring beaker; (grey solid line), Type 1; and (black dashed line), Type 2 bottles,
cooled down in a household type refrigerator. The refrigerator had an average air temperature of 6.3
ºC.
Chapter 7
148
Figur e 2B. Example of temperature profiles of reconstituted powdered infant formula in (black solid
line), 1-liter measuring beaker; (grey solid line), Type 1; and (black dashed line), Type 2 bottles,
cooled down in an air-ventilated refrigerator. The refrigerator had an average air temperature of 6.3
ºC.

With the three containers, 25 cooling experiments were performed in nine different
household refrigerators and 9 experiments were performed in an air-ventilated refrigerator.
All cooling profiles (3×25 + 3×9) were fitted to Equation 1 to obtain the overall heat
transfer coefficient (α [W m
-2
°C
-1
]), a parameter independent of the reconstitution
temperature, the air temperature, and within a limited range independent of the size of the
object. No consistent differences were found between the α-values determined for the
refrigerators that were empty and those that contained also other materials, therefore the
data were pooled and used as a single data set. Figure 3A shows the α-values with their
standard deviation as a function of air temperature in household refrigerators. An effect of
air temperature on the heat transfer coefficient (α) was not observed indeed. The α-values
ranged from 5.4 to 12.9 [W m
-2
°C
-1
] in this series of experiments in which temperature
settings and filling rates of the 9 household refrigerators were variable.
Growth during cooling

149

Figur e 3A. Overall heat transfer coefficient (α) (Wm
-2
°C
-1
) with their standard deviations, as
determined in 9 different household refrigerators at various temperatures.
Markers indicate (×, +), 1-liter measuring beaker; (■, □), Type 1 bottle; (▲, Δ), Type 2 bottle.



Figur e 3B. Overall heat transfer coefficient (α) (Wm
-2
°C
-1
) with their standard deviations, as
determined in a selection of 11 experiments performed in a single, empty household refrigerator (h.h.)
and 9 experiments in a single air-ventilated (vent.) refrigerator.

Chapter 7
150
Figure 3B shows a selection of the α-values, which were measured in one single household
refrigerator and one air-ventilated refrigerator, both empty except for the experimental
setup consisting of the 3 containers. In this well-defined series of experiments the range of
α-values (6.1 to 12.9 [W m
-2
°C
-1
]) in the single household refrigerator was almost as wide
as in the 9 different household refrigerators shown in Figure 3A. The fitted heat transfer
coefficients as shown in Figure 3B are significantly higher (P = 0.01) than in the household
type refrigerators. The coefficients in this well-controlled refrigerator varied over the range
of 9.9-19.0 Wm
-2
°C
-1
. Average values and confidence intervals are shown in Table 3.

Table 3. Estimated overall heat transfer coefficient (α) [Wm
-2
°C
-1
] of bottles Type 1 and Type 2,
each containing 120 ml, and a measuring beaker containing 1000 ml of reconstituted infant formula as
measured in one household-type refrigerator and one air-ventilated refrigerator

Type of
r efr iger ator
Aver age
over all
α
[W m
-2
°C
-1
]
Number of
exper iments
Standar d
deviation
of aver age
95%-
confidence
inter val
[W m
-2
°C
-1
]
household 9.9 11 1.4 6.8 – 13.0
Measuring
beaker
air-ventilated 14.0 9 2.4 8.4 – 19.7
household 10.0 11 1.5 6.7 – 13.3
Bottle Type
1
air-ventilated 15.7 9 2.5 9.8 – 21.5
household 7.7 11 1.5 4.4 – 11.1
Bottle Type
2
air-ventilated 13.9 9 2.8 7.4 – 20.4

The average values for the beaker and the Type 1 bottle corresponded to cooling rates of
1.0 × 10
-4
and 2.0 × 10
-4
s
-1
, identical to the cooling rates assumed in the FAO/WHO
microbial risk assessment model (25). For bottle Type 2, α was significantly lower in
household refrigerators (ANOVA univariate analysis, P=0.001) than the corresponding
values of the measuring beaker and bottle Type 1. The Type 2 bottle, which is the smallest
baby bottle available on the Dutch market, was filled to its maximum volume. It might be
that the area available for heat transfer was slightly less than assumed in Equation 2, since
the bottle is rounded at the bottom and not a perfect cylinder. Another explanation is that
the lower α might be due to the fact that in this narrow bottle, with a radius of 0.02 m,
internal heat transfer can not be overlooked. When a circular flow does develop in a
Growth during cooling

151
container during cooling, the internal heat transfer coefficient (α) can be estimated to be in
the range of 150 to 250 W m
-2
°C
-1
using the theory of free convection from a vertical plate
(5) assuming a temperature difference of 0.5 °C between core and the wall. If in a narrow
bottle such a circular flow does not develop, the internal heat transfer coefficient depends
solely on the ratio of the thermal conductivity of the reconstituted infant formula (k
liquid
,
0.52 W m
-1
°C
-1
) and radius of the container (0.02 m). The resulting internal heat transfer
coefficient is 26 W m
-2
°C
-1
, and thus is in the same order of magnitude as the resulting
overall coefficient (see Figure 3). It may therefore have contributed significantly to heat
transfer resistance. To rule out the potential effect of insufficient liquid mixing, the
experiments testing the growth of Cronobacter spp. during cooling were continued using
bottle Type 1. Furthermore this bottle size and type is used by the majority of the caregivers
in The Netherlands .

Growth during cooling
On the basis of the estimated overall heat transfer coefficient, the temperature in a bottle or
measuring beaker could be predicted during the entire cooling process. Figure 4A shows
the temperature of a 1-liter beaker placed in a household refrigerator at 7.1 °C. With the
temperature at each time point known, both the specific growth rate and the lag time of
Cronobacter spp. were calculated using Equations 4 through 7.
Figure 4B shows that the specific growth rate dropped sharply during the first 5 hours of
the cooling process. In the same 5-h period, the lag time increased to values of 20 h and
more. The construction of the lag time under dynamic temperature conditions is shown in
Figure 4C. According to the inverse of the Ratkowsky model (Equation 5) the lag time was
3.2 h. An initial estimate for the k-value of 4.05 [-] taken from previous research (17) led to
a predicted lag time of 3.3 h (data not shown), while the logarithmic model (Equation 7)
predicted a lag time of 4.8 h. The resulting predicted and experimental numbers of
Cronobacter spp. under dynamic temperature conditions are shown in Figure 4D. Growth
of the organism was apparent in the 1-liter measuring beaker that was cooling down from
38 °C in the household-type refrigerator set at 7.1 °C. The increase in numbers during the
Chapter 7
152
entire 24 hours was 0.7 log CFU/ml. Figure5 shows graphs similar to Figure 4D for both
the 1-liter measuring beaker and the Type 1 bottle at 5, 7, 10 and 16 ºC. This range of
temperatures was chosen to mimic the range of temperatures that can be found in household
refrigerators, including the 30% of household refrigerators that exceed standard
temperatures (10, 31).


Figur e 4. Prediction and experimental values of growth during cooling of infant formula in a 1-
liter measuring beaker in a household refrigerator set at 7.1 °C. The reconstituted formula was
prepared from artificially contaminated powder and had a temperature of 38 °C at the start of the
experiment. A: (×), measured and (solid line), predicted temperature with 95%-confidence
interval (dashed lines). B: Predictions of (♦), specific growth rate and (lines), lag times at the
prevailing temperature. C: Fraction of lag time elapsed and D: number of C. sakazakii (×),
measured and (lines), predicted. Various lines represent: (grey dashed line), lag = 0; (black
dashed line ▪▪▪▪▪), lag time according to Ratkowsky; and (grey line), lag time predicted with the
logarithmic model.

Impact of lag time
In the predictions a small difference in lag time resulted in considerable differences in the
estimated growth. With a longer lag time (Figure 4C), the temperature of the formula
(Figure 4A) will be lower at the moment that growth commences, whereas also the specific
Growth during cooling

153
growth rate (Figure 4B) will be lower, together resulting in a considerably smaller increase
in cell count.
When the lag time was assumed to be zero (grey dashed lines), the exponential growth
function overestimated the experimental values considerably in all cases shown in Figure 5.
Although Cronobacter spp. cells are known to commence their growth quickly in infant
formula, they thus seem to require some time to adjust to their new environment and start
multiplying. The logarithmic model overestimated the lag time and thus underestimated the
growth in all cases (25). The results of the use of a k-value were often comparable to the
Ratkowsky model. The Ratkowsky model (black dashed lines) predicted the growth during
cooling at 7 and 10 °C quite well, but at other temperatures this approach underestimated
growth. Optimal fits to the experimental data were obtained with k = 1.75 at 4 °C, k = 3.49
at 7 °C, k = 4.07 at 10 °C, k = 2.06 [-] at 16 °C (grey solid lines). The best fit over all 8
experiments was obtained with k = 2.88 [-] (black solid line). These values are all within
the range of lag values previously reported for individual experiments at a constant
temperature (17, 34).
It should be noted that the experiments shown in Figures 4 and 5 were performed with
Cronobacter spp. cells that were present in the dry powder, which can be expected to be
injured and therefore show relatively long lag phases. It could be that, even longer lag times
might have been observed when lower levels of inocula were used than in the current study
and when naturally contaminated powered infant formula had been used. On the other hand,
in an additional series of experiments using an inoculum of cells grown overnight in BHI,
lag times were shorter and could be predicted best by assuming k-values ranging from 1.2
to 1.9 [-] (results not shown).

Chapter 7
154

Figur e 5. Growth during cooling of infant formula in (left) 1-liter measuring beaker and (right)
the Type 1 baby bottle at air temperatures of 5 °C, 7 °C, 10 °C, and 16 °C in an empty
household-type refrigerator. The markers represent experimental data, while the lines are
predictions, assuming (grey dashed line ▪▪▪▪), lag = 0; (light grey solid line) lag time according to
Ratkowsky; (black solid line), lag time as predicted with k = 2.88; (grey solid line), lag as
predicted with an optimal fit of the k-value to the two experiments at each specific temperature: k
= 1.75 at 5 °C; k = 3.49 at 7 °C; k = 4.07 at 10 °C; k = 2.06 at 16 °C.
Growth during cooling

155
Table 4. Prediction of the time required to cool down from 40 to 10 °C and the increase in cell counts
after 24 h in bottles Type 1 and Type 2, each containing 120 ml, and a measuring beaker containing
1,000 ml of infant formula. Infant formula was reconstituted at 40 °C and placed in refrigerators with
an air temperature of 7 °C. Overall heat transfer coefficients with their confidence intervals were
taken from Table 3 and for lag time the optimal value k =2.88 [-] was assumed. In parentheses are
95% confidence intervals
Type of
r efr iger ator
Time to r each 10 °C
[h]
Pr edicted incr ease in cell
counts in 24 h
[log CFU/ml]
household 6.8 (5.3 –10.0) 1.3 (0.8 – 2.3 )
Measuring
beaker air-ventilated 4.8 (3.5 – 8.1) 0.7 (0.2 – 1.7 )
household 3.4 (2.6 – 5.1) 0.2 (0.0 – 0.7)
Bottle Type 1
air-ventilated 2.2 (1.6 – 3.5) 0.0 (0.0 – 0.2)
household 3.6 (2.5 – 6.2) 0.3 (0.0 – 1.1)
Bottle Type 2
air-ventilated 2.0 (1.3 – 3.7) 0.0 (0.0 – 0.3)

Prediction of growth during cooling
Using the parameters for heat transfer and lag time derived from the experiments, a series
of simulations were performed, to evaluate the effect of reconstitution temperature and
refrigerator temperature. Regarding the lag time, the value k =2.88 [-] as found in this study
was assumed in all predictions.

Cooling time and growth during 24 hours
Table 4 presents an overview of the estimated time required to cool down infant formula
from 40 °C to 10 °C in refrigerators set at 7 °C and the increase in Cronobacter spp. cells in
bottles placed in such refrigerator for 24 h. In a household refrigerator, a 1-liter portion of
formula reconstituted at 40 °C needed at least 5.3 and up to 10 h to reach a temperature of
10 °C and in an air-ventilated refrigerator at least 3.5 h. Bottles needed on average 3.5 h in
the household type and 2.1 h in the air-ventilated type to cool down to 10 °C.
In 1-liter portions, growth of Cronobacter spp. was predicted to vary in both types of
refrigerators from 0.2 to 2.3 log CFU/ml. In bottles placed in household refrigerators, the
Chapter 7
156
increase in numbers was on average limited to 0.2 or 0.3 log CFU/ml. The maximum
increase in bottles placed in household refrigerators was 1.1 log CFU/ml for refrigerators
with the poorest heat transfer properties. In most air-ventilated refrigerators set at 7 °C no
increase (0.0 log on average) was predicted in bottles during 24 h, while the upper limit of
the confidence interval was 0.3 log CFU/ml. The variability of both the time to reach 10 °C
and the increase in cells over 24 h mainly originated from the considerable variability of the
overall heat transfer coefficient (Table 3).

Varying refrigerator temperature
Figure 6 shows the effect of air temperature in the refrigerator on the predicted growth of
Cronobacter spp. during cooling down from 40 °C in the 1-liter-measuring beaker and in
the Type 1 bottle filled with 120 ml. Assuming an average heat transfer coefficient of 10 W
m
-2
°C
-1
, no growth was predicted to occur in the bottle if the air temperature was 6 °C or
lower. Over 6 °C, the numbers were predicted to increase slightly within 4 h, whereas up to
24 h more significant multiplication was predicted, depending on the temperature.
In contrast to the bottle, the 1-liter container was predicted to support growth at all
refrigerator temperatures. This can be explained by the fact that the resulting temperature of
the formula (T
c
) during cooling in this container is higher then in the bottles, due to the
higher V/A ratio (Equation 1) in the 1-liter container.
At the common range of refrigerator temperatures (below 7 °C), growth was
predominantly in the first 4 h of cooling, while growth was limited in the period between 4
and 24 h. At higher temperatures, however, Cronobacter spp. was able to continue
multiplying after the first 4 h. With the refrigerator set at 16 °C a 4.2-log increase in the
bottle and an almost 5.5-log increase in the beaker were predicted.

Growth during cooling

157

Figur e 6. Increase in cell counts as a function of refrigerator temperature at (diamonds), 4 h and
(triangles), 24 hours after reconstitution at 40 °C, assuming an average (10 W m
-2
°C
-1
) heat transfer
coefficient. Open symbols and solid lines represent the 1-liter measuring beaker and closed symbols,
dotted lines the Type 1 baby bottle.

Effect of reconstitution temperature
Figure 7A shows the impact of reconstitution temperature on the predicted growth of
Cronobacter spp. in formula placed for 24 h in a household refrigerator at 4 °C with a heat
transfer coefficient that is either average, or at the upper or lower limits of its 95%-
confidence interval. In the 1-liter beaker growth was predicted to occur at reconstitution
temperatures over 25, 32 or 37 °C, depending on the assumption for the heat transfer
coefficient. Below a reconstitution temperature of 25 °C no growth was predicted under any
circumstances. For the 120-ml bottle our model did not predict growth, unless the
reconstitution was over 35 °C and the heat transfer coefficient was at its lower confidence
limit. Reconstitution at more than 50 °C was not included, since our experiments and
simulations were not suitable to measure and to predict cell numbers after thermal
inactivation which likely plays a role at these temperatures. For comparison with our study,
results obtained using the microbial risk assessment model (25) are shown in Figure 7B as
an increase in relative risk compared to a base-line scenario, which is reconstitution at
Chapter 7
158
25 °C. Simulations were limited to 24 h cooling. Although the end-points were not
identical, the microbial risk assessment model and our study concluded that in bottles
growth will either not occur (both studies), or will be limited to 0.45 log units in case of
poor heat transfer (this study). When cooling 1-liter containers, however, the reconstitution
temperature was found to affect the proliferation of Cronobacter spp. and it may well affect
growth of other micro-organisms, possible present in the reconstituted infant formula. Our
study predicted growth of Cronobacter spp. already in 1 liter of formula after reconstituting
at 26 °C when placed in a household refrigerator with the poorest heat transfer properties.
It should be noted that the latter situation is not a worst case scenario. The heat transfer
coefficient assumed in our study was measured in an empty refrigerator. Refrigerators in a
home situations may have even poorer heat transfer properties, because the refrigerator may
be packed with other food products, and the door may be opened to frequently, or bottles
may be placed in the refrigerator door, where heat transfer is at its minimum (15).
Moreover, the Cronobacter spp. cells used in this study were present in the powder and can
be expected to have a longer lag phase than actively growing cells. When, formula is
contaminated during reconstitution with cells that are actively growing, a shorter lag phase
and quicker multiplication may be envisioned.

Growth during cooling

159

Figur e 7. The effect of reconstitution temperature on (A) increase in log count (this study) and (B)
increase in relative risk (25) during cooling for 24 hours in a refrigerator with an air temperature of
4°C. Heat transfer coefficients are assumed to be (triangles), average (10 W m
-2
°C
-1
); (crosses) at
lower (6.7 W m
-2
°C
-1
); and (squares) at upper (13 W m
-2
°C
-1
) limit of confidence interval. Open
symbols represent a 1-liter beaker and closed symbols the Type 1 bottle filled with 120 ml of infant
formula.
Chapter 7
160
Possible contr ol measur es
Focusing on the cooling phase only, there are several control measures that may help to
prevent and minimize growth of Cronobacter spp. in situations where bottles and/or
syringes of infant formula have to be prepared in advance from powdered formula for use
within 24 h. For example:
- Cooling reconstituted infant formula in portion size containers only, such as bottles,
may prevent multiplication quite drastically as compared to cooling of 1–liter portions.
- Replacing household-type refrigerators with air-ventilated refrigerators with better heat
transfer properties is effective in reducing growth during cooling. However, air-
ventilated refrigerators can not totally prevent growth in 1–liter portions of prepared
formula with reconstitution temperatures around 40 °C.
- Our study predicts no growth during cooling after formula reconstitution at or below
25 °C. This may suggest that limiting the reconstitution temperature to 25 °C might be
an effective control measure, provided that reconstitution is followed by immediate
cooling of bottle-sized portions in a refrigerator. The formulae used in this study
dissolved well at lower temperatures; for those powders that may not there may be not
dissolve at room temperature and need to be reconstituted at 40 °C or more, quick
cooling of bottle-sized portions (for instance using running water or ice) may be an
alternative to prevent growth during cooling. Reformulation to allow powders to
dissolve at lower temperature would, however, be preferred.
- Although lowering the set points of refrigerators in neonatal care units and/or
households to 2 °C or less might seem a likely control measure, our results suggests
that this may have only a very limited effect on multiplication during cooling. Only
when the current set point is above 7 °C such measure might have a significant effect.

This study provides measurements of both heat transfer characteristics and microbiological
growth of Cronobacter spp. in containers of reconstituted infant formula, along with a
model that is able to estimate the variability of these processes taking place simultaneously
Growth during cooling

161
under dynamic temperature conditions. Powdered infant formula manufactures, neonatal
health care professionals and (inter)governmental organizations may take benefit from the
specific findings to further improve guidelines and best handling practices (33).
The lag time, which is known to be highly variable, has a remarkable impact on the overall
growth opportunities of Cronobacter spp., as it affects both the moment in time and the
temperature at which the organism starts to grow. The predictive model proposed can be
used to estimate the proliferation of Cronobacter spp. and potentially of other micro-
organisms that might be present in powdered infant formula. Using the predictive model,
exposure of consumers can be simulated for different scenarios of preparation and
consumption of infant formulae, thereby aiding governments and industries to identify
effective control measures to protect the vulnerable consumers of this product.

Acknowledgements
The authors wish to Kishor Khatri and Els Peters for their contribution to the experimental
part of this work. This project was financially supported by Nestlé Research Center,
Lausanne, Switzerland.

Refer ences
1. Agostoni, C., I. Axelsson, O. Goulet, B. Koletzko, K. F. Michaelsen, J. W. Puntis,
J. Rigo, R. Shamir, H. Szajewska, D. Turck, Y. Van den Plas, and L. T. Weaver.
2004. Preparation and handling of powdered infant formula: a commentary by the
ESPGHAN Committee on Nutrition. J. Ped. Gastroenterol. Nutr. 39:320-322.
2. Anonymous. Polycarbonate. Available at:
http://en.wikipedia.org/wiki/Polycarbonate. Accessed 18 February 2008.
3. Anonymous. Liquids and Fluids - Specific Heat Capacities. A comprehensive list
of some common liquids and fluids as acetone, oil, paraffin, water and many more
and their specific heat capacities. Available at:
http://www.engineeringtoolbox.com/specific-heat-fluids-d_151.html. Accessed 18
February 2008.
4. Biering, G., S. Karlsson, N. C. Clark, K. E. Jonsdottir, P. Ludvigsson, and O.
Steingrimsson. 1989. Three cases of neonatal meningitis caused by Enterobacter
sakazakii in powdered milk. J. Clin. Microbiol. 27:2054-2056.
Chapter 7
162
5. Churchill, S. W., and H. H. S. Chu. 1975. Correlating equations for laminar and
turbulent free convection from a vertical plate. Int. J. Heat Mass Transfer.
18:1323-1329.
6. Clark, N. C., B. C. Hill, C. M. O'Hara, O. Steingrimsson, and R. C. Cooksey.
1990. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial
outbreaks. Diagn. Microbiol. Infect. Dis. 13:467-472.
7. Fanning, S., and S. J. Forsythe. 2008. Isolation and identification of Enterobacter
sakazakii. p. 27-59. In: J. M. Farber, and S. J. Forsythe (eds.), Enterobacter
sakazakii (emerging issues in food safety) ASM Press, Washington, D.C., New
Brunswick, New Jersey.
8. FAO/WHO. 2006. Enterobacter sakazakii and Salmonella in powdered infant
formula: meeting report. Microbiological Risk Assessment Series No. 10 Rome,
Italy.
9. Farmer III, J. J., M. A. Asbury, F. W. Hickman, D. J. Brenner, and the
Enterobacteriaceae study group. 1980. Enterobacter sakazakii: A new species of
"Enterobacteriaceae" isolated from clinical specimens. Int. J. Syst. Bacteriol.
30:569-584.
10. Flynn, O. M. J., I. Blair, and D. McDowell. 1992. The efficiency and consumer
operation of domestic refrigerators. Int. J. Refrig. 15:307-312.
11. Heuvelink, A. E., M. Ahmed, F. D. Kodde, J. T. M. Zwartkruis-Nahuis, and E. de
Boer. 2002. Enterobacter sakazakii in melkpoeder. De Ware(n)chemicus. 32:17-
30.
12. Iversen, C., M. Lane, and S. J. Forsythe. 2004. The growth profile,
thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant
formula milk. Lett. Appl. Microbiol. 38:378-382.
13. Iversen, C., A. Lehner, N. Mullane, E. Bidlas, I. Cleenwerck, J. Marugg, S.
Fanning, R. Stephan, and H. Joosten. 2007. The taxonomy of Enterobacter
sakazakii : proposal of a new genus Cronobacter gen. nov. and descriptions of
Cronobacter sakazakii comb. nov. Cronobacter sakazakii subsp sakazakii, comb.
nov., Cronobacter sakazakii subsp malonaticus subsp nov., Cronobacter turicensis
sp nov., Cronobacter muytjensii sp nov., Cronobacter dublinensis sp nov and
Cronobacter genomospecies I. BMC Evol. Biol. 7:64.
14. James, S. J., and J. Evans. 1992. Consumer handling of chilled foods: Temperature
performance. Int. J. Refrig. 15:299-306.
15. James, S. J., and J. Evans. 1992. The temperature performance of domestic
refrigerators. Int. J. Refrig. 15:313-319.
16. Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004. Occurrence of Enterobacter sakazakii in the food production
environments and households. The Lancet. 363:39-40.
Growth during cooling

163
17. Kandhai, M. C., M. W. Reij, C. Grognou, M. Van Schothorst, L. G. M. Gorris, and
M. H. Zwietering. 2006. Effects of preculturing conditions on lag time and
specific growth rate of Enterobacter sakazakii in reconstituted powdered infant
formula. Appl. Environ. Microbiol. 72:2721-2729.
18. Laguerre, O., S. Ben Amara, and D. Flick. 2005. Experimental study of heat
transfer by natural convection in a closed cavity: application in a domestic
refrigerator. J. Food Eng. 70:523-537.
19. Lai, K. K. 2001. Enterobacter sakazakii infections among neonates, infants,
children, and adults - Case reports and a review of the literature. Medicine.
80:113-122.
20. Leclercq, A., C. Wanegue, and P. Baylac. 2002. Comparison of Fecal Coliform
Agar and Violet Red Bile Lactose Agar for Fecal Coliform Enumeration in Foods.
Appl. Environ. Microbiol. 68:1631-1638.
21. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
22. Muytjens, H. L., H. C. Zanen, H. J. Sonderkamp, L. A. Kollee, I. K. Wachsmuth,
and J. J. Farmer, 3rd. 1983. Analysis of eight cases of neonatal meningitis and
sepsis due to Enterobacter sakazakii. J. Clin. Microbiol. 18:115-120.
23. Nazarowec-White, M., and J. M. Farber. 1997. Incidence, survival, and growth of
Enterobacter sakazakii in infant formula. J. Food Prot. 60:226-230.
24. Pagotto, F. J., M. Nazarowec White, A. Bidawid, and J. M. Farber. 2003.
Enterobacter sakazakii: infectivity and enterotoxin production in vitro and in vivo.
J. Food Prot. 66:370-375.
25. Paoli, G., and E. Hartnett. 2006. Overview of a risk assessment model for
Enterobacter sakazakii in powdered infant formula. Prepared for WHO/FAO.
Available at: www.mramodels.org, Accessed 1 Augustus 2009.
26. Reij, M. W., and M. H. Zwietering. 2008. Integrating concepts: a case study using
Enterobacter sakazakii in infant formula. p. 177-204. In: D.W. Schaffner (Ed.),
Microbial Risk Analysis ASM Press, Washington DC, New Brunswick, New
Jersey.
27. Rosset, P., V. Noel, and E. Morelli. 2007. Time-temperature profiles of infant milk
formula in hospitals and analysis of Enterobacter sakazakii growth. Food Control.
18:1412-1418.
28. Rosso, L., J. R. Lobry, S. Bajard, and J. P. Flandrois. 1995. Convenient model to
describe the combined effects of temperature and pH on microbial growth. Appl.
Environ. Microbiol. 61:610-616.
29. Smeets, L. C., A. Voss, H. L. Muytjens, J. F. G. M. Meis, and W. J. G. Melchers.
1998. Genetische karakterisatie van Enterobacter sakazakii isolaten van
Chapter 7
164
Nederlandse patiënten met neonatale meningitis. Ned. Tijdschr. Med. Microbiol.
6:113-115.
30. Spells, K. E. 1960. The thermal conductivities of some biological fluids. Phys.
Med. Biol. 5:139-153.
31. Terpstra, M. J., L. P. A. Steenbekkers, N. C. M. De Maertelaere, and S. Nijhuis.
2005. Food storage and disposal: Consumer practices and knowledge. Brit. Food
J. 107:526-533.
32. Urmenyi, A. M. C., and M. D. Berne. 1961. Neonatal death from pigmented
coliform infection. The Lancet. 1:313-315.
33. WHO. 2007. Safe preparation, storage and handling of powdered infant formula;
guidelines. Available at
http://www.who.int/foodsafety/publications/micro/pif_guidelines.pdf. Accessed 13
February, 2009.
34. Zwietering, M. H., J. T. de Koos, B. E. Hasenack, J. C. de Wit, and K. Van 't Riet.
1991. Modeling of bacterial growth as a function of temperature. Appl. Environ.
Microbiol. 57:1094-1101.

Inactivation of Cronobacter spp.

165
Chapter 8
166
Abstr act
The aim of the study was to determine the survival of two strains of Cronobacter
(Enterobacter sakazakii) and six other bacterial strains inoculated into dry powdered infant
formula (PIF) stored for 22 weeks at several temperatures between 7 and 42 ºC. The
experimental setup involved a relatively high initial concentration of bacteria, around 10
4

CFU/g of powder, and enumeration of survivors with a minimum detection level of 100
CFU/g. For all strains tested, it was found that the number of bacterial cells decreased
faster with increasing temperature. Cronobacter spp. cells generally survived better at high
temperatures (37 and 42 ºC) than the other bacteria, while such a difference in survival
was not apparent at other temperatures. To describe the effect of temperature on survival,
both the Weibull distribution model and the log-linear model were tested. At 22 °C, decline
rates of 0.011 and 0.008 log units/day were found for Cronobacter sakazakii ATCC 29544
and Cronobacter strain MC 10, respectively. Assuming a linear relationship between log
transformed D-values and temperature, z-values estimated for C. sakazakii ATCC 29544
and Cronobacter MC10, were 13.3 and 23.5 ºC, respectively. Such differences found in
resistance among Cronobacter spp. would be relevant to consider when establishing
quantitative risk assessments on consumer risks related to PIF.




This Chapter has been accepted for publication as
“Inactivation r ates of Cronobacter spp. and selected other bacter ial str ains in
powder ed infant for mulae stor ed at differ ent temper atur es”

Kandhai, M.C., M.W. Reij, M. Van Schothorst, L.G.M. Gorris,

and M.H. Zwietering.
Journal of Food Protection, 2010.
Inactivation of Cronobacter spp.

167
Intr oduction
The International Commission on Microbiological Specifications for Foods (ICMSF) has
placed Enterobacter sakazakii in the category of “severe hazards for restricted population”,
being rarely involved in food-borne disease among neonates (14). Over the past fifty years,
an overall mortality rate of 19% was estimated, based on data presented in the Food and
Agricultural Organization and the World Helath Organization report (11). Only recently,
this Gram-negative micro-organism has been reclassified (16, 17) as 6 species in a new
genus, Cronobacter gen. nov., within the family of Enterobacteriaceae. Five of the new
species are Cronobacter sakazakii, C. turicensis, C. malonaticus, C. muytjensii and C.
dublinensis. The sixth species is referred to as genomospecies I and currently includes two
representative strains. The collective of species originally known as E. sakazakii, can be
referred to as Cronobacter spp. The designation Cronobacter spp. is used consistently in
the subsequent part of this paper when the micro-organisms are referred to in general terms
or when historic information that is not specific to the various newly recognized species is
cited.
Cronobacter spp. are opportunistic pathogens that have caused illness particularly in
premature and neonatal infants. In several cases, the source of infection has been identified
as powdered infant formula (PIF) (4, 13, 26). Levels of Cronobacter spp. in PIF are thought
to be low, when present at all, with the mean level in contaminated product possibly being
in the order of 10
-3
CFU/g (8, 10). Such levels, however, reflect operations that lack
stringent hygiene measures as reductions ranging from about 10
-4
to 10
-6
CFU/g may be
possible in adequately hygienic operations (5, 11). The prevalence of these and other micro-
organisms in PIF may depend on many factors, including factors related to manufacturing
process, choice of ingredients and postprocess handling. Prevalence figures for
Cronobacter spp. in PIF are scarce, but may differ widely as they have been reported to be
about 1% and 3% (data quoted by Food and Agricultural Organization/World Health
Organization in 2008) (11), 6.7% (29) and less recently 14% (22). Cronobacter spp. has not
been consistently isolated from PIF. As an example in one study it could not be isolated in
milk- or soy based infant formulae or the infant herb drinks (24).
Chapter 8
168
Cronobacter spp. are vegetative micro-organisms that are not particularly heat resistant
(23). Reconstitution with water of 70 ºC before final use of the dry PIF will decrease the
number of micro-organisms by at least 100,000 fold, rendering the final product safe (10).
It has been reported that the micro-organism will not survive pasteurization treatments
normally given during manufacture of wet-processed infant formulae (15, 23). However,
recent research had shown that several isolates of Cronobacter spp. were able to survive
spray drying, a process that is not intended to be a pasteurization treatment (1). Though PIF
manufacturing is advocated to be undertaken respecting high hygiene conditions and
keeping the environment essentially dry, it is important to note that PIF is not manufactured
to be a sterile product. Contamination after pasteurization can occur through at least two
routes. One such route is via heat-sensitive ingredients, which may have to be added after
pasteurization and may introduce Cronobacter spp. into the final product when they are
present in the ingredients. A second route is environmental contamination. In the case
where micro-organisms are present in the factory environment, recontamination after
pasteurization may occur from such environmental sources (5). The presence of the micro-
organisms in the factory environment may pose a special problem where factory
environments become wet, for instance during wet cleaning, allowing Cronobacter spp. to
multiply rapidly (18). Because the temperature in the factory environments may allow the
micro-organisms to proliferate in residues to relatively high numbers in a short time, the
rate of survival in dried-up material is a key factor determining the extent of
recontamination possible during PIF manufacturing. Quite long survival of Cronobacter
spp. in dry residues and powder has been reported by several authors (3, 6, 12), but the
dynamics have not yet been described mathematically. Conceivably, low water activity (a
w
)
is an important determinant of the survival of Cronobacter spp. in dry factory environments
and the same may hold true for the final product during storage and retail. Unfortunately,
the impact of a
w
on the growth potential of these micro-organisms under realistic conditions
has not been reported in great detail.
In this study, the survival rates of two strains of Cronobacter spp. was determined in PIF
stored at a number of different temperatures and compared to those of six other micro-
organisms. The specific aim was to quantify key parameters that express the inactivation
Inactivation of Cronobacter spp.

169
rate of Cronobacter spp. due to desiccation in PIF at low a
w
at several different
temperatures. Such parameters are considered important for microbiological risk
assessments and for product/process design studies. Formula manufacturers may be able to
use the data in establishing adequate Good Hygienic Practices (GHP) and Hazard Analysis
Critical Control Point (HACCP) plans.

Mater ials and Methods
Preparation of the inoculum
The following strains, Cronobacter sakazakii ATCC 29544 (Enterobacter sakazakii),
Cronobacter strain MC10 (Enterobacter sakazakii MC10), Escherichia coli O2K- ec0001,
Enterobacter cloacae eb0001, Klebsiella pneumoniae kl0001, Salmonella serovar
Enteritidis (phage type 4), Bacillus cereus ATCC 14579 and Staphylococcus aureus ATCC
14458 SEB were used. The stock cultures were stored at -20 ºC in cryo vials (Greiner Bio-
one GmbH, Frickenhausen, Germany), containing 0.3 ml 87% glycerol (Fluka-Chemica
GmbH, Buchs, Switzerland) and 0.7 ml of the bacterial culture. Cronobacter MC10 is a
clinical isolate, kindly provided by Dr. Harry Muytjens (University Medical Center St.
Radboud, Nijmegen, The Netherlands). Bacillus cereus ATCC 14579 is an air isolate, St.
aureus ATCC 14458 SEB is a human feces isolate, while the other strains were food
isolates. These strains were obtained from the culture collection of the Laboratory of Food
Microbiology (Wageningen University, Wageningen, The Netherlands). All the strains
were pre-cultured by adding 100 ul of the stock to 100 ml of Brain Heart Infusion Broth
(BHI; Becton Dickinson and Co., Le Pont de Claix, France) and incubating at 37 °C for 22
hours to obtain a mature culture. Grown cells were centrifuged for 10 min at 20 °C at 2958
x g (MSE, Mistral 3000i, Leicester, United Kingdom) and washed twice in 40 ml of 0.85%
physiological salt solution. The obtained bacterial suspension with a final concentration of
around 10
7
CFU/ml was transferred into tubes of a perfume sprayer (Designed by Gérard
Brinard, DA Drogisterij, Leusden, The Netherlands) to artificially contaminate PIF (18).

Chapter 8
170
Inoculation of the powdered infant formula
Powdered infant formula for infants from 0 to 6 months was bought locally and aseptically
divided in sterilized square boxes (Greiner bio-one GmbH, Alphen aan de Rijn, The
Netherlands), each containing 20 grams. Powdered infant formula was sprayed six times,
using a perfume sprayer with an individual bacterial suspension and rigorously shaken,
resulting in a final concentration of bacterial cells in the dry powder between 10
4
and 10
5

CFU/g. The a
w
of the contaminated powder was determined using a water activity (a
w
)
meter (Novasina Aw-box, Pedak, Heythuysen, The Netherlands). The contaminated powder
was divided over small plastic containers and placed in desiccators containing saturated
lithium chloride (LiCl; Normapur, VWR International Prolab, Fontenay sous Bois France),
which has a constant a
w
of 0.11. The a
w
of the powdered infant formula measured
immediately after spraying with the bacterial suspension remained below 0.3. The
desiccators with the inoculated PIF were incubated at the following different temperatures:
7, 15, 22, 30, 37, and 42 °C for 22 weeks.
The survival of the micro-organisms during storage at 7, 15, 22 and 30 °C was determined
by sampling the contaminated PIF once every two weeks, while contaminated powder
stored at 37 and 42 ºC was sampled once every week. As the decline in the numbers of
bacteria was found to progress very rapidly at 37 and 42 ºC and very few data-points
reflecting this decline were obtained, an additional experiment was conducted with five of
the strains of bacteria (i.e. Cronobacter sakazakii ATCC 29544, Cronobacter MC10, E.
coli, E. cloacae, and St. aureus) with storage at 37 ºC. In this additional experiment,
samples were taken every day, except on week-end days, for 16 days.
Samples of 1 g of PIF were reconstituted in 9 ml peptone saline solution (NaCl, 8.5 g/L,
and neutralized bacteriological peptone, 1 g/L; Oxoid Ltd, Basingstoke, England), and
appropriate dilutions were surface plated in duplicate onto tryptone soya agar (TSA; Oxoid
Ltd., Basingstoke, England) by using a spiral plater device (Eddy Jet, IUL instruments, IKS
BV, Leerdam, The Netherlands). Inoculated plates were incubated for 20 to 24 h at 37 ºC
and typical colonies were counted by manual enumeration, the detection limit was 100
CFU/g.
Inactivation of Cronobacter spp.

171
Determination of the D-value
To describe the rate of survival of the two strains of Cronobacter spp. and the other
bacterial strains evaluated over time, both the Weibull distribution model (Equation 1) and
the log-linear function (Equation 2) were applied to the bacterial counts and used to
determine the decimal reduction time D
w
and D
l
(20, 28).

( ) ( )
p
w
D
t
N
t
N
|
|
.
|

\
|
÷ =
0
log log (1)
( ) ( )
|
|
.
|

\
|
÷ =
l
D
t
N
t
N
0
log log (2)
where t [days] is the time, N(t) [CFU/g] is the bacterial concentration at time t, N
0
is the initial
bacterial concentration at time t = 0, the start of the incubation period. D [days] is the time of the first
decimal reduction (time required to reduce the population by 1 log unit from the initial level at t = 0)
as obtained by the log-linear model (D
l
) and the Weibull distribution model (D
w
).

Parameter p [-] (p > 0) in equation 1 has no direct biological relevance; it is a parameter
that describes the shape of the inactivation curve. If p = 1 then equation 1 is identical to
equation 2 and the decline of the Log count is linear (Equation 2).
In order to obtain reliable estimates for D
w
and D
l
parameters, the experimental data had to
meet the following requirements:
1) a minimum of three data points were required,
2) at least two data points had to be above the detection limit of 100 CFU/g.
Survival data that did not meet these requirements were excluded from fitting using the
Weibull model.
All parameters in the Weibull distribution model and the log-linear model were obtained
via minimizing the residual sum of squares (RSS) using the solver option in Microsoft
Office Excel 2003 and were verified using Table curve 2D, for Windows 2.03. (Jandel
Scientific, Erkrath, Germany). The goodness of fit of both the Weibull model and the log-
Chapter 8
172
linear model was assessed by calculating the mean square error (MSE). This is done by
dividing the RSS value by the degrees of freedom (DF). In addition, for each p- value the
95% confidence interval was calculated; when p = 1 falls within the 95% confidence
interval, then the log-linear model was also suitable to describe the data.

Calculation of the z-value
Traditionally, it has been assumed that the temperature dependence of the D-value versus
the temperature is log linear and can be expressed with the z-value. The z-value is the
increase in temperature (ºC) that is needed to reduce or extend the D-value by a factor of
ten. The estimated D-values were therefore log transformed in order to obtain a linear
relationship with temperature (27). Equation 3 was used to estimate the z-values.

( ) ( )
|
|
.
|

\
| ÷
÷ =
z
T T
D D
ref
ref refmodel
log log
(3)
where log (D
T
) [log days], is the logarithm of the D-value [days], log (D
ref model
) is the log D
ref
value
at T
ref
, and T
ref
is the reference temperature [ºC].

Results
This study assessed the survival of eight strains of bacteria, i.e. Cronobacter sakazakii
ATCC 29544, Cronobacter MC10, E. coli, E. cloacae, Kl. pneumoniae, Salmonella, B.
cereus and S. aureus, in powdered infant formula during storage for 22 weeks at 7, 15, 22,
30, 37 and 42 ºC. A key parameter determining the potential survival in PIF is water
activity, and this parameter was monitored during storage. After 5 days of storage at 30 ºC,
37 ºC and 42 ºC, the average water activity had dropped to 0.129 (-). The water activity
decreased more slowly at 7, 15 and 22 ºC, reaching values of 0.286, 0.180, and 0.152 [-]
after 5 days storage. Since the water activity varies with temperature, and products were
stored at different temperatures some level of inaccuracy may be associated to the results
reported here.
Inactivation of Cronobacter spp.

173
Figures 1 and 2 show the enumeration results for the various bacterial strains at the various
temperatures over time. Symbols on the detection limit, 100 CFU/g, indicate that the results
were either on or below the detection limit. At room temperature (Figure 1; results at 22 ºC)
the trend for all strains was gradual inactivation, although the data for individual strains
showed a somewhat erratic profile. Notably, the first strain that declined below the
detection limit was B. cereus at 7 weeks of incubation. The other strains were detectable
significantly longer, with S. aureus detectable up to 17 weeks, Kl. pneumoniae and
Salmonella up to 18 weeks and E. coli and E. cloacae up to 20 weeks. After 22 weeks, at
the end of storage, C. sakazakii and Cronobacter MC10 were still detectable, with the latter
at the higher level of the two.
Compared to room temperature storage, the apparent decline at 37 and at 42
o
C progressed
very rapidly for all bacteria investigated, while a somewhat faster decline could be seen at
30
o
C and a somewhat slower decline was noticeable at 15 and 7
o
C. At the two highest
temperatures evaluated (37 and 42 ºC), Cronobacter spp. strains, E. cloacae, E. coli and S.
aureus could be recovered incidentally beyond the first week of storage. All the other
strains tested were below the detection limit of 100 CFU/g already after one week. Because
of the limited number of data points obtained in this experiment, decline rates at 42 ºC
could only be estimated for C. sakazakii and Cronobacter MC10 and these were calculated
to be 0.07 and 0.08 log units/days, respectively.
In order to gather more data points at 37
o
C, which would allow to more accurately quantify
the decline in cell numbers during the initial incubation period, a second experiment was
performed with five of the strains, namely C. sakazakii, Cronobacter MC10, E. coli, E.
cloacae, and S. aureus. In this experiment, incubation was for a period of 16 days, with
more frequent monitoring of the levels of bacteria than in the first experiment. The results
obtained are depicted in Figure 2. For all the strains tested in the second experiment, at least
a 2-log CFU/g decrease was found within the 16 days of the incubation period. Solid and
broken lines indicate fitting of the data-points with a nonlinear model (the Weibull model
shown in Equation 1) and a linear model (the log-linear model shown in Equation 2). Data
fitting did include data points on or below the detection limit. Figure 2 indicates that for
several bacteria, inactivation in PIF at 37
o
C shows a rather nonlinear curve.
Chapter 8
174
For the various data obtained in the first and second storage experiment the times for the
first decimal reduction, where possible, were calculated as a reflection of the inactivation
rate (Table 1). To quantify the progressive decline in cell numbers during storage, two
different modeling approaches were used to establish times for the first decimal reduction
(D). Decimal reduction times were estimated by fitting the log-transformed data with either
the Weibull distribution model (Equation 1), giving D values referred to as D
w
, or the log-
linear model (Equation 2), giving D values designated D
l
. In both cases, D expresses the
rate of the first decimal reduction, which for the log-linear model is equal to the rate of
subsequent decimal reductions, whereas for the Weibull model this D
w
value can change
depending on the value for parameter p. With 0 < p <1, a fitted curve shows a decline in
rate over time and thus a slow down of the speed of inactivation. From Table 1 it can be
deduced for all bacteria investigated that, while there were sometimes notable differences
between the absolute values of D
w
and D
l
, they were often in the same order of magnitude
for individual bacteria at specific temperatures but more importantly consistently showed
higher initial inactivation rates with an increase in storage temperature.
Inactivation of Cronobacter spp.

175

Figur e 1. Enumeration of various bacterial strains in powdered infant formula at 42 ºC, 37 ºC, 30 ºC,
22 ºC, 15 ºC and 7 ºC for a period of 22 weeks. + = Cronobacter sakazakii ATCC 29544; ^ =
Cronobacter MC10 ; ▲ = E. cloacae eb0001; ■ = E. coli ec0001; ○ = S. aureus 14458 SEB; ● = B.
cereus ATCC 14579; □ = Kl. pneumoniae kl0001 ; A = Salmonella.
Chapter 8
176

Figur e 2. Survival of various bacterial strains in powdered infant formula at 37 ºC for a period of 16
days, the fit of the Weibull model (
________
) and the log linear regression (----------). + = Cronobacter
sakazakii ATCC 29544 (A); ^ =Cronobacter spp. (B) ; ■ = E. coli ec0001 (C); ▲ = E. cloacae
eb0001 (D); ○ = S. aureus 14458 SEB (E).
Inactivation of Cronobacter spp.

177
Chapter 8
178

Inactivation of Cronobacter spp.

179
Chapter 8
180
Inactivation of Cronobacter spp.

181
Comparing the D
w
-values in Table 1 for the various bacteria, St. aureus did not show a
consistent decrease in numbers at 22 ºC, resulting in a D
w
-value of 169 days, whereas E.
cloacae, B. cereus and Kl. pneumoniae showed very fast decimal reduction (D
w
-value of
22, 14 and 13 days, respectively). The D
w
-values at 22 ºC calculated for the two
Cronobacter strains (78 and 67 days) were comparable with the D
w
-value estimated for the
E. coli strain (D
w
-value of 64 days) and somewhat higher than that of Salmonella (D
w
-value
47 days). For individual micro-organisms, D
w
-values at 22 ºC were generally higher at 7 or
15 ºC, while these values were lower at 30 ºC or above. For the two strains of Cronobacter
spp., D
w
- and D
l
values at 22 ºC were quite different, being 78 days and 95 days,
respectively 67 days and 119 days for C. sakazakii and Cronobacter spp. MC10,
respectively.
The fit of the two models with the data points, expressed in mean square error (MSE)
values, showed that the Weibull model generally provided the better fit as compared to the
log-linear model (Table 1). The estimated values for MSE
w
and MSE
l
,

respectively for the
Weibull model and the log-linear regression model, were similar in many instances, with
better fits often at the lower storage temperatures. In 31 (84%) of the 37 cases (as indicated
in the former last column in Table 1), the Weibull model showed equal or lower MSE
values than the log-linear model meaning that, overall, this model provided for the better fit
of the data, reflecting a rapid first decline followed by successively slower inactivation.
Table 1 shows the values computed with the Weibull model for parameter p (Equation 1).
In the last column an indication is given whether the value for parameter p is equal to 1
within the confidence interval of the data obtained and thus Equation 1 equals Equation 2,
the log-linear model. From the total of 37 cases, the log-linear model is valid 19 times
(51%) on the basis of parameter p. However, generally, differences in MSE values between
models were relatively small and also on the basis of the p parameter in the Weibull model
it appeared that either model might be applicable in certain cases.
In an effort to characterize the effect of temperature on the decline of bacteria in PIF in
quantitative terms, the log transformed first decimal reduction times calculated with the
Weibull model (D
w
-values), were plotted against the incubation temperature (Figure 3).
Chapter 8
182
D
w
values rather than D
l
values were chosen as the Weibull model fitted the data generally
better than the log-linear model. Assuming a linear relation for the temperature impact on
survival, z-values were calculated with Equation 3 using 22
o
C as the reference temperature.

Figur e 3. The log transformed estimated first decimal reduction time (D
w
-values) for several micro-
organisms in powdered infant formula at various temperatures. + = Cronobacter sakazakii ATCC
29544 ; ^ = Cronobacter MC10 ; ▲ = E. cloacae eb0001; ■ = E. coli ec0001; ○ = S. aureus 14458
SEB; ● = B. cereus ATCC 14579; □ = Kl. pneumoniae kl0001; A = Salmonella.

Table 2 presents the calculated z-values and the log D
ref model
values for the various bacterial
strains investigated. From these data the inactivation rate of strain C. sakazakii ATCC
29544 (z=13.3 °C) seemed to be the most temperature depended while the inactivation rate
of Kl. pneumoniae was the least temperature dependent (z = 43.3 °C) during incubation in
powdered infant formula. Estimated z-values for Cronobacter MC10, B. cereus, E. cloacae
and E. coli were all very close together (range: 23.3 to 25.2 °C) while the z-value for St.
aureus was relatively low (19.2 °C) and that of Salmonella relatively high (28.2 °C).


Inactivation of Cronobacter spp.

183
Table 2. Estimates of the z-values for Cronobacter sakazakii ATCC 29544; Cronobacter MC10; E.
cloacae eb0001; E. coli ec0001; S. aureus 14458 SEB; B. cereus ATCC 14579; Kl. pneumoniae
kl0001; Salmonella at 22 ºC
Str ain z-value [ºC] Log (D
r ef model
[days])
a

C. sakazakii ATCC 29544 13.3 1.77
Cronobacter MC10 23.5 1.59
Salmonella 28.2 1.58
B. cereus ATCC 14579 23.5 1.23
Kl. pneumoniae kl0001 43.3 1.40
S. aureus 14458 SEB 19.2 1.81
E. cloacae eb0001 23.3 1.25
E. coli ec0001 25.2 1.59
a
D
ref
with T
ref
22 ºC.

Discussion
In this study we investigated the dynamics of several different bacteria possibly associated
with powdered infant formula including two strains of Cronobacter spp., the genus that has
been identified as an opportunistic pathogen potentially present in PIF and a risk factor for
specific vulnerable consumers. PIF is not designed and processed to be a sterile product and
as a consequence storage, handling and preparation before consumption have to take
account of the presence of potentially harmful bacteria. Though much has been done in
manufacturing operations in terms of process optimization and control of hygiene, at
present it is not possible to totally eliminate Enterobacteriaceae, including Cronobacter
spp. from the processing environment in practice (10, 21). In effect, bacteria may be present
in PIF after manufacture and packaging. However, adherence to proper hygiene and
manufacturing practices tailored to PIF production should ensure levels of, for instance,
Cronobacter spp. to be below 10
-4
CFU/g (5). Levels of Cronobacter spp. in contaminated
PIF have been reported, in three surveys, in the range of 0.002 to 0.92 CFU/g (8). Whether
and how the level of such bacteria changes from the manufacturing phase, during storage
Chapter 8
184
up to consumer use, may depend on factors such as the length of storage and storage
temperature.
In this study it was found for all bacteria, that levels in PIF declined over time and that the
rate of decline was influenced by storage temperature (Figure 1). At temperatures of 37 and
42
o
C, reductions were generally very rapid, but then these temperatures are usually not
temperatures at which PIF is stored in practice. However, such temperatures (and higher)
are not uncommon in environments surrounding spray-dryers. At 30
o
C, the rate of decline
was found to be somewhat more rapid that at 22
o
C. The latter is a realistic storage
temperature generally, while the former could be a valid storage temperature in countries
with a tropical climate. Decline at 7 and 15
o
C progressed more slowly than at 22
o
C. In
terms of achieving a reduction in the level of potentially hazardous micro-organisms
remaining in PIF when present from the manufacturing phase, storage at room temperature
or above may thus have an advantageous effect over cold storage. However, there may be a
detrimental impact on product quality and/or nutritional properties of high temperature
storage.
For the various bacteria tested, the estimated values for D-values were found to decrease
with increasing temperature (Table 1) putting the observations made regarding Figures 1
and 2 on a quantitative basis. Consistently, higher storage temperatures seem to cause or
contribute to a progressively rapid die-off of micro-organisms. At 37 and 42
o
C, loss of
viability is particularly rapid compared to storage at 22
o
C. The underlying mechanism for
this may relate to metabolic exhaustion at elevated temperatures as reported before (19).
Vegetative cells tend to initiate every possible repair mechanism to overcome the impact of
temperature stress on metabolism and cell integrity, exhausting their energy reserves, as
well as genetic and structural building blocks. Notably, rather the low water activity of the
product would render the micro-organisms metabolically inactive within a short space of
time. Comparing the five strains of which the decline rate could be derived at 37
o
C (St.
aureus, E. cloacae, E. coli and both Cronobacter spp. strains), it seems that the D
w
-value of
the two Cronobacter spp. strains were slightly lower (2 days at 37 ºC) than those of the
others (3 to 5 days). For C. sakazakii and Cronobacter MC10, the estimated D-values at 37
ºC were 2 days (for D
w
) and 7 and 12 days (for D
l
), respectively. D-values between 30 ºC
Inactivation of Cronobacter spp.

185
and 7 ºC ranged between 49 and 713 days and between 49 and 332 days for C. sakazakii
and Cronobacter MC10, respectively. As indicated by the 95% confidence interval data for
the D
w
- and D
l
-values in Table 1, these ranges occasionally varied considerably. Also the
D
w
- and D
l
-values estimated for the other six bacteria at the six different incubation
temperatures evaluated showed significant individual variation. Nevertheless, the data
indicated that at 22 and 30
o
C, temperatures that can be room temperature in practice,
progressive inactivation of most bacteria may take place in the course of storage. At 7 and
15
o
C, overall, inactivation was limited to about 2 log cycles for the various bacteria.
Notably, at these lower temperatures, the two strains of Cronobacter spp. remained viable
and well over the detection limit of the experimental set-up for the whole 22 weeks storage
period. The fact that these and other bacteria tested here survived under the dry conditions
prevailing in PIF, indicates that the micro-organisms posses a certain level of resistance or
tolerance to dryness. Many other studies reported resistance to low a
w
(2, 3, 6, 12, 15), and
together these findings underscore that that survival of Cronobacter spp. in PIF is a matter
to be carefully considered in assessing the safe preparation of infant formula for final
consumption.
Previous studies on the inactivation of Cronobacter spp. in dry powdered infant formulae
only considered storage at room temperature and expressed the decline as a rate in log units
per day. To compare the current findings with those of previous studies, the decimal
reduction times found at 22
o
C were converted to decline rates (log units/day) by taking the
reciprocal of the D-values. Thus, decline rates observed in this study at room temperature
were found to be 0.011 and 0.013 log units per day for C. sakazakii and 0.008 and 0.015 log
units per day for Cronobacter MC10, as calculated from the D
l
- and D
w
-values,
respectively. These values compare well with the decline rate of 0.014 log units per day
reported in literature (7, 10). However, these literature values were computed considering
the inactivation curves to be biphasic. While the first rapid phase of inactivation was
characterized by a decline rate of 0.014 log units per day, the second phase showed a much
lower decline rate for Cronobacter spp. of 0.001 log units per day (10). However, as seen
frequently in the current study for this bacterium and several others investigated, decline
rates may gradually slow down in the course of storage, a (so-called) tailing effect that was
Chapter 8
186
most pronounced at low storage temperatures (7 and 15 ºC). Rather than choosing one
decline rate, e.g. the most rapid as the best case or the slowest as the most conservative
case, it would be more realistic to consider the actual dynamic change in decline rate over
storage in predictions, for instance as part of a microbiological risk assessment study. The
data obtained in the current investigation show that a dynamic change in decline rate can be
reflected well by using the Weibull model. In Table 3, our findings considering the impact
of storage temperature on the decline of Cronobacter spp. in PIF are projected as times
required for a certain decimal reduction using both the Weibull and the log-linear model.
The projected times are in months and are best interpreted considering that the maximum
shelf-life of PIF is about 30 months, though it is likely that in most cases PIF is consumed
within a shorter time frame (10). Comparing times predicted using the Weibull model, as
this generally fitted the dynamic behavior of most bacteria evaluated here best, it can be
seen that at 7
o
C, an 1-log inactivation in the level of Cronobacter spp. may occur within 3
to 24 months of storage. At 22
o
C, a 1-log reduction may occur within 2 to 3 months while
a 3-log reduction may take place in 11 to 25 months. At elevated temperatures, reduction is
faster than at room temperature, while under refrigeration, survival will exceed the
maximum shelf life of the product. Interpretation of these calculations should take into
account the significant variation in the estimated D-values determined in this study and that
samples were taken during a 22-week period. Extrapolations beyond that time frame have
been indicated in parentheses in Table 3 and should be interpretated as orders of magnitude
only.

Inactivation of Cronobacter spp.

187
Chapter 8
188
Since inactivation kinetics data have been accumulated for several different temperatures in
the present investigation, it has been possible to estimate z-values which can be used to
relate inactivation kinetics between temperatures. Considering the better fit of the Weibull
distribution model with experimental data, z-values for the various bacteria were
established on the basis of the D
w
-values at different storage temperatures. As shown in
Table 2, z-value of 13.3 ºC and 23.5 ºC were estimated for C. sakazakii and Cronobacter
spp. MC10, respectively. It should be noted that z-values reported in this paper are based on
D
w
-values as estimated in dry PIF and do not apply to earlier reported thermal inactivation
in liquid (25). Thus the given D/z-concept is specific for inactivation in dry PIF at moderate
temperatures.

In conclusion, quantification of survival rates of Cronobacter spp. in dry PIF may provide
useful data for microbiological risk assessments and for product/process design studies, as
it allows determining likely levels of the micro-organism in PIF subjected to different
scenarios of time and temperature during storage and distribution. Our findings suggest that
there may be wide differences in the inactivation rates of Cronobacter spp. in dessicated
PIF, Cronobacter cells are rapidly inactivated in dry powder at temperatures of 37 °C and
42 °C, but that inactivation rates at ambient temperatures or lower tend to progressively
slow-down. In most cases, however, inactivation rates may be higher than the rather slow
decline rate of 0.001 log units per day assumed in the Food and Agricultural Organization/
World helath Organization risk assessment study (11). The Weibull model in conjunction
with the inactivation data generated in this study, may be useful for reflecting the dynamic
reduction of Cronobacter spp. in PIF, for instance, in microbiological risk assessments.
Formula manufacturers may be able to use the model and other results in establishing
adequate Good Hygienic Practices (GHP) and Hazard Analysis Critical Control Point
(HACCP) plans. Several observations may need to be considered in the design of Good
Manufacturing Practices (GMP) and HACCP systems for PIF manufacture, including the
impact of storage at temperatures over ambient on Cronobacter spp. inactivation, D- and z-
values and the quantitative substantiation of the very long survival time of Cronobacter
spp. in PIF. Further work may need to address interstrain differences of Cronobacter spp. in
Inactivation of Cronobacter spp.

189
their ability to survive dry environments. Our data, collected from two clinical isolates in
this study, should not be assumed to represent a worst-case scenario of the most
desiccation-resistant strains that may be present in the environment.

Acknowledgements
The financial support by Nestlé Research Center, Lausanne, Switzerland, is gratefully
acknowledged. The authors thank Samia Amrouche for carrying out part of the experiments
and Dr. H.L. Muytjens for providing the Cronobacter strain MC10.

Refer ences
1. Arku, B., N. Mullane, E. Fox, S. A. Fanning, and K. Jordan. 2008. Enterobacter
sakazakii survives spray drying. Int. J. Dairy Technol. 61:102-108.
2. Breeuwer, P., A. Lardeau, M. Peterz, and H. Joosten. 2003. Desiccation and heat
tolerance of Enterobacter sakazakii. J. Appl. Microbiol. 95:967-973.
3. Caubilla-Barron, J., and S. J. Forsythe. 2007. Dry stress and survival time of
Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered
infant formula. J. Food Prot. 70:2111-2117.
4. Clark, N. C., B. C. Hill, C. M. O'Hara, O. Steingrimsson, and R. C. Cooksey.
1990. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial
outbreaks. Diagn. Microbiol. Infect. Dis. 13:467-72.
5. Cordier, J. L. 2008. Production of powdered infant formulae and microbiological
control measures. p. 145-185. In: J. M. Farber, and S. J. Forsythe (eds.),
Enterobacter sakazakii (emerging issues in food safety), ASM Press, Washington,
D.C., New Brunswick, New Jersey.
6. Edelson-Mammel, S. G., and R. L. Buchanan. 2005. Survival of Enterobacter
sakazakii in a dehydrated powdered infant formula. J. Food Prot. 68:1900-1902.
7. Edelson-Mammel, S. G., and R. L. Buchanan. 2004. Thermal inactivation of
Enterobacter sakazakii in rehydrated infant formula. J. Food Prot. 67:60-63.
8. Fanning, S., and S. J. Forsythe. 2008. Isolation and identification of Enterobacter
sakazakii. p. 27-59. In: J. M. Farber, and S. J. Forsythe (eds.), Enterobacter
sakazakii (emerging issues in food safety), ASM Press, Washington, D.C., New
Brunswick, New Jersey.
Chapter 8
190
9. FAO/WHO. 2004. Enterobacter sakazakii and other micro-organisms in powdered
infant formula: meeting report. Microbiological Risk Assessment Series No. 6,
Rome, Italy.
10. FAO/WHO. 2006. Enterobacter sakazakii and Salmonella in powdered infant
formula: meeting report. Microbiological Risk Assessment Series No. 10 Rome,
Italy.
11. FAO/WHO. 2008. Enterobacter sakazakii (Cronobacter spp.) in powdered follow-
up formulae: meeting report. Microbiological Risk Assessment Series No. 15
Rome, Italy.
12. Gurtler, J. B., and L. R. Beuchat. 2007. Survival of Enterobacter sakazakii in
powdered infant formula as affected by composition, water activity, and
temperature. J. Food Prot. 70:1579-1586.
13. Himelright, I., E. Harris, V. Lorch, and M. Anderson. 2002. Enterobacter
sakazakii infections associated with the use of powdered infant formula-
Tennessee, 2001. JAMA. 287:2204-2205.
14. ICMSF. 2002. Micro-organisms in Foods 7: Microbiological Testing in Food
Safety Management. Springer Publishers, New York.
15. Iversen, C., M. Lane, and S. J. Forsythe. 2004. The growth profile,
thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant
formula milk. Lett. Appl. Microbiol. 38:378-382.
16. Iversen, C., A. Lehner, N. Mullane, E. Bidlas, I. Cleenwerck, J. Marugg, S.
Fanning, R. Stephan, and H. Joosten. 2007. The taxonomy of Enterobacter
sakazakii: proposal of a new genus Cronobacter gen. nov. and descriptions of
Cronobacter sakazakii comb. nov. Cronobacter sakazakii subsp sakazakii, comb.
nov., Cronobacter sakazakii subsp malonaticus subsp nov., Cronobacter turicensis
sp nov., Cronobacter muytjensii sp nov., Cronobacter dublinensis sp nov and
Cronobacter genomospecies I. BMC Evol Biol. 7:64.
17. Iversen, C., N. Mullane, B. McCardel, B. D. Tal, A. Lehner, S. Fannin, R.
Stephan, and H. Joosten. 2008. Cronobacter gen. nov., a new genus to
accommodate the biogroups of Enterobacter sakazakii, and proposal of
Cronobacter sakazakii gen. nov., comb. nov., Cronobacter malonaticus sp. nov.,
Cronobacter turicensis sp. nov., Cronobacter muytjensii sp. nov., Cronobacter
dublinensis sp. nov., Cronobacter genomospecies 1, and of three subspecies,
Cronobacter dublinensis subsp. Dublinensis subsp. nov., Cronobacter dublinensis
subsp. lausannensis subsp. nov. and Cronobacter dublinensis subsp. lactaridi
subsp. nov. . Int. J. Syst. Evol. Microbiol. 58:1442-1447.
18. Kandhai, M. C., M. W. Reij, C. Grognou, M. Van Schothorst, L. G. M. Gorris, and
M. H. Zwietering. 2006. Effects of preculturing conditions on lag time and
specific growth rate of Enterobacter sakazakii in reconstituted powdered infant
formula. Appl. Environ. Microbiol. 72:2721-2729.
Inactivation of Cronobacter spp.

191
19. Leistner, L., and L. G. M. Gorris. 1995. Food preservation by hurdle technology.
Trends Food Sci. Technol. 6:41-46.
20. Mafart, P., O. Couvert, S. Gaillard, and I. Leguerinel. 2002. On calculating
sterility in thermal preservation methods: application of the Weibull frequency
distribution model. Int. J. Food Microbiol. 72:107-113.
21. Mullane, N. R., P. Whyte, P. G. Wall, T. Quinn, and S. Fanning. 2007.
Application of pulsed-field gel electrophoresis to characterise and trace the
prevalence of Enterobacter sakazakii in an infant formula processing facility. Int.
J. Food Microbiol. 116:73-81.
22. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
23. Nazarowec-White, M., and J. M. Farber. 1997. Thermal resistance of Enterobacter
sakazakii in reconstituted dried-infant formula. Lett. Appl. Microbiol. 24:9-13.
24. O' Brien, S., B. Healy, C. Negredo, W. Anderson, S. Fanning, and C. Iversen.
2009. Prevalence of Cronobacter species (Enterobacter sakazakii) in follow-on
infant formulae and infant drinks. Lett. Appl. Microbiol. 48:536-541.
25. Shaker, R. R., T. M. Osaili, A. S. Abu Al-Hasan, M. M. Ayyash, and S. J.
Forsythe. 2008. Effect of desiccation, starvation, heat, and cold stresses on the
thermal resistance of Enterobacter sakazakii in rehydrated infant milk formula. J.
Food Sci. 73:M354-M359.
26. Van Acker, J., F. de Smet, G. Muyldermans, A. Bougatef, A. Naessens, and S.
Lauwers. 2001. Outbreak of necrotizing enterocolitis associated with Enterobacter
sakazakii in powdered milk formula. J. Clin. Microbiol. 39:293-297.
27. Van Asselt, E. D., and M. H. Zwietering. 2006. A systematic approach to
determine global thermal inactivation parameters for various food pathogens. Int.
J. Food Microbiol. 107:73-82.
28. Van Boekel, M. A. J. S. 2002. On the use of the Weibull model to describe
thermal inactivation of microbial vegetative cells. Int. J. Food Microbiol. 74:139-
159.
29. Yoo, M. K., S. S. Kim, and S. Oh. 2005. Isolation and genotyping of Enterobacter
sakazakii from powdered infant formula manufactured in Korea. J. Food Sci.
Biotechnol. 14:875-877.
Chapter 8
192

Discussion and concluding remarks

193
Chapter 9
194
Intr oduction
Cronobacter spp., previously referred to as Enterobacter sakazakii, is more and more
widely recognized as an opportunistic foodborne pathogen that can potentially cause illness
in vulnerable sub-populations of consumers. Cronobacter infection is a rare illness and no
active surveillance systems exist, though notification of the illness is mandatory in a few
countries. All known species of Cronobacter have been linked retrospectively in either
infants or adults to clinical cases of infection (24). Sub-populations at risk for Cronobacter
infection can be found among infants and elderly consumers (53). As detailed in the
Chapter 1, the number of cases of confirmed Cronobacter infections in children worldwide
is limited (n=156), as is the number of confirmed death (n=29). In this regard, Cronobacter
spp. may not appear to be as important a pathogen as many other foodborne pathogens that
challenge the safety of our food to date, but because it particularly affects very young
infants it is causing a high level of public concern and outrage.
Although the natural source or habitat of Cronobacter spp. is not clearly understood, it is
increasingly apparent that this group of micro-organisms is present in many different
environments and can be found in a wide range of foods. Epidemiological studies have
implicated powdered infant formulae (PIF) as a prime vehicle for possible exposure of
infants to Cronobacter species (1, 77, 82). Since an outbreak of Cronobacter infections
occurred in the United States (Figure 1), there have been recalls of Cronobacter spp.
contaminated infant formulae in several countries around the world (28).
Like many other governmental agencies, the United States Food and Drug Administration
(FDA) issued an alert to health care professionals regarding the risk associated with
Cronobacter infections among neonates fed with milk-based powdered infant formulae
(26). For the last decade there has been a strong and focused effort amongst all relevant
stakeholders (i.e. governments, health case professionals, industry, academia, consumers) to
establish harmonized safety criteria, good manufacturing practices and sound guidelines for
PIF preparation and use in order to protect the vulnerable consumers from the threat of
infection with Cronobacter spp..

Discussion and concluding remarks

195

Figur e 1. Reported outbreak related to Cronobacter spp. in 2002 in the United States (14).

Because the ability to prevent and control Cronobacter spp. is intimately associated to the
ability to detect and identify the micro-organism in foods and food environments, the
progress in methodologies is first discussed. Following that, new insights from the research
reported here as well as from other sources as related to our ability to establish sound
assessments of consumer exposure to the micro-organism, are described. Finally, potential
exposure is simulated for different scenarios of preparation and consumption of infant
formulae, using the results of this research and relevant external data to compile particular
simulations and to draw conclusions relating to prevention and control of the risk to infants
of Cronobacter spp. associated with powdered infant formulae.

Methodology
Before the new systematic categorization and nomenclature as described in Chapter 1,
Cronobacter spp. belonged to the genus Enterobacter that contains species such as E.
cloacae, Pantoea spp. (formerly E. agglomerans), E. aerogenes and E. gergoviae. The
differentiation of Cronobacter spp. among these species is based on biochemical reactions,
Chapter 9
196
and serological and molecular techniques (25). A key feature for Cronobacter spp., as
compared to other Enterobacter species, is the presence of α-glucosidase (60). Based on
this feature and the fact that Cronobacter spp. typically produce yellow pigmented colonies
when grown on tryptone soy agar, early in this research a new practical method for
screening Cronobacter spp. in foods and food environments was developed. This method is
elaborated on in Chapter 2 and it has been used with a range of isolates obtained from
environmental samples taken in several different (food) factories and food products
(Chapter 3 and 4). The biochemical characteristic of α-glucosidase production has remained
the basis for many subsequently developed detection media. Selective media developed
using this feature include “Enterobacter sakazakii Isolation Agar” – ESIA (AES, France),
R&F
®
(E. sakazakii chromogenic plating medium) and Druggan-Forsythe-Iversen-medium
(DFI) (37). Notably, other Enterobacteriaceae species may possess α-glucosidase after 24
hours, but only after induction of the enzyme by a suitable substrate (Chapter 2).
Over the years, methods for the isolation, detection and enumeration of
Enterobacteriaceae, Cronobacter spp. and Salmonella spp.have been developed to various
degrees of specificity, sensitivity and robustness. Many new methods have been introduced
and further refined over the last decade by academics and media manufacturers (Table 1),
in particular methods aiming to detect and trace Cronobacter spp. in powdered infant
formulae.
All methods have a number of steps such as pre-enrichment and (selective) enrichment.
Sometimes a resuscitation step is used for the recovery of stressed cells or cells injured by
exposure to chemical or physical treatments used in food production or processing (79). In
the (selective) enrichment step the specific growth conditions of the target organism(s) are
met while growth of other micro-organisms is inhibited to the greatest extent possible. An
enrichment procedure may be an essential step in enabling detection of low contamination
levels as it can promote the amplification of a single cell to levels of > 10
3
CFU/ml of
enrichment broth. In an enrichment procedure, typically, a 1:10 dilution of the food matrix
is made, thus reducing interferences from the food in the detection assay (9). Reducing the
detection level has been an important aim for method development and currently
Cronobacter spp. can be detected at 1 CFU/100g of PIF, when 100 g is pre-enriched.
Discussion and concluding remarks

197
The initial detection methods established by the US FDA (2002) (78) and ISO (TS 22964)
(36) are currently being re-evaluated, as knowledge regarding Cronobacter further
improves. For example, the reliance on yellow-pigmentation (25) in ISO TS 22964 is no
longer accepted as valid because a high proportion of strains (up to 25%) are non-
pigmented on tryptone soy agar (6). It was also shown that some Cronobacter strains (2%)
(43) are unable to grow in standard Enterobacteriaceae enrichment (EE) broth and,
therefore, would be missed using only the general Enterobacteriaceae test procedures.
Improvements in detection methods for Cronobacter have been focussing much on
optimising the isolation of the micro-organism from powdered infant formulae. Most
published methods to date include a pre-enrichment step, which is followed by a selective
enrichment broth prior to chromogenic agar plating. Currently, the international standard
method for detection of Cronobacter from milk products issued by the International
Organization for Standardization (ISO) and designed by the International Dairy Federation
(IDF) involves a two-step enrichment (36). The selective enrichment method used in this
international standard method has been adapted from the one which has been developed in
the research described in this thesis as described in Chapter 3.
Conventional methods are generally more time consuming than the molecular methods
described more recently (2, 55, 62, 76, 85). However, the conventional methods showed to
be efficient even when low levels of Cronobacter cells were found in a sample (0.36 to 66
CFU/100 g) (59). Iversen et al., (40) showed that current cultural media do not support the
growth of all Cronobacter strains and, therefore, efforts have to be made to improve the
cultural methods such that all strains are captured. Although real-time PCR methods for
specific detection of Cronobacter spp. that target 16S rRNA of the micro-organism are
highly sensitive and useful for rapid and direct detection of Cronobacter spp. in foods, they
are rather qualitative and still rely on a time consuming (pre-)enrichment step. Moreover,
for this method a population of 10
2
CFU/ml is required to obtain a 100% detection
probability (21). Recently, a fluorescence in situ hybridisation (FISH) technique has been
described for the detection of Cronobacter spp. This method is based on binding of specific
probes to target nucleic acid regions and is able to detect Cronobacter spp. at a population
Chapter 9
198
level of 10
-2
CFU/ml after an enrichment of 8 h (2). Note that an enrichment step is
necessary to detect such low levels of the micro-organism.
Further research on methodology should best focus on improving the enrichment step,
ideally providing a more specific and sensitive enrichment procedure, as this so far remains
an important bottle-neck. A key physiological parameter of importance to consider in
further methodology development is the lag time (i.e. the time required for adjustment of
the microbial cells to a new environment after transfer from another, including recovery
from an injured state) of the target micro-organism, as it may add several hours to the
overall detection time. Further optimising the sensitivity and speed of detection remains a
challenging task, considering the relatively low level of the micro-organism amidst many
other micro-organisms and given the heterogeneous distribution within a bulk production of
PIF and the complexity and diversity of the food matrix itself. Molecular based detection
technologies offer opportunities for increased sensitivity but they may be sensitive to
product matrix interference. A major drawback of applying molecular based detection
technologies as compared to traditional culture based methods is that they can not clearly
differentiate living from dead bacteria. Another disadvantage is that the micro-organism
itself is not isolated and can not be studied further. Therefore, while PCR methods for
Cronobacter spp. detection may offer advantages in identification of the micro-organism,
they need to be used in conjunction with culture based confirmation methods.

Discussion and concluding remarks

199
Chapter 9
200
Occur r ence of Cronobacter spp.
Chapter 4 and 5 showed that Cronobacter spp. can be found in the production environment
of various (dry) food products, including the production facilities for milk powder,
chocolate, and cereals and also in various foods present on the Dutch market. Notably,
Cronobacter spp. were also isolated from vacuum cleaner bags collected in households.
Several other studies have since investigated the prevalence of Cronobacter spp. in a wide
variety of products (4, 13, 27, 39, 45, 54). Overall, these studies have established the
ubiquitous or widespread occurrence of Cronobacter spp. in food and beverages other than
infant formulae, with the presence of the micro-organism being confirmed in food
ingredients from plant origin, like cereals, fruits and vegetables, legume products, herbs and
spices, as well as in foods of animal origin like milk, meat and fish and products made from
these foods (27). Cronobacter spp. were isolated from scraping or sweeping surfaces in the
production-line environment of factories producing milk powder, cereals, chocolate, potato
flour, and pasta, as well as in vacuum cleaner bags from domestic environments (Chapter
4). Cronobacter spp. have also been isolated from household cleaning cloths (52) and
kitchen brushes (61). Chapter 5 showed that the prevalence of Cronobacter spp. on human
skin and in human fecal samples may be rather low, at ~1%. The isolation of Cronobacter
spp. from human samples is not new, as it was earlier isolated from various clinical sources,
including a child’s stool (25). More recently, Cronobacter spp. could be isolated from a
diarrhoeal child's stool (69) and from stool samples of hospitalised infants (0.5%
prevalence) and patients in the age group 61-70 year (51). These findings may suggest that
humans may be carrier of the micro-organism without any symptoms. Consequently, it
should be recognized that PIF can be contaminated with Cronobacter spp. through human
vectors both during manufacture and preparation.
Although Cronobacter spp. have been isolated from various environmental and clinical
sources and a wide range of food products, so far only powdered infant formulae have been
epidemiologically linked with Cronobacter infections (1,07,016,061,065,077). PIF
manufacturing industries around the world have tightened control of Cronobacter spp.
contamination in PIF and the PIF batches that are marketed generally comply with stringent
safety standards established for the presence on Cronobacter spp.. However, the bacterium
Discussion and concluding remarks

201
has been isolated from powdered infant formulae by several investigators (13, 22-24, 32,
39, 57, 59).
It needs to be stressed that PIFs are not designed to be sterile products and that they can
occasionally be contaminated with micro-organism, including pathogens. Micro-organisms
that can occasionally be present, depending on many factors relating to choice of
ingredients, processing, and other manufacturing activities, include Bacillus spp.,
Clostridium spp., Staphylococcus

spp., and Enterobacteriaceae, especially Cronobacter
spp. and Salmonella spp. (23).
Regulatory agencies worldwide (22, 23, 83) have issued warnings to professional and
domestic carers that the safety of powdered infant formulae needs to include careful
handling and preparation of PIF for consumption. While generally healthy infants should be
at no elevated risk consuming product prepared according to instructions provided by
manufacturers, certain more sensitive infant groups may require that additional care is taken
in the preparation and handling of infant formulae. Most notably, this is the case for
consumption of PIF by infants less than two months of age, especially those that are of low
birth weight, premature and or otherwise immunocompromised.
It is generally accepted that preventive measures implemented to control Salmonella spp.
occurring in PIF, comprise the basis of possible control measures and are a prerequisite to
address and control Cronobacter spp. in infant formulae production facilities (17). Until
now, contrary to what can be achieved for Salmonella spp., Cronobacter spp. can not be
totally eliminated from milk powder production facilities (30, 48, 58) possibly due to its
ubiquitous nature. Surveys in powdered milk facilities have identified the presence of
Cronobacter spp. throughout the manufacturing operation. Results obtained by sub-typing
of Cronobacter isolates revealed a genetic diversity of isolates, with most isolates being
detected in air filtering samples (42, 56). Prevalence figures of 39% have been recorded for
Cronobacter spp. in certain environmental samples from processing lines. Notably, such
high prevalence figures may be due to ongoing improvements in enrichment and
biochemical methods. Some isolates recovered from the PIF final product reported were
identical to those from the environmental samples on the basis of similarity of PFGE
Chapter 9
202
profiles, and could be found over a year period in product samples, which may suggest that
Cronobacter spp. can persist in the production environment (30).
Future research on occurrence might focus on determining the spatial and statistical
distribution of contamination, especially in PIF.

Contamination r outes of Cronobacter spp.
There are three main routes by which micro-organisms may enter PIF during its production
and at reconstitution:
1) through exposing the product to contaminated equipment, the environment and
people handling the product in the course of the manufacture,
2) through adding dry ingredients after the pasteurization step,
3) through contamination during reconstitution prior to feeding, from the
environment or due to poorly cleaned feeding bottles or poorly maintained
equipment in hospitals and homes.

Because of the importance to control infections caused by Cronobacter spp. potentially
present in PIF, it is important that the occurrence of Cronobacter spp. in the product ready
for consumption is prevented or that its presence in the prepared product is controlled to a
level that is considered acceptable for consumer safety. Preventive and control measures
need to be taken 1) during production of powdered infant formula (i.e. regarding
contamination of product ingredients and recontamination from environmental sources), 2)
during reconstitution of PIF (e.g. regarding recontamination from environmental sources
and product preparation conditions) and 3) before the reconstituted formula is consumed
(e.g. regarding product handling and storage and consumption conditions). Together,
prevention and control measures need to ensure that exposure of consumers to Cronobacter
spp. is minimized.
Regarding the possible occurrence of micro-organisms in PIF, it is worthwhile to discuss a
number of characteristics/issues such as their survival in dry powdered infant formulae,
Discussion and concluding remarks

203
their inactivation by pasteurisation and their growth in reconstituted infant formulae Such
considerations can lead risk managers to specific advice to caregivers of at risk consumers,
such as health care professionals in neonatal care units, and to further improve the existing
guidelines and handling practices (83).
Cronobacter spp. have the ability to survive under dry conditions (8) as has been
documented in the research described in Chapter 8 and it can do so for a long period.
Previous studies conclusively showed that Cronobacter spp. can be present in various dry
environments (46) and that Cronobacter spp. cells are able to survive for at least twelve
months in powdered infant formulae (12, 18). The micro-organism is clearly able to survive
in very dry environments such as those with a water activity < 0.2 [-]. In factories, the
incidence of Cronobacter spp. could increase after a wet cleaning process or due to
uncontrolled presence of water, for instance following an emergency shower (17).
Conceivably, therefore, the incidence of Cronobacter spp. in infant formula production
factories should be reduced by effectively eliminating water throughout the manufacturing
operation, by applying appropriate control measures such as the application of adequate dry
cleaning procedures within the processing environment and on the processing equipment.
Pasteurization, a heat process applied typically in the range of 60-80 ºC for a few minutes,
can lead to elimination of specific micro-organisms. Pasteurization is indeed effective in
destroying Cronobacter spp. cells during dry powdered infant formulae manufacture even
if the thermal tolerance of individual Cronobacter spp. isolates may differ (41, 64, 67).
Recent studies showed that Cronobacter spp. strains could survive the spray drying process
(3). Acidification could reduce the concentration of Cronobacter cells in different types of
infant formulae even further (44).
Notably, occasional contamination of PIF may occur after the pasteurization step, either via
addition of contaminated ingredients, such as vitamins, or via the factory environment,
though in general the incidence of Cronobacter in PIF remains low with levels, reportedly
between 0.002 to 0.92 CFU/g (21, 63). In 2007, there has been an increase of Rapid Alert
System for Food and Feed (RASFF) notifications concerning potential microbiological
Chapter 9
204
contaminations mainly due to some batches of vitamins contaminated with Cronobacter
spp. (20).
A factor possibly contributing to exposure of infants to Cronobacter spp. that has received
rather little scientific attention is biofilm formation. Some Cronobacter strains can produce
exopolysaccharides and thus it is possible that biofilm formation may occur on surfaces
commonly associated with PIF feeding equipment (41, 49). Biofilms may be present in a
variety of environments, such as process equipment, and kitchens sinks (33) and such
biofilms may be a source of contamination. In fact, formation of biofilms may protect
bacteria from aggressive conditions, such as resulting from detergents and disinfectants.
Disinfectants are used routenily in hospitals, day care facilities and food service kitchens,
however they have been shown to be quite inadequate in eliminating biofilms comprised of
Cronobacter spp. (50). Biofilm may also be formed in enteral feeding, as described further
on.

PIF r econstitution and gr owth of Cronobacter spp.
Reconstituted infant formula can either be used immediately or, if indicated on the package,
can be covered, refrigerated and used within 24 h (83). Previous results (Chapter 4 an 5)
showed the ubiquitous occurrence of Cronobacter spp., and the possible outgrowth of the
Cronobacter spp. in reconstituted infant formula (Chapter 6) (23). It is likely that storage at
ambient temperatures may lead to substantial growth. Cells of Cronobacter spp. may
recover quickly after dry stress (23), in Chapter 6 the quick recovery of Cronobacter cells
at 37 ºC is highlighted. The bacteria were furthermore found to be able to grow at 7 ºC,
with an optimum being between 37 and 43 ºC (39, 47, 63). In Chapter 6 the effect of pre-
culturing conditions on the growth was investigated at several temperatures between 8 and
47 ºC, and the growth data obtained were used to develop predictive growth models for
Cronobacter spp. The key growth parameters were quantified applying the secondary
growth model of Rosso, resulting in T
min
= 3.6 ºC, T
max
= 47.6 ºC, u
opt
= 2.31 h
-1
, T
opt
= 39.4
ºC. These parameters could be used to calculate the proliferation of Cronobacter spp. in
reconstituted infant formula at any temperature below the maximum growth temperature of
Discussion and concluding remarks

205
47.6 ºC. Studying the microbial growth of Cronobacter spp. in reconstituted infant
formulae, using cells with a different pre-culturing background (being pre-cultured to lag
phase, exponential phase and stationary phase, respectively), it was found that the various
physiological phases of the cells did not significantly influence its growth in reconstituted
powdered infant formulae.
Future research might focus on determining the lag phase of the organism in naturally
contaminated powder, although this will not be easy due to initially very low levels.
Chapter 7 focused on the cooling time required for reconstituted infant formula to reach the
set refrigerator temperature. A model was established to predict the cooling profiles of
reconstituted infant formulae given different containers or refrigerators used. The growth
parameters obtained from the research in Chapter 6, were combined with this model into
one that allows predicting growth of Cronobacter spp. during the cooling process of
prepared formulae irrespective of container volume. The model enabled to show
quantitatively to what extent practical control measures such as proper cooling or limiting
the time between preparation and consumption could help reduce growth of Cronobacter
spp., thereby likely leading to a reduction in risk as was also reported elsewhere (23, 67).
Cronobacter spp. infections have occurred in both hospitals and outpatient settings. Many
professionals and home cares for infants lack information on proper handling, storage,
preparation and use practices should be performed (31). In practice, various practices are
followed in hospitals and in the home, including mixing PIF with cold or hot water,
immediate feeding or storage for short or extended times with or without refrigeration, re-
warming or feeding without re-warming, or several feedings from a single container or a
single feeding taking an extended amount of time (11, 23, 31, 70). Based on the results in
this thesis, the relative risk associated with the various preparation practices can be assesed.

Chapter 9
206
Risk mitigation options
Several risk mitigation options have been proposed for various stakeholders (10, 23, 67, 83)
in order to prevent Cronobacter infections due to consumption of PIF. These include the
following:
− For manufacturers
- Utilizing good manufacturing and good hygiene practices to further
minimize the entry of Cronobacter spp. into in manufacturing
environment and avoid their persistence/growth (10, 23),
- Improving product labeling, e.g. including harmonized guidance to create
awareness among caregivers that PIF is not a sterile product and could
cause illness in infants, and providing clear information to caregivers for
them to adequately control microbiological hazards in reconstituted
formula (10).
− For governments and intergovernmental bodies:
- Setting a microbiological criteria for Cronobacter spp. and/or suitable
indicator micro-organisms in order to reduce the risk of Cronobacter
infection (23),
- Educating health care professionals to provide effective hygiene training
to parents and professional caregivers to ensure that PIF is prepared,
handled and stored properly (10).
− For hospitals:
- Prevention of prolonged storage time of prepared formulae,
- Restricting feeding times, especially regarding tube feeding,
- Exclusively using sterilized liquid infant formula for high risk infants,
- Quick cooling of reconstituted formula using ice or running water and
storing only small volumes of prepared formulae in refrigerators (83),
- Reconstituting PIF with water of at least 70 ºC, in order to take benefit of
the heat to inactivate viable Cronobacter spp. possibly present (23, 67).

Discussion and concluding remarks

207
In the next paragraphs four mitigation options are discussed in some more detail, where
relevant, employing the estimated parameters and predictive models developed in this
thesis (Chapter 6 and 7) to determine the effectiveness of the mitigation options in reducing
absolute or relative risks.

Risk mitigation option 1: Preparation of PIF with water of at least 70 °C
The Food and Agricultural Organization anf the World Health Organization (FAO/WHO)
guidelines for safe preparation, storage and handling of powdered infant formula (83)
advise several steps to reconstitute PIF, namely 1) heat-treat the bottle at each feeding; 2)
wash hands with warm water and soap before preparation; 3) use boiled water of a
temperature no less than 70 ºC; 4) add the PIF; 5) quickly cool prepared formula to feeding
temperature by holding the bottle under running tap water, or by placing in a container of
cold or iced water; 6) check the temperature before feeding in order to avoid scalding the
infant’s mouth; 7) discard any feed that is not consumed within 2 hours after feeding. If
prepared formula is not to be consumed within two hours after preparation it should be
quickly cooled immediately after preparation and stored in a refrigerator (at a maximum
temperature of 5 ºC) and used within 24 h (83). These guidelines have been implemented or
advised in some countries, e.g. Italy (11) and Ireland (75), while in other countries, such as
in the Netherlands, caregivers are still advised to prepare one or two bottles in advance
following the instructions on the label, and place the bottles immediately in a refrigerator
with a temperature of no more than 4 ºC with a maximum storage time of 8 h (81).
Following the advice provided in FAO/WHO guidelines to reduce the risk of Cronobacter
infection from PIF, reconstituting PIF with water > 70 ºC should result in at least a 4-log
reduction of Cronobacter spp.(67).
A study in Italy on home preparation of PIF showed that only 22% (n=29) of caregivers use
water of at least 70 ºC when preparing infant formula. Furthermore, after preparing a bottle,
10% of the parents stored the bottle at room temperature and 3% in a bottle warmer; 16% of
the parents used the stored formula within 4 h after preparation (11). These observations
Chapter 9
208
indicate that the guidelines are implemented only by a certain percentage of the caregivers.
Evedently, it may take quite some time and effort for training caregivers to properly
implement best practices in reliable way.
A key beneficial effect of using water of at least 70 ºC is the on-site pasteurization of the
bottle. Baby bottles may be heavily contaminated if small amounts of infant formulae are
not immediately discarded after feeding, left at room temperature (70) and not properly
cleaned and heat inactivated before re-use. This may allow Cronobacter spp. to outgrow to
hazardous numbers (Chapter 6). In the Netherlands, it is advised to clean bottles using a
dishwasher with a washing temperature of at least 55 ºC or using warm soapy water in
combination with a dish brush (80). Note that Cronobacter spp. may be introduced into a
bottle due to its presence in for example a contaminated dish cloth or dish brush used for
cleaning.
A potential disadvantage of reconstitution with hot water is its effect on heat sensitive
vitamins. Furthermore, the caregiver or the infant may be wounded by scalding. Notably,
the effect of the high temperature on dissolving the powder should also be considered. It
would be relevant to dedicate further research on suitable control measures to minimizing
an impact on vitamins and to minimizing the risk of scalding.
Another important factor that needs to be taken into account is the potential activation of
Bacillus spores. Infant formulae are frequently contaminated with low numbers of Bacillus
spores (5). It has been reported that reconstitution at temperatures ranging from 65 to 95 ºC
may lead to germination of endospores of B. cereus, which may cause foodborne illnesses
in infants such as diarrhea (73, 74). Thus, the high temperatures as recommended may
encourage the proliferation of other bacterial cells, germination of spores and possible toxin
production. For this reason, more research is encouraged to investigate the effect of using
hot water for the preparation of PIF on sporulation of spores, such as B. cereus.
Discussion and concluding remarks

209
Risk mitigation option 2: The relative risk reduction achieved when
reconstitution is not with water of 70 °C but with water of 65 or 60 °C
The expert consultation of the FAO/WHO (2006) pointed out that reconstitution with liquid
of at least 70 ºC is an effective risk mitigation strategy (23). A relevant risk management
question then is: how much less risk reduction is achieved at lower reconstitution
temperature?
In order to investigate the effect of different water temperatures, the following scenarios
were calculated: use of water of 40, 50, 60, 65 and 70 ºC to reconstitute the powder in
either a can or a bottle and to store it in both a refrigerator with still air or convection.
Furthermore, the effect of rapid cooling on the relative risk was estimated. The baseline
scenario for relative risk calculations was preparation of the formula with water of 40 ºC in
a bottle and storing the bottle in a refrigerator with still air at 6 ºC for 8 hours, as shown in
Table 2. The web-based tool of the risk assessment model of the FAO/WHO (68) was used
to assess the relative risk associated with the various preparation and handling scenarios.
The initial Cronobacter spp. concentration was assumed to be 1 CFU/g.

Table 2. Preparation and storing conditions for scenario’s
(Room) temper atur e [ºC] Time [h]
Preparation 25
a
0.25
Refrigeration
b
6 8
Re-warming 37 0.5
Feeding 30 0.5
a
Please note that the temperature of the water used for reconstitution is between 40 and 70 ºC
whereas preparation was at room temperature is 25 ºC.
b
Refrigeration conditions used in the various scenarios included immediate refrigeration after
preparation in still air or convection refrigerator. Where rapid cooling was part of the option it was
part of the refrigeration step.

For the given preparation and storage scenarios, the temperature of the formula was
calculated by the web-tool at successive time points between preparation and consumption.
Chapter 9
210
From these, the duration of the lag time, growth or inactivation rates were estimated and the
resulting cell concentration calculated. The probability of infection (P) resulting from
consumption of an estimated average dose of Cronobacter spp. can be described by the
exponential dose-response model (Equation 1) (71).

( ) rD exp P ÷ ÷ = 1 (1)
where r [1/CFU] is the exponential dose-response parameter and D [CFU] the estimated dose
consumed.

The value of the dose-response parameter r has not been established with any certainty.
Values between 10
-10
and 10
-5
have been suggested

(23). For the low values of the dose (D),
it is assumed that the risk (P) and the dose (D) are linearly correlated. Therefore, even
without knowing the exact value of the r value a relative risk can be determined.
Figure 2 shows the estimated relative risk for the various scenarios. At a reconstitution
temperature of 40 ºC, there is a slight reduction in estimated risk for reconstituting PIF in
the bottles for rapid cooling and convection refrigeration compared with the baseline
scenario. Reconstitution of PIF in a can shows an increase (100%) in the relative risk.
When water of 50 ºC is used for reconstitution in a bottle, the relative risk increases (32%)
for the scenario with still air and the bottle while it does not change for convection
refrigeration but it is greatly reduced (38%) in the case of rapid cooling (for example using
ice or running water). Note that these comparisons are made in relation to the baseline
scenario. The relative risk increases a lot (247%) when a can volume is used rather than a
bottle. Rapid cooling of PIF reconstituted at 50 ºC in a bottle, could lead to a decrease in
relative risk of 68% compared to cooling in a refrigerator with still air.
Reconstituting PIF with water of 60 ºC gives a variable effect in risk, with a reduction for
all storage scenarios involving bottles but with increased risk for formula preparation in a
can followed by storage in a still air refrigerator.

Discussion and concluding remarks

211
Figur e 2. Increase in relative risk of different preparation, storage and handling practices for prepared
formula as simulated using the web based tool (68). Note that the chosen setting (at 40 ºC), as shown
in Table 2, was used as baseline scenario for comparing relative risk levels.

At the higher temperatures, 65 and 70 ºC, there was no apparent effect of convection and
rapid cooling for both the bottle and can on the relative risk. In all cases a similar risk
reduction of at least 4 log units or more was observed. Reconstituting at a temperature
somewhat colder than 70 ºC, still can be expected to achieve a reduction in Cronobacter,
subsequently in relative risk reduction. However, when the temperature of the water is 60
ºC, instead of 70 ºC, a slight inactivation of Cronobacter spp. is achieved and the relative
risk may even be increased compared with the baseline scenario in case PIF is prepared and
stored in a can. Results from these simulations can be compared with reported reductions
achieved when water of 60 ºC and 70 ºC was used to reconstitute PIF, where a 1.2- and
5.26-log reduction of Cronobacter populations was obtained (67).
Therefore, reconstituting PIF with water of more than 70 ºC may have a safety margin, as
this action would yield at least a 4 to 6-log reduction of Cronobacter spp., depending on the
type of formula used and the type of Cronobacter strain present.
Chapter 9
212
Risk mitigation option 3: Impact of hang-time in case of continuous tube-
feeding on consumer risk
The WHO guidelines for safe preparation, storage and handling of PIF do not consider the
use of enteral feeding tubes, which can be in place for various time periods, ranging from
< 6 h to > 48 h (35), at a relatively high temperature (warm rooms may be 35
o
C) and
therefore may be a cause of large outgrowth of bacteria, or may accommodate micro-
organisms as these may possibly attach and form biofilms. Enteral feeding tubes may be
used to especially feed low birth weight infants in neonatal care units, which are the typical
weak and the group at greatest risk (84).
Depending on the regime in the hospital, tube-fed formula is refreshed every 2 to 3 hours
(34). Research has shown that Cronobacter spp. can form biofilms on plastic surfaces (41)
and attach to enteral feeding tubes at ambient temperatures (49). It is therefore reasonable
to assume that Cronobacter spp. will multiply in the tube, which is at about body
temperature and also in the feeding container (syringe) which is at ambient temperature.
Considering the advice from the FAO/WHO expert meetings (22, 23) and the EFSA panel
(19) to limit the hang time of tube-fed prepared infant formulae to 4h, the impact of the
following scenarios on the level of Cronobacter spp. was estimated: the feeding container
was assumed to be a baby bottle containing 120 ml of reconstituted infant formula was
assumed to contain an initial concentration of Cronobacter spp. of 1 CFU/ml. Therefore,
the estimated parameters for the baby bottle from Chapter 7 were presumed to be similar
for the feeding container. The tube feeding started immediately after preparation and the
formula was prepared with lukewarm water (40 ºC) in a baby bottle. The feeding time was
simulated up to 6 hours in a warm room (35 ºC) and in a cold room (20 ºC). The growth of
Cronobacter spp. was predicted, using the specific growth rate calculated with the
secondary growth model of Rosso with parameters as determined in Chapter 6. The
increase in cell count was predicted with the exponential growth function, both neglecting
the lag time and including the lag time according to the optimal fit of the k-value:
k = u x λ = 2.88 as described in Chapter 7.
Discussion and concluding remarks

213
Figure 3 gives the temperature profile, together with the predicted increase in cell count
during tube feeding in a cold room (20 ºC), while the temperature profile and predicted
increase in cell count during tube feeding in a warm room are depicted in Figure 4. As can
be seen, the decrease in temperature from 40 to 20 ºC is slow and takes more than 6 hours,
while cell numbers increase in time. The exponential growth function, which neglects lag
time, overestimates growth as it predicts the increase to start immediately. When
considering a lag time, growth will start after 2 hours. The predicted growth shows a 1-log
unit increase in 4 h.
Tube feeding in a warm room (35 ºC), may result in a considerably larger increase in cell
numbers as can be seen in Figure 4. After 4 hours, an increase of at least 4-log units is
predicted with the exponential growth function. This prediction must be assumed to be an
overestimation, as lag time is neglected (Chapter 7). Including the lag time, as estimated
with a k-value of 2.88, an increase of 2.5 log units is predicted in 4 h. These predictions
show that the room temperature (Figures 3 A and 4A) and thus, the temperature of the
formula being fed may significantly affect growth of Cronobacter spp., and needs to be
considered for tube feeding infant formula to sensitive infants.


Chapter 9
214




Figur e 3. Temperature profile (A) and the number of Cronobacter spp. (B) in a bottle used for tube
feeding in a room which is at 20 ºC. Grey line = prediction with exponential growth function; black
line = prediction with optimal fit k=2.88.
Discussion and concluding remarks

215




Figur e 4. Temperature profile (A) and the number of Cronobacter spp. (B) in a bottle used for tube
feeding in a room which is at 35 ºC. Grey line = prediction with exponential growth function; black
line = prediction with optimal fit k=2.88.
Chapter 9
216
Risk mitigation option 4: Use sterile formula for the population at greatest
risk.
The two FAO/WHO expert meetings have established that all infants at or below 12 months
are at particular risk for Cronobacter infections (22, 23). Among these infants, those at
greatest risk are neonates (< 28 days), particularly pre-term infants, low birthweight (LBW)
infants (< 2500 g), and immuno-compromised infants, and those infants that are less than 2
months of age. Infants of HIV-positive mothers are also at risk, because they may
specifically require infant formula and be more susceptible to infection (10, 24).
Table 3 gives an overview of data on reported cases adapted from the FAO/WHO expert
meetings. Cases are divided in subgroups, being all infants, the group at greatest risk, low
birthweight infants, and neonates. It is assumed that all infants have been formula fed only
and that all cases were due to the presence of Cronobacter spp. in PIF; both assumptions
are conservative and may cause overestimation of risks. It was estimated what the risk
reduction could be of providing sterile formula to the whole or to a subset of the group at
greatest risk. A 100% risk reduction was assumed when using sterile formula for each
individual infant.
Using liquid sterile formula for all infants of the group at greatest risk, would possibly lead
to a 100% reduction of Cronobacter infection cases for this group and a 72% (111/155),
reduction in the total group of infants. Subsequently, a mortality reduction of 90% (26/29)
would be achieved. Feeding only the LBW group with sterile formula is expected to have
reduced the morbidity of Cronobacter infection with 37% (57/155), and the motility by
52% (15/29). If all neonates, in both homes and household settings, are fed with sterile
formula only a 48% (75/155) reduction of Cronobacter infections and a 24% (7/29)
reduction in deaths would be reached.
Discussion and concluding remarks

217
Table 3. Estimation of the risk reduction, feeding the group particularly at risk and the at greatest risk
group with sterile formula. Data adapted from WHO/FAO 2008 report (24)
No. cases Recover ed Outcome
unknown
Number of
deaths
All infants (≤ 12 months of age
and including age not known)
a

155 85 41 29
Group indicated to be at greatest
risk (≤ 2 months)
a

111 71 14 26
LBW group (< 2500g) 57 38 4 15
Neonates (< 28 days)
a
75 49 9 7
Age at moment of illness not
known
10 2 6 2
a
in all groups is age indicated to be age at onset of the Cronobacter infection.

These estimates show that it would be worthwhile to consider using sterile formula for the
group at greatest risk, especially for the LBW group a larger reduction in the mortality rate
was estimated compared to the neonate group. Although prematurity and low birth-weight
are the most important risk factors for diseases, like necrotising enterocolitis in newborn
infants, also other strategies, next to using sterile formula, should be considered to address
effect in preventing illness.

Concluding r emar ks
In conclusion; results in this thesis showed that Cronobacter spp. is an ubiquitous
pathogen, because of its widespread occurrence. It was isolated from a wide variety of of
food production environments and food products, including PIF. Therefore, it should be
considered that PIF is not a sterile product. Most of reported Cronobacter infections are
related to PIF, although infections were also reported in adults, showing that young infants
are at greatest risk. For this reason, care should be taken while preparing, storing and
handling PIF, as results in this thesis showed that potential rapid growth Cronobacter spp.
in reconstituted infant formula is possible, depending on storage and feeding practices.
Overall, the results described in this thesis can be used to estimate the exposure and thus the
Chapter 9
218
probability of illness due to Cronobacter spp. after the consumption of reconstituted infant
formulae.

Refer ences
1. Acker van J., de Smet F. , G. Muyldermans, A. Bougatef, A. Naessens, and S.
Lauwers. 2001. Outbreak of necrotizing enterocolitis associated with Enterobacter
sakazakii in powdered milk formula. J. Clin. Microbiol. 39:293-297.
2. Almeida, C., N. F. Azevedo, C. Iversen, S. Fanning, C. W. Keevil, and M. J.
Vieira. 2009. Development and application of a novel peptide nucleic acid probe
for the specific detection of Cronobacter genomospecies (Enterobacter sakazakii)
in powdered infant formula. Appl. Environ. Microbiol. 75:2925-2930.
3. Arku, B., N. Mullane, E. Fox, S. A. Fanning, and K. Jordan. 2008. Enterobacter
sakazakii survives spray drying. Int. J. Dairy Technol. 61:102-108.
4. Baumgartner, A., M. Grand, M. Liniger, and C. Iversen. 2009. Detection and
frequency of Cronobacter spp. (Enterobacter sakazakii) in different categories of
ready-to-eat foods other than infant formula. Int. J. Food Microbiol. 136:189-192.
5. Becker, H., G. Schaller, W. Von Wiese, and G. Terplan. 1994. Bacillus cereus in
infant foods and dried milk products. Int. J. Food Microbiol. 23:1-15.
6. Besse, N. G., A. Leclercq, V. Maladen, C. Tyburski, and B. Lombard. 2006.
Evaluation of the International Organization for Standardization-International
Dairy Federation (ISO-IDF) draft standard method for detection of Enterobacter
sakazakii in powdered infant food formulas. J. AOAC Int. 89:1309-1316.
7. Block, C., O. Peleg, N. Minster, A. Simhon, I. Arad, and M. Shapiro. 2002.
Cluster of neonatal infections in Jerusalem due to unusual biochemical variant of
Enterobacter sakazakii. Eur. J. Clin. Microbiol. Infect. Dis. 21:613-616.
8. Breeuwer, P., A. Lardeau, M. Peterz, and H. Joosten. 2003. Desiccation and heat
tolerance of Enterobacter sakazakii. J. Appl. Microbiol. 95:967-973.
9. Brehm-Stecher, B., C. Young, L. Jaykus, and M. L. Tortorello. 2009. Sample
preparation: The forgotten beginning. J. Food Prot. 72:1774-1789.
10. CAC. 2008. Code of hygienic practice for powdered formulae for infants and
young children. CAC/RCP 66-2008.
11. Carletti, C., and A. Cattaneo. 2008. Home preparation of powdered infant formula:
is it safe? Acta Paediatr. 97:1131-1132.
12. Caubilla-Barron, J., and S. J. Forsythe. 2007. Dry stress and survival time of
Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered
infant formula. J. Food Prot. 70:2111-2117.
Discussion and concluding remarks

219
13. Cawthorn, D. M., S. Botha, and R. C. Witthuhn. 2008. Evaluation of different
methods for the detection and identification of Enterobacter sakazakii isolated
from South African infant formula milks and the processing environment. Int. J.
Food Microbiol. 127:129-138.
14. CDC. 2002. Enterobacter sakazakii infections associated with the use of powdered
infant formula - Tennessee, 2001. MMWR. 51:297-300.
15. Chen, Y., T. S. Hammack, K. Y. Song, and K. A. Lampel. 2009. Evaluation of a
revised US Food and Drug Administration method for the detection and isolation
of Enterobacter sakazakii in powdered infant formula: pre-collaborative study. J.
AOAC Int. 92:862-872.
16. Clark, N. C., B. C. Hill, C. M. O'Hara, O. Steingrimsson, and R. C. Cooksey.
1990. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial
outbreaks. Diagn. Microbiol. Infect. Dis. 13:467-472.
17. Cordier, J. L. 2008. Production of powdered infant formulae and microbiological
control measures. p. 145-185. In: J.M. Farber and S.J. Forsythe (eds.),
Enterobacter sakazakii (emerging issues in food safety) ASM Press, Washington,
D.C., New Brunswick, New Jersey.
18. Edelson-Mammel, S. G., and R. L. Buchanan. 2005. Survival of Enterobacter
sakazakii in a dehydrated powdered infant formula. J. Food Prot. 68:1900-1902.
19. EFSA. 2004. Opinion of the scientific panel on biological hazards on the request
from the commission related to the microbiological risks in infant formulae and
follow-on formulae. The EFSA Journal. 113:1-35.
20. European Commission. 2007. The Rapid Alert System for Food and Feed
(RASFF) Annual Report 2007. p. 65. In: Official Journal of the European Union.
Luxembourg.
21. Fanning, S., and S. J. Forsythe. 2008. Isolation and identification of Enterobacter
sakazakii. p. 27-59. In: J. M. Farber and S. J. Forsythe (eds.), Enterobacter
sakazakii (emerging issues in food safety) ASM Press, Washington, D.C., New
Brunswick, New Jersey.
22. FAO/WHO. 2004. Enterobacter sakazakii and other micro-organisms in powdered
infant formula: meeting report. Microbiological Risk Assessment Series No. 6,
Rome, Italy.
23. FAO/WHO. 2006. Enterobacter sakazakii and Salmonella in powdered infant
formula: meeting report. Microbiological Risk Assessment Series No. 10 Rome,
Italy.
24. FAO/WHO. 2008. Enterobacter sakazakii (Cronobacter spp.) in powdered follow-
up formulae: meeting report. Microbiological Risk Assessment Series No. 15
Rome, Italy.
Chapter 9
220
25. Farmer III, J. J., M. A. Asbury, F. W. Hickman, D. J. Brenner, and the
Enterobacteriaceae study group. 1980. Enterobacter sakazakii: A new species of
"Enterobacteriaceae" isolated from clinical specimens. Int. J. Syst. Bacteriol.
30:569-584.
26. FDA. Date, 11 April 2002, Health professionals letter on Enterobacter sakazakii
infections associated with use of powdered (dry) infant formulas in neonatal
intensive care units. Available at:
http://www.fda.gov/Food/FoodSafety/ProductSpecificInformation/InfantFormula/
AlertsSafetyInformation/ucm111299.htm. Accessed 5 August, 2009.
27. Friedemann, M. 2007. Enterobacter sakazakii in food and beverages (other than
infant formula and milk powder). Int. J. Food Microbiol. 116:1-10.
28. Friedemann, M. 2009. Epidemiology of invasive neonatal Cronobacter
(Enterobacter sakazakii) infections. Eur. J. Clin. Microbiol. Infect. Dis. 28:1297-
1304.
29. Guillaume-Gentil, O., V. Sonnard, M. C. Kandhai, J. D. Marugg, and H. Joosten.
2005. A simple and rapid cultural method for detection of Enterobacter sakazakii
in environmental samples. J. Food Prot. 68:64-69.
30. Hein, I., B. Gadzov, D. Schoder, H. Foissy, B. Malorny, and M. Wagner. 2009.
Temporal and spatial distribution of Cronobacter isolates in a milk powder
processing plant petermined by Pulsed-Field Gel Electrophoresis. Foodb. Path.
Dis. 6:225-233.
31. Herbold, N. H., and E. Scott. 2008. A pilot study describing infant formula
preparation and feeding practices. Int. J. Environ. Health Res. 18:541-459.
32. Heuvelink, A. E., M. Ahmed, F. D. Kodde, J. T. M. Zwartkruis-Nahuis, and E. de
Boer. 2002. Enterobacter sakazakii in melkpoeder. De Ware(n)chemicus. 32:17-
30.
33. Hood, S. K., and E. A. Zottola. 1995. Biofilm in food processing. Food Control.
6:9-18.
34. Hurrell, E., E. Kucerova, M. Loughlin, J. Caubilla-Barron, and S. J. Forsythe.
2009. Biofilm formation on enteral feeding tubes by Cronobacter sakazakii,
Salmonella serovars and other Enterobacteriaceae. Int. J. Food Microbiol.
136:227-231.
35. Hurrell, E., E. Kucerova, M. Loughlin, J. Caubilla-Barron, A. Hilton, R.
Armstrong, C. Smith, J. Grant, S. Shoo, and S. J. Forsythe. 2009. Neonatal enteral
feeding tubes as loci for colonisation by members of the Enterobacteriaceae. BMC
Infectious Diseases. 9:146.
36. ISO. 2006. ISO/TS 22964 and IDF/RM 210-first edition 2006-02-01: Milk and
milk product- Detection of Enterobacter sakazakii.
Discussion and concluding remarks

221
37. Iversen, C., P. Druggan, and S. J. Forsythe. 2004. A selective differential medium
for Enterobacter sakazakii, a preliminary study. Int. J. Food Microbiol. 96:133-
139.
38. Iversen, C., P. Druggan, S. Schumacher, A. Lehner, C. Feer, K. Gschwend, H.
Joosten, and R. Stephan. 2008. Development of a novel screening method for the
isolation of "Cronobacter" spp. (Enterobacter sakazakii). Appl. Environ.
Microbiol. 74:2550-2553.
39. Iversen, C., and S. J. Forsythe. 2004. Isolation of Enterobacter sakazakii and other
Enterobacteriaceae from powdered infant formula milk and related products.
Food Microbiol. 21:771-777.
40. Iversen, C., and S. J. Forsythe. 2007. Comparison of media for the isolation of
Enterobacter sakazakii. Appl. Environ. Microbiol. 73:48-52.
41. Iversen, C., M. Lane, and S. J. Forsythe. 2004. The growth profile,
thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant
formula milk. Lett. Appl. Microbiol. 38:378-382.
42. Iversen, C., A. Lehner, C. Fricker-Feer, K. Gschwend, and R. Stephan. 2009.
Genotyping of Cronobacter (Enterobacter sakazakii) strains isolated from an
infant formula processing plant. Archiv. Fur Lebensmittelhygiene. 60:66-72.
43. Joosten, H., J. Marugg, R. Stephan, A. Klijn, T. Jackson, and C. Iversen. 2008. A
rapid and reliable alternative to ISO 21528-1 : 2004 for detection of
Enterobacteriaceae. Int. J. Food Microbiol. 125:344-346.
44. Joosten, H. M., and A. Lardeau. 2004. Enhanced microbiological safety of
acidified infant formulas tested in vitro. SAJCN. 17:87-92.
45. Jung, M. K., and J. H. Park. 2006. Prevalence and thermal stability of
Enterobacter sakazakii from unprocessed ready-to-eat agricultural products and
powdered infant formulas. Food Sci. Biotechnol. 15:152-157.
46. Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004. Occurrence of Enterobacter sakazakii in the food production
environments and households. The Lancet. 363:39-40.
47. Kandhai, M. C., M. W. Reij, C. Grognou, M. van Schothorst, L. G. M. Gorris, and
M. H. Zwietering. 2006. Effects of preculturing conditions on lag time and
specific growth rate of Enterobacter sakazakii in reconstituted powdered infant
formula. Appl. Environ. Microbiol. 72:2721-2729.
48. Kandhai, M. C., M. W. Reij, v. K. Puyvelde, O. Guillaume-Gentil, R. R. Beumer,
and M. Van Schothorst. 2004. A new protocol for the detection of Enterobacter
sakazakii applied to environmental samples. J. Food Prot. 67:1267-1270.
49. Kim, H., J. H. Ryu, and L. R. Beuchat. 2006. Attachment of and biofilm formation
by Enterobacter sakazakii on stainless steel and enteral feeding tubes. Appl.
Environ. Microbiol. 72:5846-5856.
Chapter 9
222
50. Kim, H., J. H. Ryu, and L. R. Beuchat. 2007. Effectiveness of disinfectants in
killing Enterobacter sakazakii in suspension, dried on the surface of stainless steel,
and in a biofilm. Appl. Environ. Microbiol. 73:1256-1265.
51. Kim, J. B., S. H. Cho, Y. B. Park, J. B. Lee, J. C. Kim, B. K. Lee, H. K. Lee, and
H. S. Chae. 2008. Surveillance of stool samples for the presence of Enterobacter
sakazakii among Korean people. Yonsei Medical Journal. 49:1017-1022.
52. Kusumaningrum, H. D. 2003. Behaviour and cross-contamination of pathogenic
bacteria in household kitchens - relevance to exposure assessment. In: PhD Thesis.
Laboratory of Food Microbiology Wageningen University, Wageningen, The
Netherlands.
53. Lai, K. K. 2001. Enterobacter sakazakii infections among neonates, infants,
children, and adults - Case reports and a review of the literature. Medicine.
80:113-122.
54. Leclercq, A., C. Wanegue, and P. Baylac. 2002. Comparison of fecal coliform
agar and violet red bile lactose agar for fecal coliform enumeration in foods. Appl.
Environ. Microbiol. 68:1631-1638.
55. Malorny, B., and M. Wagner. 2005. Detection of Enterobacter sakazakii strains by
real-time PCR. J. Food Prot. 68:1623-1627.
56. Mullane, N., B. Healy, J. Meade, P. Whyte, P. G. Wall, and S. Fanning. 2008.
Dissemination of Cronobacter spp. (Enterobacter sakazakii) in a powdered milk
protein manufacturing facility. Appl. Environ. Microbiol. 74:5913-5917.
57. Mullane, N. R., J. Murray, D. Drudy, N. Prentice, P. Whyte, P. G. Wall, A. Parton,
and S. Fanning. 2006. Detection of Enterobacter sakazakii in dried infant milk
formula by cationic-magnetic-bead capture. Appl. Environ. Microbiol. 72:6325-
6330.
58. Mullane, N. R., P. Whyte, P. G. Wall, T. Quinn, and S. Fanning. 2007.
Application of pulsed-field gel electrophoresis to characterise and trace the
prevalence of Enterobacter sakazakii in an infant formula processing facility. Int.
J. Food Microbiol. 116:73-81.
59. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of
powdered substitutes for breast milk with regard to members of the family
Enterobacteriaceae. J. Clin. Microbiol. 26:743-746.
60. Muytjens, H. L., J. Van der Ros-van de Repe, and H. A. Van Druten. 1984.
Enzymatic profiles of Enterobacter sakazakii and related species with special
reference to the alpha-glucosidase reaction and reproducibility of the test system.
J. Clin. Microbiol. 20:684-686.
61. Muytjens, H. L., H. C. Zanen, H. J. Sonderkamp, L. A. Kollee, I. K. Wachsmuth,
and J. J. Farmer, 3rd. 1983. Analysis of eight cases of neonatal meningitis and
sepsis due to Enterobacter sakazakii. J. Clin. Microbiol. 18:115-120.
Discussion and concluding remarks

223
62. Nair, M. K. M., and K. S. Venkitanarayanan. 2006. Cloning and Sequencing of the
ompA gene of Enterobacter sakazakii and development of an ompA-targeted PCR
for rapid detection of Enterobacter sakazakii in infant formula. Appl. Environ.
Microbiol. 72:2539-2546.
63. Nazarowec-White, M., and J. M. Farber. 1997. Incidence, survival, and growth of
Enterobacter sakazakii in infant formula. J. Food Prot. 60:226-230.
64. Nazarowec-White, M., and J. M. Farber. 1997. Thermal resistance of Enterobacter
sakazakii in reconstituted dried-infant formula. Lett. Appl. Microbiol. 24:9-13.
65. Noriega, F. R., Kotloff, K.L., Martin, M.A., Schwalbe, R.S. 1990. Nosocomial
bacteremia caused by Enterobacter sakazakii and Leuconostoc mesenteroides
resulting from extrinsic contamination of infant formula. Pediatr. Infect. Dis. J.
9:447-449.
66. O'Brien, S., B. Healy, C. Negredo, S. Fanning, and C. Iversen. 2009. Evaluation of
a new one step enrichment in conjunction with a chromogenic medium for the
detection of Cronobacter spp. (Enterobacter sakazakii) in powdered infant
formula. J. Food Prot. 72:1472 -1475.
67. Osaili, T. M., R. R. Shaker, M. S. Al-Haddaq, A. A. Al-Nabulsi, and R. A. Holley.
2009. Heat resistance of Cronobacter species (Enterobacter sakazakii) in milk and
special feeding formula. J. Appl. Microbiol. 107:928-935.
68. Paoli, G., and E. Hartnett. 2006. Overview of a risk assessment model for
Enterobacter sakazakii in powdered infant formula. Prepared for WHO/FAO.
Available at: http://www.mramodels.org/esakmodel/ESAKRAModelWizard.aspx.
Accessed 23 October 2009.
69. Raghav, M., and P. K. Aggarwal. 2007. Isolation and characterization of
Enterobacter sakazakii from milk foods and environment. Milchw. Milk Sci. Int.
62:266-269.
70. Redmond, E. C., and C. J. Griffith. 2009. The importance of hygiene in the
domestic kitchen: Implications for preparation and storage of food and infant
formula. Perspectives in Public Health. 129:69-76.
71. Reij, M. W., I. Jongenburger, E. Gkogka, L. G. M. Gorris, and M. H. Zwietering.
2009. Perspective on the risk to infants in the Netherlands associated with
Cronobacter spp. occurring in powdered infant formula. Int. J. Food
Microbiol.136:232-237.
72. Restaino, L., E. W. Frampton, W. C. Lionberg, and R. J. Becker. 2006. A
chromogenic plating medium for the isolation and identification of Enterobacter
sakazakii from foods, food ingredients, and environmental sources. J. Food Prot.
69:315-322.
73. Rowan, N. J., and J. G. Anderson. 1998. Effectiveness of cleaning and disinfection
procedures on the removal of enterotoxigenic Bacillus cereus from infant feeding
bottles. J. Food Prot. 61:196-200.
Chapter 9
224
74. Rowan, N. J., J. G. Anderson, and A. Anderton. 1997. Bacteriological quality of
infant milk formulae examined under a variety of preparation and storage
conditions. J. Food Prot. 60:1089-1094.
75. Safefood and Health Service Executive. 2009. How to prepare your baby's bottle
available at:
http://www.safefood.eu/Global/Publications/Consumer%20information/PIF%20le
aflet.pdf?epslanguage=en. Accessed 30 September 2009.
76. Seo, K. H., and R. E. Brackett. 2005. Rapid, specific detection of Enterobacter
sakazakii in infant formula using a real-time PCR assay. J. Food Prot. 68:59-63.
77. Simmons, B. P., M. S. Gelfand, M. Haas, L. Metts, and J. Ferguson. 1989.
Enterobacter sakazakii infections in neonates associated with intrinsic
contamination of a powdered infant formula. Infect. Control Hosp. Epidemiol.
10:398-401.
78. U.S. Food Drug and Administration (FDA). 2002. Isolation and enumeration of
Enterobacter sakazakii from dehydrated powdered infant formula. Accessed 10
February, 2009.
79. Van Schothorst, M. 1976. Resuscitation of injured bacteria in foods. p. 317-328.
In: Skinner F. A., and W. B. Hugo (eds.), Inhibition and inactivation of vegatative
microbes. Academic Press, Inc. (Londen), Ltd., Londen.
80. Voedingscentrum. 2008. Klaarmaken van flesvoeding. Available at:
http://www.voedingscentrum.nl/nl/eten-gezondheid/zwangerschap-en-
kind/flesvoeding/klaarmaken-van-flesvoeding.aspx. Accessed 28 September 2009.
81. Voedingscentrum. 2009. Flesvoeding. Avalaible at:
http://www.voedingscentrum.nl/nl/eten-gezondheid/zwangerschap-en-
kind/flesvoeding.aspx. Accessed 21 September 2009.
82. Weir, E. 2002. Powdered infant formula and fatal infecton with Enterobacter
sakazakii. Can. Med. Assoc. J. 166:1570.
83. WHO. 2007. Guidelines on Safe preparation, storage and handling of powdered
infant formula. Available at:
http://www.who.int/foodsafety/publications/micro/pif_guidelines.pdf. Accessed 13
February, 2009.
84. Yu, V. Y. H. 1999. Enteral feeding in the pre-term infant. Early Hum. Dev. 56:89-
115.
85. Zhou, Y. H., Q. P. Wu, X. K. Xu, X. J. Yang, Y. W. Ye, and J. M. Zhang. 2008.
Development of an immobilization and detection method of Enterobacter
sakazakii from powdered infant formula. Food Microbiol. 25:648-652.




225

226


227
Samenvatting
Het genus Cronobacter bestaat uit 6 verschillende species (= soorten = spp.) en is
onderdeel van de familie van de Enterobacteriaceae. Vóór 2007 werd Cronobacter
aangeduid als Enterobacter sakazakii, en tot 1980 werd het genus aangegeven met de naam
“geelgekleurde Enterobacter cloacae”. Het zijn Gram-negatieve, staafvormige bacteriën
die bij specifieke groepen gevoelige personen ziekte kunnen veroorzaken. De meeste
Cronobacter stammen maken een kenmerkend geel pigment aan als ze op Trypton Soya
Agar gekweekt worden.
Cronobacter heeft ziektes veroorzaakt in alle leeftijdsgroepen, maar het merendeel van de
infecties komt voor bij zuigelingen jonger dan één jaar. In het bijzonder baby’s jonger dan
twee maanden, die te vroeg zijn geboren, ondergewicht of een verzwakt afweersysteem
hebben, lopen een verhoogd risico. Cronobacter infecties zijn zeldzaam, maar kunnen
resulteren in vormen van neonatale meningitis (hersenvliesontsteking bij pasgeborenen),
sepsis (bloedvergiftiging) en necrotiserende enterocolitis (ernstige ontsteking en afsterving
van het maagdarmkanaal). Van de gerapporteerde patiënten is 19% overleden tussen 1961
en 2008 en van degenen die een infectie overleven, houdt een groot gedeelte blijvende
neurologische complicaties. In nationaal en internationaal verband wordt sinds enige jaren
aandacht besteed aan de mogelijkheid dat de bacterie via poedervormige
zuigelingenvoeding ziekte kan veroorzaken.
Cronobacter is uit verschillende voedingsmiddelen en omgevingen geïsoleerd.
Zuigelingenvoeding is echter het enige levensmiddel dat als mogelijke infectiebron
geïdentificeerd is. In dit proefschrift worden in hoofdstuk 2 en 3 methoden beschreven om
snel en nauwkeurig vast te stellen of er Cronobacter voorkomt in een product of in een
omgeving (detectie) en of de gevonden bacteriesoort inderdaad een Cronobacter is
(identificatie). De ontwikkelde selectieve methoden zijn toegepast in diverse producten en
omgevingen, waaronder huishoudens. Het blijkt dat Cronobacter spp. voorkomt in onder
andere fabrieken waar graanproducten, aardappelzetmeel of melkpoeder verwerkt worden.
Eveneens werd Cronobacter spp. geïsoleerd uit vuil uit stofzuigerzakken, verzameld in
verschillende huishoudens.

228
Hoofdstuk 4 en 5 van dit proefschrift geven meer inzicht in het voorkomen van
Cronobacter spp., eventuele besmettingsroutes en bieden tevens mogelijkheden voor
maatregelen om Cronobacter spp. infecties te voorkomen.
Eerder onderzoek heeft aangetoond dat Cronobacter spp. in vergelijking tot andere
bacteriën uit de Enterobactericeae familie, goed kan overleven in droge omgevingen. De
bacterie kan dan ook via herbesmetting vanuit productieruimten, met name in de
afvulruimten, in zuigelingenvoeding terechtkomen. Een andere besmettingsroute is via het
toevoegen van besmette, droge ingrediënten aan de zuigelingenvoeding, waarna tijdens het
verdere productie proces geen hittebehandeling meer plaatsvindt.
Aangezien Cronobacter spp. in de fabrieksomgeving aanwezig kan zijn, is het zaak te
voorkómen dat de bacterie in zuigelingenvoeding aanwezig is, door toepassing van
kwaliteitssystemen zoals Good Manufacturing Practices (GMP) en Good Hygiene Practices
(GHP). De in dit proefschrift beschreven resultaten kunnen onder ander door
zuigelingenvoeding fabrikanten worden geïmplementeerd in de bestaande Hazard Analysis
Critical Control Point (HACCP) plan om zo de veiligheid beter te beheersen. Ondanks alle
voorzorgsmaatregelen, kan het toch voorkomen dat Cronobacter spp. in de
zuigelingenvoeding aanwezig is. Het is belangrijk te realiseren dat zuigelingenvoeding niet
een steriel product is en dat onjuiste bereiding en bewaring van de voeding de kans op
besmetting kunnen vergroten. Over de groei van Cronobacter spp. in aangemaakte
zuigelingenvoeding was tot voor kort niet veel bekend. Resultaten uit een experimentele
dierstudie doen suggereren dat hoge aantallen levende cellen van Cronobacter spp. (meer
dan 100.000 cellen per voeding ofwel > 10
5
) nodig zouden zijn om ziekte of dood te
veroorzaken. Deze hoge aantallen zijn nooit in droge poedervormige zuigelingenvoeding
aangetroffen, dankzij een goede beheersing van de productie en de hygiëne bij het vullen
van de verpakkingen. Besmetting kan echter ook optreden tijdens het bereiden van de
zuigelingenvoeding zowel in het ziekenhuis als thuis. Wanneer lage aantallen Cronobacter
spp. voorkomen in de poedervormige zuigelingenvoeding kunnen deze redelijk goed
overleven (hoofdstuk 8) en eventueel verder uitgroeien nadat de poedervormige
zuigelingenvoeding aangemaakt is met water. De snelheid en mate van uitgroei is sterk
afhankelijk van de temperatuur van de aangemaakte voeding. In hoofdstuk 6 worden


229
voorspellende modellen gebruikt om de groeiparameters, waaronder de groeisnelheid van
Cronobacter spp. in aangemaakte zuigelingenvoeding, te kunnen bepalen bij diverse
temperaturen. Cronobacter spp. hebben optimale groei tussen 37 en 39 ºC.
Aangemaakte zuigelingenvoeding kan worden bewaard in de koelkast. De temperatuur van
huishoudkoelkasten is echter niet altijd beneden de 7 °C, wat in Nederland de
voorgeschreven maximum temperatuur is. Een nog belangrijker probleem is echter dat de
voeding niet onmiddellijk dezelfde temperatuur heeft als de lucht in de koelkast. Na het
plaatsen in de koelkast duurt het geruime tijd voordat de koelkasttemperatuur bereikt is. De
verkregen groeiparameters uit hoofdstuk 6 zijn in hoofdstuk 7 gebruikt om een model te
ontwikkelen om de proliferatie van Cronobacter spp. in aangemaakte zuigelingenvoeding
te voorspellen gedurende het afkoelproces. Voorspellingen laten zien dat de groei van
Cronobacter spp. kan worden voorkomen door de aangemaakte zuigelingenvoeding in
kleine porties ter grootte van één voeding te bewaren in de koelkast, en door de
poedervormige zuigelingenvoeding aan te maken met water dat een temperatuur van 25 ºC
of minder heeft.
De in dit proefschrift beschreven resultaten zijn gebruikt in de expert meetings van de FAO
(Food and Agricultural Organization of the United Nations) en de WHO (World Health
Organization) waarbij het risico van Cronobacter spp. in zuigelingenvoeding is onderzocht.
De ontwikkelde modellen kunnen worden toegepast voor het kwantificeren van de groei
van Cronobacter spp. in de zuigelingenvoeding onder verschillende aanmaak- en
bewaaromstandigheden om zo de effecten van opslag en bereiding van zuigelingenvoeding
door te rekenen, bijvoorbeeld in het kader van microbiologische bepalingen
(Microbiologisch Risk Assessments) die ontwikkeld worden om de kans op ziekte na
consumptie te kunnen schatten.


230


231
Acknowledgements
Finally, I reached the last pages of this thesis. I would therefore like to take the opportunity
to thank the various people who helped to finalize this dissertation.
First and foremost, I would like to thank my supervisors and co-supervisor for their input
during the past years. Mick thank you for giving me the opportunity to conduct this
research. Although you were at a distance you kept on following the progress of my work. I
have learned a lot from your practical experiences, especially from the infant formula
manufacturing environment. Martine, being my daily supervisor you always could improve
my draft presentations, reports and manuscripts. Leon and Marcel, almost two years from
the start you joined this project, thank you for supervising my work. Leon thank you for the
useful discussions, your insights in the microbiological food safety contributed to this broad
research approach. Marcel, your experience on predictive modeling came along at the right
moment during my research.
Also, the members of Scientific Advisory Board and co-authors contributed positively to
my research project. Frans Rombouts, Olivier Guilluame-Gentil, Rijkelt Beumer, Wilma
Hazeleger, Tjakko Abee, Annette Heuvelink, Enne de Boer, Han Joosten, Mats Peters,
Pieter Breeuwer, Ellen Brinkman and Roelina Dijk thank you for the valuable discussions.
I would also like to thank all the other staff of the Nestlé Food Safety Microbiology group
at the Nestlé Research Center in Lausanne, Switzerland for the nice collaboration and
sharing their experience on Cronobacter spp., resulting in jointed publications.
Together with Martine, I supervised several BSc and MSc students. Ingrid, Cecil, and
Etienne your work on the growth of Cronobacter spp. in reconstituted infant formulae
resulted in Chapter 6. Kishor and Samia part of your work is described in Chapter 7 and 8.
Dennis, Dominique, Gerben, Debbie, and Tvzetomira your work contributed to get more
insight in the behaviour of Cronobacter spp.. Thank you all for your interest in conducting
your thesis within my research project.
Furthermore, I would like to thank all my colleagues of the Laboratory of Food
Microbiology, for the nice conversations during the coffee and lunch breaks. Over the past

232
years many of you have left the department. Wilma, I enjoyed the time and the nice
conversations we had during the Food Micro 2004 Congress in Slovenia. Gerda van Laar,
Els, Ida, Maarten, and Ingrid thank you for listening to my complaints every now and then.
Gerda van Laar, I would also like to thank you for arranging all administration activities in
the project. Ingrid, thank you for the practical assistance, especially with all the enrichment
experiments.
Sabita, Annette and Esther thank you for reading through the various Chapters. Gerda van
Donkersgoed, we shared one office at RIKILT, thank you for your motivation to finish this
thesis.
Esther, I learned a lot from you, sharing the same office with you for some years was a very
pleasant time. I hope our friendship will last forever. Thank you for being one of my
paranymphs.
The cover is beautiful. Thank you, Dorothée, for designing it and for helping out with the
layout of this thesis.
Dear family thank you for all your support, now I can answer your question that I finally
have finished my “study”. Late aunt Rinia and Ine, you were always there for me during the
past years. Thank you! Geeta thank you for your support and the discussions we had
especially on the last part of this thesis.
And last but not least, Fons thank you for your patience, love and encouragement. Now we
will have more time to spend together, and together with Aditya, Sarada and Ishaan.



233
List of Publications
Peer-reviewed papers
Kandhai, M. C., A. E. Heuvelink, M. W. Reij, R. R. Beumer, R. Dijk, M. Van Schothorst,
and L. G .M. Gorris. 2010. A study into the occurrence of Cronobacter spp. in The
Netherlands between 2001and 2005. Food Control, doi:10.1016/j.foodcont.2010.01.007.
Kandhai, M. C., M. W. Reij, M. Van Schothorst, L. G. M. Gorris and M. H. Zwietering.
2010. Inactivation rates of Cronobacter spp. and other bacteria in dry powdered infant
formula at various temperatures. Accepted for publication in Journal of Food Protection
Kandhai, M. C., C. J. H. Booij, and H. J. Van der Fels-Klerx. 2010. A Delphi based expert
study to identify indicators for emerging Fusarium related mycotoxins in wheat based
supply chains. Accepted for publication in Risk Analysis.
Van der Fels-Klerx, H. J., S. Dekkers, M. C. Kandhai, S. M. F. Jeurissen, C. J. H. Booij,
and C. De Heer. 2010. Expert study to select indicators for identification of re-emerging
mycotoxins. NJAS - Wageningen Journal of Life Sciences, doi:10.1016/j.njas.2010.02.003.
Kandhai, M. C., P. Breeuwer, M. H. Zwietering, L. G. M. Gorris, and M. W. Reij. 2009.
Growth of Cronobacter sakazakii under dynamic temperature conditions during bottle
cooling. Journal of Food Protection, 72:2489-2498.
Van der Fels-Klerx, H. J., M. C. Kandhai, S. Brynestad, M. Dreyer, T. Borjesson, M.
Martins, E. Morrison, and C. J. H. Booij. 2009. Development of an European system for
identification of emerging mycotoxins in wheat based supply chains. World Mycotoxin
Journal, 2:2; 119-127.
Van der Fels-Klerx, H. J., M. C. Kandhai, and C. J. H. Booij, 2008. A conceptual model
for identification of emerging risks, applied to mycotoxins in wheat based supply chains.
World Mycotoxin Journal, 1:1, 13-22.
Kandhai, M. C., M. W. Reij, M. Van Schothorst, and M. H. Zwietering. 2006. The effect
of pre-culturing conditions on the lag phase and the growth rate of Enterobacter sakazakii
in reconstituted infant formula. Applied and Environmental Microbiology, 72:4, 2721-2729.

234
Guillaume-Gentil, O., V. Sonnard, M. C. Kandhai, J. D. Marugg, and H. Joosten. 2005. A
simple and rapid cultural method for detection of Enterobacter sakazakii in environmental
samples. Journal of Food Protection, 68: 64-69.
Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004. Occurrence of Enterobacter sakazakii in the food production
environments and households. The Lancet, 363: 39-40.
Kandhai, M. C., M. W. Reij, K. Van Puyvelde, O. Guillaume-Gentil, R. R. Beumer, and
M. Van Schothorst. 2004. A new protocol for the detection of Enterobacter sakazakii
applied to environmental samples. Journal of Food Protection, 67: 1267-1270.
Kandhai, M. C., M. W. Reij, L. G. M. Gorris, O. Guillaume-Gentil, and M. Van
Schothorst. 2004. Enterobacter sakazakii in factories and households. The Lancet. 364:
414-415.

Non peer-reviewed scientific publications
Kandhai, M. C., P. Breeuwer, E. G. Biesta-Peters, L. G. M. Gorris, M. H. Zwietering, and
M. W. Reij. 2009. Growth of Cronobacter spp. under dynamic temperature conditions
occurring during cooling of reconstituted powdered infant formula. In: V.K. Juneja & D.W.
Schaffner (Eds.), 6
th
International Conference Predictive Modeling in Foods Meeting
Abstracts and Information Booklet, Washington DC, USA. p. 46.
Kandhai, M. C., H. J. Van der Fels-Klerx, and C. J. H. Booij. 2008. Early Identification of
Emerging Mycotoxins in Wheat. In: The 21
st
international ICFMH Symposium, Food
Micro 2008, Evolving microbial food quality and safety, Aberdeen, Scotland, 1-4
September 2008. p. Z4.
Van Asselt, E. D.; H. Van der Voet, H. J. Van der Fels-Klerx, and M. C. Kandhai. 2008.
Salmonella serotypes in the broiler supply chain. In: The 21
st
international ICFMH
Symposium, Food Micro 2008, Evolving microbial food quality and safety, Aberdeen,
Scotland. p. AA3.


235
Dekkers, S., H. J. Van der Fels-Klerx, S. M. F. Jeurissen, M. C. Kandhai, C. J. H. Booij,
C. de Heer, and P. M. J. Bos. 2008. Development of a model to assess the occurrence of
mycotoxins in wheat, maize and nuts using a holistic approach-draft report. RIVM report
320111002/RIKILT report 2008.004/PRI report 186. Available at:
http://www.rivm.nl/bibliotheek/rapporten/320111002.html
Van der Fels-Klerx, H. J., C. J. H. Booij, and M. C. Kandhai. Indicators for emerging
mycotoxins. 2008. In: Van der Fels-Klerx, H. J. and C. J. H Booij (Eds.). European network
of information sources for an identification system for emerging mycotoxins in wheat based
supply chains (MYCONET). RIKILT – Institute of Food Safety, Wageningen UR,
Wageningen, The Netherlands. Report R2008.008. Available at:
http://www.rikilt.wur.nl/NR/rdonlyres/BDEEDD31-F58C-47EB-A0AA-
23CB9956CE18/83031/R2008008.pdf.
Kandhai M. C., M. W. Reij, and M. H. Zwietering. 2005. Enterobacter sakazakii: a risk
for infants? Prospects for health, well-being and safety. Fourth Nizo Dairy conference, The
Netherlands, p. 78.
Kandhai M. C., M. W. Reij, L. G. M. Gorris., M. Van Schothorst, and M. H. Zwietering.
2004. The effect of pre-culturing conditions on the lag phase of Enterobacter sakazakii. In:
New tools for improving microbial food safety and quality. Food Micro 2004 Congress
Slovenia, p. 335.
Kandhai M. C. 2003. Enterobacter sakazakii en zuigelingen voeding Symposium
'Voedselvergiftigingen en -infecties' : EFFI symposium , 2003-05-15 / Stichting EFFI, -
Wageningen : Stichting EFFI , 2003. p. 1–4.

236


237
Cur iculum Vitae
Morea Chantal Kandhai was born on April 3
rd
1974 in Paramaribo, Suriname, as the first
daughter of Johan Kit Rameshchandra Narain and late Jien-Tjauw Melie Narain-Tjon Sien
Foek. She completed her secondary education (Lyceum II in Paramaribo) in 1993. From 1994
until 2000 she studied Food Science at the Wageningen University (then the Agricultural
University). As part of her MSc study thesis research projects were carried out within the
Food chemistry and Food Microbiology specialization. During her internship at Numico
Research, Wageningen, The Netherlands, she worked on the effect of antioxidant
supplementation on the antioxidant status of healthy volunteers. In April 2001 she started
with her PhD project entitled: “Detection, occurrence, growth and inactivation of
Cronobacter spp. (Enterobacter sakazakii)” at the European chair in Food Safety
Microbiology and the Laboratory of Food Microbiology of the Wageningen University. The
results obtained during this research are described in this thesis. In 2006 she worked part time
as education coordinator at the Graduate school VLAG. From 2007 till 2009 she worked as a
researcher at RIKILT - Institute of Food Safety, on microbiological safety of various food
supply chains. Currently, Chantal is working at the Board for the Authorisation of Plant
Protection Products and Biocides (Ctgb).

238
Completed Tr aining Activities
Courses
VLAG Management on food safety and microbiological risks (2002)
VLAG Physiology of food associated micro-organisms (2004)
VLAG Food Fermentation (2004)
Basic and advanced statistic course (2003/2004)
Bionumerics software course (2002)

Meetings
EFFI symposium: Overleving en na-besmetting (2001)
NVVM symposium (2001)
HACCP en de garantie van voedselveiligheid (2002)
Safe: new emerging pathogens, Brussel, Belgium (2003)
EFFI symposium: Voedsel vergiftigingen en – infecties (2003)
Fourth International Conference on Predictive Modeling in foods, Quimper, France (2003)
NVvM symposium: Microbiologische voedselveiligheid: van gen tot wet (2003)
Nationale Hygiënedag (2003)
EFFI symposium: Microbiologisch onderzoek van Levensmiddelen (2003)
EFFI symposium: Microbiële gevaren: voorkomen en prevention (2004)
EUROFORUM: 5
e
Nationaal Congres Voedselveiligheid (2004)
NVVM symposium: Microbiologisch onderzoek van levensmiddelen: eigen werk (2004)
Food Micro: New tools for improving microbial food safety and quality, Slovenia (2004)
NIZO Dairy Conference: Prospects for health, well-being and safety, Papendal (2005)
EFFI symposium: Micro-organismen op weg (2006)

General courses
English for PhD students (2001)
Endnote course (2001)
Presenting in English (2003)
PE&RC Time and project management (2004)
Career perspectives (2004)
OWU Gespreksvaardigheden (2005)
OWU Basis cursus didactiek (2004/2005)
OWU Afstudeervak organiseren en begeleiden (2003)

Other activities
Preparation research proposal (2001)
Training at Nestlé Research Centre in Lausanne, Switzerland (2001)
Foreign students trip Slovakia, June 2004
General meetings Laboratory of Food Microbiology (2001, 2002, 2003, 2004, 2005)


239

240
The research described in this thesis was financially supported by Nestlé Research Center,
Lausanne, Switzerland.
















Cover picture and design by Dorothée Becu; www.becu.nu

Printed by GVO printers & designers B.V. | Ponsen & Looijen, Ede, The Netherlands.

Propositions
1. In the last decennium, detection methods for Cronobacter spp. have been improved, which has led to an increase in the isolation of Cronobacter spp. from a wide variety of food products and environments. This fact in itself is not a reason to consider Cronobacter spp. as an emerging pathogen. The use of sterile liquid formulae is an effective control measure to reduce the probability of infections in low and very low birth weight infants due to Cronobacter spp.. As, generally, pathogens are relevant in foods only at such low numbers that multiplication is required to allow their detection, the enrichment step must be regarded as the most important step in the entire detection procedure. Yet, relatively little research effort is dedicated to optimizing this procedure. Improvements in food safety management can only successfully be achieved in a food enterprise when there is a true food safety culture in the food enterprise. Children are an extraordinary gift and enhance their mother's career by improving her multitasking skills. Patience is the most important virtue in joint scientific writing.

2.

3.

4.

5.

6.

Propositions belonging to the thesis, entitled: “Detection, occurrence, growth, and inactivation of Cronobacter spp. (Enterobacter sakazakii)” M.C. Kandhai Wageningen, 23 April 2010.

Detection, occurrence, growth and inactivation of Cronobacter spp. (Enterobacter sakazakii)

M.C. Kandhai

Thesis committee Thesis supervisors
Prof. dr. L.G.M. Gorris Chairholder of the European Chair in Food Safety Microbiology Wageningen University Prof. dr. ir. M.H. Zwietering Professor of Food Microbiology Wageningen University

Thesis co-supervisor
Dr. ir. M.W. Reij Researcher and Lecturer at Laboratory of Food Microbiology Wageningen University

Other members
Prof.dr.ir. M.A.J.S. van Boekel, Wageningen University Prof. S. Fanning, UCD Center for Food Safety, Ireland Prof. dr. ir. A.H. Havelaar, University of Utrecht Dr. ir. A.E. Heuvelink, Food and Consumer Product Safety Authority, Zutphen

This research was conducted under the auspices of the Graduate School of Voeding, Levensmiddelentechnologie, Agrobiotechnologie en Gezondheid (VLAG).

Detection.C. Kropff.J. Kandhai Thesis Submitted in fulfilment of the requirements for the degree of doctor at Wageningen University by the authority of the Rector Magnificus Prof. M. dr. . occurrence. in the presence of the Thesis Committee appointed by the Doctorate Board to be defended in public on Friday 23 April 2010 at 4 PM in the Aula. (Enterobacter sakazakii) M. growth and inactivation of Cronobacter spp.

with summaries in Dutch and English ISBN 978-90-8585-556-9 . Kandhai Detection. NL (2010) With references. Wageningen University. Wageningen. 240 pages. Thesis. growth and inactivation of Cronobacter spp. (Enterobacter sakazakii).M.C. occurrence.

for tomorrow might never come” .To Fons. Whatever time you have is yours. Aditya. Sarada & Ishaan “Life is constantly changing.

At the onset of the work for this thesis. The detection method which was developed is based on two features of Cronobacter spp.Abstract The genus Cronobacter consists of Gram-negative. have been documented as a rare cause of outbreaks and sporadic cases of neonatal meningitis. necrotizing enterocolitis. Furthermore. and sepsis in infants with a high mortality. facultative anaerobic bacteria. Quantitative data on product contamination at manufacture. in environmental samples and a variety of food products manufactured or marketed in The Netherlands. Among these infants. in environmental samples. in reconstituted infant formula was necessary in order to provide data to be used in Microbiological Risk Assessment (MRA) dedicated to this particular food product. had been isolated from milkbased powdered formulae which have a direct link to the sub-population at greatest risk. more insight into the growth behavior of Cronobacter spp. Cronobacter spp. non-spore forming. combined: their yellow pigmented colonies when grown on tryptone soy agar and their constitutive -glucosidase. Cronobacter spp. which can be detected in a 4-h colorimetric assay. there was a need to more closely investigate whether and where Cronobacter spp. A selective enrichment method was developed for the rapid and reliable enrichment and detection of Cronobacter spp. The initially developed method and refinements thereof were applied for routine screening for the presence of Cronobacter spp. However. motile. and also growth after reconstitution are required in order to assess the risk associated with . those at greatest risk are infants less than 2 months of age. low birth weight (LBW) infants (< 2500 g). and immuno-compromised infants. The main objective of this study was to develop isolation and detection methods that would allow quick and reliable investigation into the occurrence of the micro-organism in potential sources. particularly pre-term infants. The detection method described in this thesis has been the basis for a series of media for Cronobacter spp. occurs in environments in which these powdered are manufactured and packed but also to investigate other sources which could lead to exposure of vulnerable sub-populations. during preparation. and was originally defined as one species “Enterobacter sakazakii” within the genus Enterobacter in 1980. that have recently been commercialized.

under dynamic temperature conditions.3 ± 18. The estimated lag time of the microorganisms was found to vary from 83. during cooling. predictive growth models were developed that capture key growth parameters. Minimum– and maximum temperatures estimated with the Secondary Rosso equation were 3. The research described in this thesis contributes to the existing knowledge on the natural habitat of Cronobacter spp. in reconstituted infant formula stored at any temperature below 47 C.7 h at 10 ºC to 1. The survival of two Cronobacter strains in dry powdered infant formula (PIF) was tested and compared to the survival of six other bacterial strains after inoculation and storage at several temperatures between 7 and 42 ºC. It is evident that storage of reconstituted formula in a refrigerator may require a significant amount of the time before the formula reach the targeted refrigeration temperature. during cooling strongly depends on the size of the container used for storage and that it may be prevented by limiting the volume to be cooled to portion-size only or by reconstituting at temperatures of 25 °C or lower.Cronobacter spp.6 ºC and 47. As growth rates of Cronobacter at refrigeration temperatures are relatively small.e. tools are needed to asses the micro-organisms growth potential as well as its inactivation (thus. Therefore. Next to that. its survival) due to specific control measures applied.73 ± 0. The models for growth rates and lag times as a function of temperatures obtained during this study allow estimating the potential growth of Cronobacter spp. Differences were found in the rate of survival that can be due to difference in the resistance to inactivation in dry environments between Cronobacter species. caregivers are advised to store reconstituted infant at low temperature as a control measure to prevent microbial growth. In this thesis. exposure. The effect of temperature on survival in PIF.6 ºC. Predictions showed that proliferation of Cronobacter spp. a mathematical model was built to predict the temperature profile and the resulting growth of Cronobacter spp. The models developed . respectively. and its occurrence and behavior in PIF. was described using both the Weibull distribution model and the log-linear model. i. which could be relevant to consider when establishing quantitative risk assessments on consumer risks related to PIF.43 h at 37 ºC and could be described with the hyperbolic model and reciprocal square root relation.

for quantifying the growth of Cronobacter spp. International and national governmental bodies may use these predictive models in risk assessments and to establish guidelines for health care professionals to provide effective hygiene training to parents and professional caregivers to ensure that PIF is prepared handled and stored appropriately. . in reconstituted formulae under various conditions can be applied in risk assessments set-up to estimate the probability of vulnerable sub-populations becoming ill after consuming infant formulae.

in The Netherlands The effect of pre-culturing conditions on the lag time and specific growth rate of Cronobacter spp.Contents Abstract Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Introduction A new protocol for the detection of Cronobacter spp. in environmental samples Occurrence of Cronobacter spp. under dynamic temperature conditions occurring during cooling of reconstituted powdered infant formula Survival of Cronobacter spp. in reconstituted infant formula Growth of Cronobacter spp. and concluding remarks 6 11 43 55 69 77 109 Chapter 7 133 165 193 227 231 233 237 238 Chapter 8 Chapter 9 Samenvatting Acknowledgement List of publications Curriculum vitae Overview of completed training activities . in comparison to other bacteria in dry infant formulae Discussion. in food production environments and households A study into the prevalence of Cronobacter spp. applied to environmental samples A simple and rapid cultural method for enrichment of Cronobacter spp.

.

Introduction 11 .

scientific communities. 12 . the International Commission on Microbiological Specifications for Food (ICMSF) categorised Cronobacter spp. as a severe hazard for restricted populations. 27. causing bacteraemia and meningitis and associated with necrotizing enterocolitis (NEC) (26. (Enterobacter sakazakii) is an opportunistic food-borne pathogen. In 2002. Interest has been increasingly raised in the public domain with notable activities of Codex Alimentarius. First report of an outbreak related to Cronobacter spp. 54). national governments. Pages 313-315 Figure 1. Volume 277. Issue 7172. that has been linked with serious infections in infants. Most cases of Cronobacter disease have been detected among newborn and very young infants in hospital nurseries and neonatal intensive care units (28). At the time the micro-organism was referred to as “the yellow pigmented Enterobacter cloacae”. 11 February 1961. The first reported cases attributed to this organism occurred in 1958 in England (See Figure 1) and resulted in the death of two infants (74). In most outbreaks. health care providers and the infant food industry. the source of Cronobacter spp. in which a source could be established. 69.Chapter 1 General Cronobacter spp. since an outbreak involving powdered infant formula (PIF) in the USA in 2002 (36). was proven to be (dry) powdered infant formulae (PIF) (5. 75. The Lancet. in scientific literature. 6. 76).

low birthweight (LBW) infants (< 2500 g). Information relevant for compiling a quantitative MRA relevant to Cronobacter spp. the various steps of this MRA 13 . survival and growth. in PIF is presented. In this thesis the aspects that have been investigated are presented within the framework of the MRA. 2) exposure assessment. and immunocompromised infants. such as inactivation. More recently. There is evidence that consumer groups vulnerable to Cronobacter infections are premature and low-birth-weight infants and those aged < 28 days (8. 26-28). and those infants that are less than 2 months of age. Suitable predictive models capture insights in the dynamics of the target micro-organisms. and possible risk-mitigating control measures. Most value is derived from being able to quantify individual parameters. A hazard is defined as a biological. Among these infants. as they may specifically require infant formula and may be more susceptible to infection (12. particularly pre-term infants. In Figure 2. 22). Microbiological Risk Assessment approach Microbiological risk assessment (MRA) is widely used to characterise health risks associated to the potential presence of microbial hazards in foods. The magnitude of the associated risk is obtained as well as insight in risk-contributing factors. Infants of HIVpositive mothers are also at risk for Cronobacter infection. An MRA consists of four stages: 1) hazard identification. 3) hazard characterization and 4) risk characterization (49).Introduction resulting in life-threatening diseases or substantial chronic sequelae or effects of illnesses of long duration (37). 22). 21. To determine the magnitude of the risk through MRA predictive models are often used as tools in such assessments. the Codex Committee for Food Hygiene (CCFH) of Codex Alimentarius and the BIOHAZ scientific expert panel of the European Food Safety Authority (EFSA) established that all infants at or below 12 months are at particular risk for Cronobacter infections (21. chemical or physiological agent in food (or condition of food) with the potential to cause an adverse health effect (11). Resulting parameters can be applied to estimate the consumer risk and to give directions to which preventive measures should be advised (51). those at greatest risk are neonates (< 28 days).

in powdered infant formula. Schematic view of elements of microbiological risk assessment for Cronobacter spp. 14 . Grey boxes indicate the topics that were investigated in more detail in this thesis.Chapter 1 approach are given with sub-topics indicated in the various boxes. All boxes are described in this introduction section. The grey boxes indicate the research topics focused on in this thesis. Figure 1.

Cronobacter infections are very rare and often underreported. 27) that have provided a scientific basis for creating a new code for hygienic practices by CCFH (12). aimed at providing risk management guidance at the international level. invade the blood stream (bacteraemia and/or sepsis) and/or invade the central nervous system. pharmaceutical treatment. Since 1961 and up to July 2008. A neonate is a newborn infant. An immunocompromised infant may have an immunodeficiency of any kind. especially in developing countries (23). low birth weight.Introduction Epidemiology The epidemiology of Cronobacter spp. and a premature is an infant that was born prior to 37 weeks of gestation. clinical conditions. According to Codex Alimentarius. Specific international attention has been given to the safety of food for infants and young children as it relates to the possible presence of Cronobacter spp. at least 29 cases (19%) resulted in death (28). while young children are persons from 12 months up to the age of three years (36 months) (12). HIV status.e. Of these 156 cases. and may therefore be particularly vulnerable to opportunistic infections in addition to normal infection that could affect anyone. 15 . infections from all parts of the world have been reported in the published literature and in reports submitted by public health organizations and laboratories. products for infants up to 12 months of age are considered. A number of factors could contribute to an infant’s immune status. less than four weeks old. 156 documented cases of Cronobacter spp. In this code. causing celebritis and/or meningitis. may affect the intestines (causing necrotizing enterocolitis). nutritional status such as vitamin status. Because the prevalence of these factors varies among countries. is poorly understood. The Food and Agriculture Organization of the United Nations (FAO) and the World Health Organization (WHO) have established two international risk assessments on the topic (26. and premature birth. an infant is a person not more than 12 months of age. Meningitis has a high mortality rate (42%) and many of the survivors (74%) suffer from neurological complications (66). there is a wide variation in the prevalence of immunocompromised infants globally. in powdered formulae intended for this consumer group. Diseases caused by Cronobacter spp. i.

environmental sources of contamination should not be excluded (58). invasive infections seems to be much higher among infants than older age groups such as young children. Though Cronobacter spp. children or adults (26. the illness started as soon as 3 to 4 days after the initial exposure to the implicated formula (2. Only in two outbreaks. documented Cronobacter spp. as published by the Food and Agriculture Organization of the United Nations (FAO) and the World Health Organization (WHO) (28). It has been reported that in outbreaks associated with powdered infant formulae. 33. infection. however both the route of exposure and the incubation period are generally unclear. Many infections in newborns are transmitted from mother to child. In most cases that infection has occurred. 75). 48). The group of infants at risk for necrotizing enterocolitis. especially those with a low birth weight (< 2500 g) or a very low birth weight (< 1500 g). however. a clear relationship was shown between Cronobacter spp. For several years passage of the organism through the mother’s birth canal was therefore suspected to be the source of Cronobacter spp. A recent survey showed a mortality rate of invasive neonatal Cronobacter infections of 27%. as well as insect or animal reservoirs. are thought to be at greater risk for severe infection than more mature infants. 69. while Cronobacter bacteremia has occurred in all age groups (8). 48) may indicate potential other sources such as home environments. other than from powdered infant formula and human illness are less well understood. infections in immunocompromised adults (34. Cases of Cronobacter spp. occurring in neonatal intensive care units. seems to be similar to the group of neonates which is defined to be at risk for bacteraemia and meningitis (66). The incidence of Cronobacter spp. While relationships between sources of Cronobacter spp. 15). The lethality was also calculated for Cronobacter meningitis (42%). Table 1 shows an overview of documented outbreaks and cases of infections in infants and children per year per country caused by Cronobacter spp. food and food manufacturing environments. infections have been associated with contaminated powdered formulae. meningitis are reported exclusively among infants. In various reports mortality rates between 40 and 80% are reported (48). Cronobacter 16 . isolates from patients and isolates from unopened cans of powdered dry infant formulae of the same batch as consumed by the patient (6.Chapter 1 Premature infants.

whereas the annual incidence rate among the low birth weight infants (< 2500 g) was found to be higher at 8. Cronobacter spp. outbreaks from 1961 to 2008 (28) as presented in Table 1. Table 1.4 per 100. except for *. invasive infections was estimated for the United States of America (26).e. Considering the overview of invasive Cronobacter spp. an another study estimated an incidence rate of 9. In the first expert meeting of the FAO/WHO on Cronobacter spp. five of which were invasive. Analyzing the tota. cases 2 1 1 2 8 1 1 11 2 1 3 Deaths 2 6 4 1 1 17 . where the child was 13 months of age Publication Year 1961 1965 1979 1981 1983 1984 1985 1987 1988 1989 1989 Country UK Denmark USA USA Netherlands Greece USA Greece USA Portugal Iceland No. an annual incidence rate of 1 per 100. In the age group 6-11 months six cases are well documented.7 per 100. Similarly. outbreaks and cases in infants and childeren reported until July 2008 (adapted from reference (28)). an average mortality rate of 19% can be calculated (i.e. Overview of the number of invasive Cronobacter spp.Introduction septicaemia (< 10%). children < 12 months of age) for Cronobacter spp. three had other medical problems. From the five invasive cases. but the incidence rate appears to be low. In the age group 12–35 months only two cases have been well described. In all cases the age at illness onset was < 1 year. number of cases did not yield consistent information about the cases per age group in the population.000 infants (27). Cronobacter necrotising enterocolitis (19%) (31).000 infants (i. 29 death in 156 cases).000 very low birth weight infants (< 1500 g) (72). in PIF. so currently it is not possible to quantify existing potential differences in the susceptibility of infants of different age groups. infections have occurred in both hospital and home settings.

cases 4 1 1 1 1 1 5 3 12 2 3 20 1 1 1 5 2 1 5 2 1 5 1 9 9 1 2 1 18 3 1 156 Deaths 2 2 1 1 1 2 1* 1 4 29 18 .Chapter 1 Table 1. Continued Publication Year 1989 1990 1991 1994 1997 2000 2000 2001 2001 2001 2002 2002 2002 2003 2003 2003 2004 2004 2004 2005 2005 2006 2006 2006 2007 2007 2007 2007 2007 2008 2008 Total Country USA USA USA Germany Scotland USA Brazil USA Belgium Israel Israel USA Belgium Brazil Hungary USA USA Hungary New Zealand USA Hungary USA Hungary France USA Canada India Spain France USA Japan No.

indicating that the pathogenicity between strains tested may differ. powdered infant formula has been identified as the source of infection. 19 . very little information is available on the virulence factors of Cronobacter spp. infection may cause a highly lethal syndrome of bacteraemia and meningitis with involvement of the central nervous system in neonates. In several cases and outbreaks. occur in the intestinal tract of humans and animals. and its pathogenic mechanisms. Pagotto et al. 8 food and 1 Type strain). its relevance to pathogenesis has not been established yet. It is known that Cronobacter spp. each of the species have been linked retrospectively to clinical cases of infection in either infants or adults (28). relatively little scientific attention has been given to unravelling the mechanism of pathogenicity or to potential virulence factors of Cronobacter spp. therefore the entry for the infection can be gastrointestinal (64). showed lethality at a dose of 107 CFU per mouse (2 out 8 suckling mice (20%) died ) (64).. These data have been extrapolated to estimate the minimum infectious dose for human infants to be 105 CFU in a number of studies. sakazakii have been shown to produce an enterotoxin (64) however. (2003) used a suckling mice model to determine the pathogenicity of 18 Cronobacter isolates (9 clinical. and are frequently a cause of nosocomial infections. causing illness in humans is not known and represents a challenging aspect requiring further investigation. Orally dosed suckling mice. showing that a dose of 108 CFU/mouse was lethal to 58 out of 69 (84%) suckling mice within 3 days after intraperitoneal administration.Introduction Pathogenicity Enterobacter species are widely distributed in nature. A minimal number of 105 CFU of certain Cronobacter strains was lethal to 5 out 20 (25%) suckling mice. To date. Currently. Some strains of C. (63). The dose-response relationship for Cronobacter spp. There is no data available that any one of the Cronobacter species is more pathogenic than another.

Cronobacter malonaticus sp. nov. Throughout this thesis the name Cronobacter spp. nov. biochemical reactions. nov. nov. comprising at least five genomospecies. 20 . cloacae in DNA-DNA hybridization. 7. Cronobacter dublinensis sp. covering the various strains which were earlier named Enterobacter sakazakii. (Biogroup 16). Cronobacter muytjensii sp. In 2007. including three subspecies. pigment production and antibiotic susceptibility. Figure 3 gives an overview of the place of the genus Cronobacter within the Enterobacteriaceae family. fluorescentamplified fragment length polymorphism (f-AFLP) and DNA-DNA hybridisation. employing full-length 16S rRNA gene sequencing. with the wild type Biogroup 1 being the most common (29). nov. The reclassification to Enterobacter sakazakii was based on differences from E. sakazakii were described. 8. is used. Also a new biogroup (biogroup 16) was identified (41). was referred to as the “yellow pigmented Enterobacter cloacae”. Cronobacter turicensis sp. 11 and 13). comb. This genomospecies has not been associated with a specific biogroup. Based on this approach and the different phenotypic profiles. the novel Cronobacter species are now divided in 16 biogroups. (Biogroup 1. 10. 15 biogroups of 57 strains of E. The following five species are distinguished: Cronobacter sakazakii gen. (Biogroup 5. Based on DNA-DNA hybridization. Cronobacter spp.4. Two Cronobacter strains appear to be a separate genomospecies and are indicated as Cronobacter genomospecies I. 12). (Biogroup 15). are Gram-negative. nov. (Biogroup 6. peritrichous non-spore forming.Chapter 1 Taxonomy Cronobacter spp. nov. ribotyping. is based on a polyphasic taxonomic approach. straight rods. The latest classification into Cronobacter gen. motile. Until 1980.. The micro-organism was named after the Japanese microbiologist Riichi Sakazaki in honor of his work on the enteric bacteriology. within the family Enterobacteriaceae and originally belonging to the genus Enterobacter. 9. there was a proposal for a novel genus Cronobacter. 14).

the reader is referred to Iversen et al. lactaridi subsp.. and Cronobacter dublinensis subsp. Healy (personal communication). Adapted from B. including (sub) species. 21 . dublinensis subsp. including their new names. 2008 (42). (Biogroup 12). (Biogroup 10). nov. nov. Schematic view of the place of Cronobacter spp. Figure 2. within the Enterobacteriaceae family. Table 2 provides a list of the Cronobacter strains used throughout this research. (Biogroup 6). and Salmonella spp. Cronobacter dublinensis subsp.Introduction A number of subspecies of Cronobacter dublinensis are: Cronobacter dublinensis subsp. lausannensis subsp. nov. For the full taxonomic description on the genus Cronobacter.

Concentrations of the organism found in powdered infant formulae were between 0. highly specific methods are required for enumeration/detection and identification that can be used for both product and environmental samples and that allow quick isolation of Cronobacter spp. lactaridi Cronobacter spp.2 and 92 CFU/100 g. According to Iversen et al. in PIF and powdered formula producing facilities. there was a special need to improve the existing methods for isolation and detection of 22 . 60). 2008 (42). Methodology Detection and isolation Cronobacter spp. Prevalences varying from 2 to 14% have been reported (35. Cronobacter spp.. Overview of the Cronobacter spp. At the start of the studies documented in this thesis. dublinensis Cronobacter muytjensii Cronobacter turicensis Cronobacter malonaticus Cronobacter dublinensis subsp. levels greater than 1 CFU/g have not been reported (25).Chapter 1 Table 2. Because of the low prevalence of Cronobacter spp.a strains investigated in this thesis Old strain taxonomy and culture collection code(s) Enterobacter sakazakii NCTC 8155 Enterobacter sakazakii NCTC 9238 Enterobacter sakazakii NCTC 11467 = ATCC 29544 = DSM 4485 Enterobacter sakazakii NCTC 9844 Enterobacter sakazakii NCTC 9846 Enterobacter sakazakii NCTC 9529 Enterobacter sakazakii DSM 18705 Enterobacter sakazakii ATCC 51329 Enterobacter sakazakii DSM 18703 Enterobacter sakazakii DSM 18702T Enterobacter sakazakii DSM 18706T Enterobacter sakazakii DSM 18707T Enterobacter sakazakii MC10 Enterobacter sakazakii MM9 a New strain taxonomy Cronobacter sakazakii Cronobacter sakazakii Cronobacter sakazakii Cronobacter dublinensis Cronobacter dublinensis Cronobacter genomospecies 1 Cronobacter dublinensis subsp. lausannensis Cronobacter dublinensis subsp. can be present in low numbers in powdered infant formulae. 56..

The third is streaking of (selective) enrichment broth onto selective solid media. The success rate of the above mentioned basic detection and isolation protocols depends on: 1) the number and the state of the micro-organisms in the sample. 57. in powdered infant formulae Preenrichment Buffered Peptone Water (BPW) De-ionized Water Enrichment (usually selective) Enterobacteriaceae enrichment (EE) broth EE broth Isolation Violet Red Bile Glucose Agar (VRBGA) plates VRBGA plates Identification API 20E and yellow colonies on Tryptone Soy Agar (TSA) plates API 20E Ref. depending on the number of cells expected in the sample. After the selective enrichment step. 2) the selectivity of the media (the balance between the inhibition of the competitors and inhibition of the target organism). Methods available in 2001 for detection of Cronobacter spp. (Chapter 9). but significantly less to the species or group of species to be isolated. 23 . There are three basic steps in the detection of micro-organism in foods. 3) the conditions of the incubation (temperature. The first one is pre-enrichment in a non-selective medium allowing recovery of sub-lethally damaged cells. shown in Table 3 (56. 73). These procedures may be used in different combinations. 4) the visual distinction of the isolation medium (the distinction between the competitive flora and the target organism) (4). presumptive Cronobacter spp. The second is selective enrichment containing one or more compounds that are inhibitory to the majority of micro-organisms. time. Many methods and media have been described since then to isolate Cronobacter spp. are still under development to increasingly better meet the need to detect levels of approximately 1 cell per 100 g of PIF. oxygen availability). isolated on agar need to be confirmed. Table 3. (56) (60) Methods for detecting Cronobacter spp.Introduction Cronobacter spp.

52. (57). These are based on amplifying specific fragments of DNA. Also these methods does not allow for isolation of the strains. Furthermore. Recently. has remained the basis for many detection media (22). 77). Recently. The PCR confirmation method depends on the availability of two short oligonucleotide primer sequences that will hybridize to opposite strands of heat-denatured DNA at either end of the region which eventually be probed. 62). Furthermore. A DNA polymerase then catalyses the extension of the primers to produce two double-stranded copies of the region of interest. molecular methods can not distinguish living from dead cells and organisms killed for example by a heat process can still be detected. DNA fingerprints obtained with standerized DNA. Molecular subtyping of bacteria by profiling either proteins or nucleic acids could be a useful tool to investigate for example epidemiological relationships of isolates (25). Much effort has been put on the development of molecular sub-typing techniques. biochemical tests. 53. also several genotypic methods including the Polymerase Chain Reaction (PCR) have been developed. detected on selective plating media. A potential disadvantage of molecular techniques can be their sensitivity that can be influenced by food components. Nevertheless. 65. from other species by biochemical differentiation. 24 .. like Pulsed Field Gel Electrophorese (PFGE) (71). described by Muytjens et al. was firstly employed in the research conducted in this thesis as a suitable confirmation method for Cronobacter spp. allowing detection of a single copy of the target sequence (18.based protocols could be applied for direct comparison of isolates in outbreaks (70). a number of PCR probes for conserved genes such as ompA have been published (59). The biochemical characteristic of α-glucosidase production by Cronobacter spp.Chapter 1 Confirmation methods To confirm presumptive Cronobacter spp. ribotyping and random amplification of polymorphic DNA (RAPD) (15. A specific feature among Enterobacter species. such methods can still be useful in rapid detection of the presence of very low numbers of target micro-organisms. namely the presence of α-glucosidase activity. isolated from PIF – producing facilities (Chapter 2). 68. such as API 20E can de used that distinguish Cronobacter spp.

specific food safety management systems such as the Hazard Analysis Critical Control Points (HACCP) system. FUF is defined as a food intended 25 . nutritious and wholesome foods with an adequate shelf-life and at reasonable costs to the consumer. by itself. Powdered formulae are divided in powdered infant formula (PIF). have been deployed successfully. the effectiveness of various control measures is monitored to determine whether the process parameters for the control measures remain within predetermined limits designed to achieve food safety (50). Implementation of general food safety management systems such as Good Hygiene Practices (GHP) and Good Manufacturing Practices (GMP) are considered in many countries as pre-requisite conditions for facilities preparing or manufacturing foods. To better assure safe food production. and dietary foods for special medical purposes. To design adequate control measures and set-up effective food safety management systems it is important to understand the different possible contamination routes of relevant hazards as well as the efficacy of specific control measures to control them. Production of Powdered Infant Formulae Whereas it is generally accepted that breastfeeding is the best option for the nutrition and the health of infants and young children. there may be situations where a mother can not breastfeed her baby or chooses no to do so. the nutritional requirements of infants during the first months of life up to the introduction of appropriate complementary feeding. Food production processes often rely on control methods such as pasteurization or sterilization to reduce the number of vegetative pathogens and/or spore-formers in the final food product. Using these food safety management systems. In such cases. PIF is defined as a breast milk substitute specially manufactured to satisfy. By taking adequate control measures and deploying sound assurance systems to manage food production. powdered formulae represent useful alternatives to replace breast feeding either partially or totally and these are therefore formulated to meet the nutritional needs of infants (78). the food industry helps to minimize risks to consumers associated with the potential presence of foodborne hazards in foods. follow-up formula (FUF).Introduction Production Food production aims at delivering safe.

all separate ingredients are dry blended to obtain the final product. For a detailed description of the production process the reader is referred to Cordier (2008) (16). The composition. in powdered infant formulae and reconstituted infant formula have been issued in joint reports by the Food and Agriculture Organization of the United Nations (FAO) and the World Health Organization (WHO) (12. applied by the European Commission. all unprocessed raw materials as well as separately processed ingredients are handled as a liquid product that is heat treated. In the wet-mixing process. intended for dietary management of patients and to be used under medical supervision (12). EC No 1441/2007. which is further used for the manufacture of different finished products. the wet-mixand the combined process. Dietary foods for special medical purposes are defined as foods for particular nutritional uses. quality and labelling requirements of powdered infant formulae are clearly defined in national. minerals. A schematic overview of PIF production is given in Figure 4. Guidance and recommendations concerning control and elimination of Salmonella spp. Production of powdered formulae can follow three procedures: the dry-mix-. Monitoring for Enterobacteriaceae is used as 26 . The type of process used depends mostly on the product manufactured and the processing facility. The dry mixing is done when heat sensitive ingredients such as vitamins. starch. regional as well as international standards or regulations (12). The most significant difference between PIF and FUF is that FUF may contain a wider variety of dry-mix ingredients. Powdered formulae are produced with stringent control measures and strategies to prevent post-process contamination. according to the formulation. The microbiological criteria.Chapter 1 for use as a liquid part of the weaning diet for infants from the 6 th month onwards and for young children. In the dry-mix process. 27). which is further handled up to the filing process. 26. Ingredients as such have all been submitted to some thermal inactivation step during their manufacture and must fulfil the same microbiological requirements as the final product. The combined process may include and combine different mixing steps to obtain the final formulation. carbohydrates and others need to be added. In the combined process the unprocessed raw material and part of the ingredients are processed according to the wet-mix process to obtain base powder. and Cronobacter spp. dried and then further handled up to the filling stage. for PIF are shown in Table 4.

and other Enterobacteriaceae may be established (22). and other Enterobacteriaceae..Introduction process criterion that indicates the hygienic status of the production. Table 4. Current European Union microbiological criteria for dried infant formulae for infants up to 6 months of age and for formulae for special medical purposes (24) Micro-organism(s) Process hygiene criteria Food safety criteria Enterobacteriaceae (in 10-g samples) Cronobacter spp. unless a correlation between microorganisms has been established at an individual plant level. while detection of Salmonella spp. (in 10-g samples) Salmonella (in 25-g samples) Microbiological limit n c m 10 0 Absence in 10 g 30 30 0 0 Absence in 10 g Absence in 25 g n is the number of units comprising the sample. c is number of acceptable sample units with values above m. at individual plant level such a correlation between Cronobacter spp. is advised. and that there is also no apparent relationship between Cronobacter spp. the batch must be tested for Cronobacter spp. and Cronobacter spp. For PIF for infants of 6 months and dietary products for special medical purposes parallel testing for Enterobacteriaceae and Cronobacter spp. The detection of Enterobacteriaceae is a process hygiene parameter. If Enterobacteriaceae are detected in any of the products samples tested in such a plant. Salmonella and other Enterobacteriaceae in PIF and FUF undertaken by the scientific Panel on Biological Hazards (BIOHAZ Panel) of the European Food Safety Authority (EFSA) concluded that there is no correlation between Salmonella spp. However.. The recent review regarding the relationship between Cronobacter spp. and other Enterobacteriaceae. 27 . include food safety parameters (24).

including both wet.Chapter 1 Figure 3. Contamination during manufacture could in particular occur in the steps highlighted in grey. 28 .and dry mixing. Schematic view of a production process for PIF.

Cronobacter spp. are Cronobacter spp. and its occurrence generally is rare.. 27). Following the production scheme depicted in Figure 4. through air containing dust or skin particles of food handlers (17). currently not considered to be feasible in the high-hygiene zone (27). 54). could enter powdered infant formulae via the processing environment or the processing line and through the addition of (contaminated) ingredients such as vitamins and starch after the heat-inactivation step (16. The level of Enterobacteriaceae has been used for decades as an indicator for process hygiene. 29 . The relevant hazards in PIF requiring appropriate control measures. Contrary to Salmonella spp.) are known to be present in powdered infant formulae processing environments and in homes or other places where PIF is reconstituted. the production steps where contamination could occur are depicted in grey. during manufacture of PIF is. In this Figure. has been detected in factory environments (46). PIF has been found to be possibly contaminated during addition of ingredients after the pasteurization process. during the packing process or during reconstitution of the powder at homes or in the hospitals (14. Indeed. sporadically. 45). Even though Good Practices (GMP and GHP) and correct HACCP plans are adhered to (26. 27). (26. 27). 16. Micro-organisms belonging to the family of Enterobacteriaceae (which includes Salmonella spp. in the processing environment implies that there may always be a chance of product contamination. such as hospitals (14.. PIF may be contaminated with this micro-organism. and Cronobacter spp. Literature data have shown that the levels of Salmonella spp. powdered formulae can also be contaminated during preparation of the feed at homes and at hospitals (27). As the microorganism occurs in homes and hospitals. 39. contamination from the environment could occur via aerosols. in the processing of PIF are much reduced in properly maintained high hygiene zones. due to its ubiquitous nature. complete elimination of Cronobacter spp. Cronobacter spp. Next to contamination via ingredients. The presence of Cronobacter spp. and Salmonella spp.Introduction Contamination routes relevant for PIF production Severe illness and sometimes death of infants has been attributed to PIF that has been contaminated with Cronobacter spp. as pointed out by the FAO and WHO consultations.

was detected on kitchen equipment used for preparing powdered infant formulae and also on apparently clean equipment (58. This processing step is particularly important because Cronobacter spp. Cronobacter spp. 19. soil and living vectors like insects or small vertebrates. Furthermore. In these factories the numbers of Cronobacter cells and the incidence of product contamination could increase after a wet cleaning process (16). 44.. In homes. fermented. is a vegetative micro-organism. As air can be contaminated with bacteria. Since Cronobacter spp. was isolated from household clothes (47). product–air contact could thus cause contamination after the pasteurization step. literature data showed that Cronobacter spp. It was also isolated from fresh vegetables and spices (3). frozen. Prevalence outside factories In order to identify potential vehicles for transmission.g. contamination may occur intrinsically. 61). In several investigations Cronobacter spp. The micro-organism seems to be ubiquitously present as it has been isolated from a wide spectrum of environmental sources and food products (22). it is important to investigate potential environmental reservoirs of Cronobacter spp. was able to survive in a dish brush (58). due to adherence to dust particles. e. Cronobacter cells are able to survive in dry environments (45) such as powdered formulae producing facilities (55). may be introduced during dry cleaning of the spray dryer and also via air that is in contact with the product after the spray drying step. water droplets or even skin particles. may survive the spray drying process (1). indicating that good hygiene practices are necessary to avoid contamination at homes and in 30 . Raw foods of animal origin may be contaminated additionally via the (faecal) micro-flora of the food source animal itself (30). the pasteurization step should be sufficient to inactivate this micro-organism (9. 71). due to its endophytical presence in plants or through contact with water. Primary Cronobacter spp. It has been detected in raw and fresh products of animal and vegetable origin as well as in processed and prepared foods such as dried. ready-to-eat and street foods.Chapter 1 Prevalence in factories An adequate heating step is very important to significantly decrease the bacterial number in raw milk. Cronobacter spp. smoked. cooked or fried products. 40.

Previous studies showed the occurrence of Cronobacter spp. Inactivation of Cronobacter spp. the preparation practices vary according to local arrangements and the availability of adequately trained personnel (23). cells were able to survive for at least twelve months in powdered infant formulae under favourable conditions (13. specific preparation and handling instructions are mentioned on the packaging label of powdered infant formulae. cells in different types of powdered formulae and vegetable based 31 . 20. and also that Cronobacter spp. it is necessary to consider possible contamination during preparation and handling of prepared formulae. generally there is a gradual die-off. in dry powder Cronobacter spp.Introduction hospital kitchen facilities. and other micro-organisms that may be present in such products. leading to infection and illness in infants. 20). In hospitals. However. 13. but there is a difference in the thermal tolerance between Cronobacter strains (19. Growth in PIF is not possible after the spray drying step because the water activity is too low to allow bacterial growth. 56. To avoid growth and further contamination. Although the micro-organism may be able to survive in PIF. Acidification could reduce the concentration of Cronobacter spp. depending on the storage time and conditions. has an unusual surviving ability under dry conditions compared to other Enterobacteriaceae (9. Product handling As the manufacture of commercially sterile PIF is not feasible using current processing technologies. cells in the powdered infant formulae (39. Nevertheless. Immediate consumption or rapid cooling and storage at low temperature are critical control measures to prevent microbial growth once PIF is reconstituted. This risk is particularly relevant when prepared food is handled or stored incorrectly. 60). Because of its ubiquitous nature. several investigations showed that professional and home caregivers have their own way of preparing the formula. 61). 32). reconstituted PIF forms an excellent growth medium for Cronobacter spp. there is a low but potential risk of infection to infants through consumption of PIF.

recontaminate and grow. in reconstituted PIF. A wide variety of factors such as inoculum size. being the lag phase. Quantification of survival rates of Cronobacter spp. This information could lead to recommendations to health professionals and parents on proper preparation. Growth of Cronobacter spp. 75). as only low numbers of the micro-organism have been detected in powdered infant formulae (25). 36. depending on the ability of cells to form new protoplasm from available nutrients. 15. Bacterial growth is defined as an orderly increase in the quantity of cellular constituents.Chapter 1 products (43. time necessary to recover from physical damage or shock after transfer from 32 . in reconstituted infant formulae at various temperatures is scarce. Most probably. in powdered infant formulae (35. In order to predict the microbial behaviour in a food product and to determine the effect of growth on the product’s safety and quality. PIF is not a sterile product and several investigations have shown the occurrence of cells of Cronobacter spp. 67). it is a necessity to have good insights in the production. bacteria can adapt to their new environment and injured bacterial cells can recover. and storage of reconstituted PIF to ensure the product is safe when fed to the infants. the stationary phase and the death phase. 39. To determine its effect on risk it is therefore necessary to quantify the growth of Cronobacter spp. multiplication of the organism has taken place in these cases. 56) leading to it being associated to outbreaks after consumption of reconstituted infant formula (5. 69. Bacterial growth can be divided in four phases. handling. In the lag phase. 27). the exponential phase. storage and handling conditions of PIF and to relate this understanding to quantitative data regarding the micro-organisms ability to survive. in reconstituted infant formulae As acknowledged by the Food and Agricultural Organization and the World Health Organization (26. 8. as it allows determining likely levels of the micro-organism in PIF dependent on for instance different scenarios of time and temperature during storage and distribution. isolates in dry PIF may provide useful data for microbiological risk assessments and for product/process design studies. Literature data about the growth potential of Cronobacter spp.

such as the specific growth rate as a function of temperature. Quantification of growth parameters. Mathematical models are useful tools to describe the growth (7). can influence the length of the lag phase. initially the cell number no longer increases but is kept in balance as a result of cells dying (due to exhaustion of available nutrients and/or accumulation of inhibitory metabolites or end products or exhaustion of space) and of cells multiplying. later in the stationary phase a gradual reduction in cell numbers may be seen as the balance is lost. and fluctuations in temperature during storage and preparation may significantly affect the potential for outgrowth. Temperature is one of the major environmental factors influencing microbial growth. However. In the case of Cronobacter infections. cells divide at a constant rate depending on the composition of the growth medium and incubation conditions such as temperature. either qualitative or quantitative. as an extended lag phase can be a means to prevent growth and multiplication of bacteria (10). in reconstituted infant formulae at various timetemperature scenarios. The likelihood and severity of illness is directly related to the number of micro-organisms ingested. The minimum infectious dose (MID) of Cronobacter spp. In the stationary phase. In other studies it is speculated that a reasonable estimate for infection might be close of that postulated for Escherichia coli 33 . In the exponential phase. using available dose-response relationships. is extrapolated from animal models (64): it estimated that high levels of the organism (> 105 CFU/feeding) are necessary to cause illness. for infants.Introduction other conditions. Hazard Characterization In the hazard characterization step the possible impact of an identified hazard on consumer is determined. though in most cases this is a very variable relationship across consumer groups and individuals as well as possibly across different strains of micro-organisms. time needed for synthesis of essential (co-)enzymes or division factors and time required for synthesis of new (inducible) enzymes that are necessary to metabolize the substrates present in the medium. The duration of the lag phase is of particular interest. help calculating possible outgrowth of Cronobacter spp. the infectious dose relationship for humans has not been clearly determined yet.

Outline of this thesis and the research conducted This research project aimed to get more insight into the habitat and growth characteristics of Cronobacter spp. risk manager could decide whether to inform the consumer to implement actions to reduce the risk on infection. isolates were confirmed using a Polymerase Chain Reaction (PCR) method and typed with Pulse Field Gel Electrophoresis (PFGE) fingerprinting analysis (Chapter 5). i. i. and to study the factors that might ultimately affect consumer risk. was 34 . The newly developed methods were then applied to dry environmental samples from households and food producing facilities (Chapter 4).. Prior to this study. in dry powder and growth after reconstitution of the powdered formula (Figure 2). Retrospective studies on epidemiological data and animal models could help to better estimate the dose-response relationship of Cronobacter spp. circa 1000 CFU (38). including regulatory measures. in particular the survival of Cronobacter spp. As reported in Chapter 6. To better explore the varied occurrence of the microorganism in food products. To further improve enumeration and detection of Cronobacter spp. no specific enrichment method was available and therefore an enrichment method was developed as described in Chapter 2.e reconstituted infant formula is consumed.e. Furthermore. in foods marketed in The Netherlands was conducted.Chapter 1 O157:H7 and Listeria monocytogenes 4b. In this survey. the effect of the pre-culturing conditions on key growth parameters of Cronobacter spp. an extensive survey on the occurrence of Cronobacter spp. a more specific detection method than available at the onset of the study was designed based on the presence of αglucosidase activity in the micro-organism (described in Chapter 3).. Results from risk characterization and especially comparisons of the impact of different scenarios of exposure and possible mitigation could be used by risk managers to decide on appropriate controls. Risk characterization In the risk characterization step the results of the previous three phases are integrated to give an estimate of the probability of becoming ill after the food product.

Introduction quantified. O. J. N. B. A. J. Breeuwer. R. and I. M. Int. Dairy Technol. 2003. Baumgartner.. Simhon. Microbiol. Detection and frequency of Cronobacter spp. 35 . and J. C. The Netherlands. S. Growth of Cronobacter spp. Preminger. Jordan. Arad. Braden. 1997. Ludvigsson.. Int. 27:2054-2056. Steingrimsson. 2009. J. Baranyi. Lardeau. Microbiol. J. 4. Clin. Bar Oz. Bowen. E. Emerg. Biering. 2006. Clark. 3. was then studied in reconstituted infant formulae cooled in refrigerators with stagnant air (normal household refrigerators) and in a refrigerator equipped with a fan (circulating air). Finally. growth parameters obtained earlier were used in different scenarios to predict the possible growth of Cronobacter spp. Iversen. Fox. 2001. 8. A. and C. 5. 90:356-358. Peleg. Desiccation and heat tolerance of Enterobacter sakazakii. Bovill. Enterobacter sakazakii infection in the newborn. Eur. 61:102-108. Arad. Jonsdottir. J. Block. 2001. Infect. Acta Paediatr.. Minster.. 21:613-616. A.. Dis. (Enterobacter sakazakii) in different categories of ready-to-eat foods other than infant formula. and C. Peleg. Beumer. J. J. Enterobacter sakazakii survives spray drying. Karlsson. 2008. N. and O. Measurements and predictions of growth for Listeria monocytogenes and Salmonella during fluctuating temperature II. the scientific insights from the thesis work are put into the perspective of general progress in the field (Chapter 9). In the general discussion. K. 2. A. in reconstituted infant formula (Chapter 7). R. E. Invasive Enterobacter sakazakii disease in infants. O. 9. Bew. C. and H. Rapidly changing temperatures. 67:131-137. and M. Wageningen. G. M. A. P. Fanning. 1989. R. S. I. B.. Joosten. C. Listeria monocytogenes: detection and behavior in food and in the environment. Laboratory of Food Microbiology Wageningen University. Microbiol. Appl. M. Peterz. 12:1185-1189. Shapiro. 2002. the survival and gradual inactivation of Cronobacter spp. 136:189-192. P. References 1. Cluster of neonatal infections in Jerusalem due to unusual biochemical variant of Enterobacter sakazakii.. A. Dis. Block. Grand. R. 7. and K.. Three cases of neonatal meningitis caused by Enterobacter sakazakii in powdered milk. 95:967-973. 6. Clin. Mullane. Infect. Food Microbiol. Liniger. Int. Food Microbiol. Arku. In: PhD thesis. B. in dry powdered infant formula under different storage conditions was investigated in detail and mathematically described (Chapter 8). N.

Fanning. Food Microbiol. J.C. Infect. 22. M... Wall. Thermal inactivation of Enterobacter sakazakii in rehydrated infant formula. and R. and R. 20. S. P. D. 2008. 444:1-14. 11. J. Evaluation of different methods for the detection and identification of Enterobacter sakazakii isolated from South African infant formula milks and the processing environment. L. C. Farber and S. and R. 13:467-472. Whyte. 67:60-63. L. Witthuhn. R. J. and S. Diagn. Edelson-Mammel. J. The EFSA Journal. Murphy. A mathematical approach toward defining and calculating the duration of the lag phase. L. Production of powdered infant formulae and microbiological control measures. and S. Food Prot. and R.M. Boom. D. L. M. Van Schothorst. and M. C.. 17. C. O. M. Buchanan. E. 13. Mullane. S. New Brunswick. 18.. B. 145-185. H. R. Botha. 2007. 2008. Den Aantrekker. Cygnarowicz. 14. Dis. N. M. Principles and guidelines for the conduct of microbiological risk assessment. The EFSA Journal. Cooksey. Cordier. Forsythe.J. CAC.). 110:127-134. Cawthorn. EFSA. 12. Code of hygienic practice for powdered formulae for infants and young children.. 2005. C. Hill. D. New Jersey. Food Microbiol. J. J. Food Microbiol. Quantifying recontamination through factory environments. G. 70:2111-2117. Food Prot. M. O'Hara. J. Forsythe (eds. Buchanan. 16. J. Clark. 113:1-35. Dry stress and survival time of Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered infant formula. Sanjaq. C. S. Buchanan.. Steingrimsson. J. 7:237-240. 2006. O'Rourke. Fischer.. 21. and M. P. M. O'Mahony. CAC. CAC/GL 30-1999. Shannon. 15. Characterization of a collection of Enterobacter sakazakii isolates from environmental and food sources. Microbiol. Kelly. 2004. Int. 1990. G. Zwietering. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial outbreaks. O'Mahony. EFSA. 1999. Drudy. 68:1900-1902. S. R. M. Caubilla-Barron. N. L.. Food Prot. 2007. 80:117-130. CAC/RCP 66-2008. 19. L. Opinion of the scientific panel on biological hazards on the request from the commission for review of the opinion on microbiological risks in infant formulae and follow-on formulae with regard to Enterobacteriaceae as indicators. M. 36 .Chapter 1 10. Enterobacter sakazakii (emerging issues in food safety) ASM Press.. D. Food Microbiol. 2008. Opinion of the scientific panel on biological hazards on the request from the commission related to the microbiological risks in infant formulae and follow-on formulae. Survival of Enterobacter sakazakii in a dehydrated powdered infant formula. P. Washington.a review. Int. R. Edelson-Mammel. 2004. 127:129-138. 1990. 2003. In: J. Int. p.

2002. J. 2007.. 2006.M. and the Enterobacteriaceae study group. L 322/12. Microbiol. D. Kornacki. Hawkins. J. Washington D. Int. J. p. 104:1-34. Italy. Food Microbiol. Clin. Enterobacter sakazakii and Other Micro-organisms in Powdered Infant Formula: meeting report. Forsythe. Microbiological Risk Assessment Series No. Ahmed. Dis. Washington. D. Fanning. Microbiological Risk Assessment Series No. 25.). A. J. Gurtler. Farmer III. Enterobacter sakazakii (emerging issues in food safety) ASM Press. J. New Jersey. M. 30. M. S.J. J. 1991. New Brunswick. 30:569-584.. 2008. R. A.C. FAO/WHO. M. Epidemiology of invasive neonatal Cronobacter (Enterobacter sakazakii) infections. 1441/2007 of 5 December 2007 amending Regulation (EC) No 2073/2005 on microbiological criteria for foodstuffs Official Journal of the European Union. M. W. Italy. Syst. Lissner. Gurtler. Enterobacter sakazakii in melkpoeder. Med.. 2008. Beuchat. 2004. J. M. 1980. 6. Isolation and identification of Enterobacter sakazakii. Powdered infant formula in developing and other countries .J. Heuvelink. J. Enterobacter sakazakii (emerging issues in food safety) ASM Press. E. B. F. 2007. New Jersey.Issues and prospects. Brenner. Eur. 35. Beuchat. F.. South. de Boer. Farber and S.. Enterobacter sakazakii and Salmonella in powdered infant formula: meeting report. and N. 24. J. 26. Rome. Food Prot. C. and S. 27-59.Introduction 23. 27. 34. 28:12971304. Enterobacter sakazakii in food and beverages (other than infant formula and milk powder). In: J. 15 Rome. Int. J. Survival of Enterobacter sakazakii in powdered infant formula as affected by composition. New Brunswick. European Commission (EC). and L. Friedemann. 28. J. and E. and J. Forsythe (eds. 32:1730. Enterobacter sakazakii: A coliform of increased concern to infant health. Friedemann. Bacteriol. De Ware(n)chemicus. R. L. 2005. Microbiological Risk Assessment Series No. J.. Italy. Estuningsih. Sanford. 84:793-795. Enterobacter sakazakii: A new species of "Enterobacteriaceae" isolated from clinical specimens. J. T. D. 37 . J. water activity. 2009.. Int. Commission Regulation (EC) No. Abdullah Sani.C.). S. B. 221-234.) in powdered followup formulae: meeting report. FAO/WHO.. 70:1579-1586. and L. 2008. Infect. 29. In: J. E. 33. Kodde. Enterobacter sakazakii bacteremia in an adult. R. Forsythe (eds. p. 2007. Zwartkruis-Nahuis. Hickman. 10 Rome. 32. Farber and S. Asbury. Enterobacter sakazakii (Cronobacter spp.. FAO/WHO. 116:1-10. R.M. and temperature. 31. P. Food Microbiol.

Dublinensis subsp. Van Schothorst. A. Microbiol. C. and Cronobacter dublinensis subsp.. 42. I. Cronobacter muytjensii sp. B. J. Lehner. C.. E. 44. J. 38.. nov. Trends Food Sci. Reij. A.. Micro-organisms in Foods 7: Microbiological Testing in Food Safety Management. Marugg. 2004.. M. C. Food Sci. N. M. and proposal of Cronobacter sakazakii gen. 2008. Mullane. A new protocol for the detection of 37. nov. Iversen.. Kandhai. 2004. Anderson. Forsythe. Joosten. S. and S. Joosten. and M. BMC Evol. Beumer. nov. lausannensis subsp. ICMSF. M. The taxonomy of Enterobacter sakazakii: proposal of a new genus Cronobacter gen. Lett. Cronobacter gen.. 45. G. and of three subspecies. Forsythe.. 14:443-454. Food Microbiol. L. and A. E. Cronobacter dublinensis sp. C.. Cronobacter muytjensii sp nov. Biol. The Lancet. nov. R. K. 363:39-40. Iversen. Iversen. nov. lactaridi subsp. 43. Risk profile of Enterobacter sakazakii. nov. Tal. and H. Prevalence and thermal stability of Enterobacter sakazakii from unprocessed ready-to-eat agricultural products and powdered infant formulas. D.. 58:1442-1447. McCardel.. Harris. New York. nov. Stephan. K. O. Cronobacter turicensis sp nov. Lehner. 41. 287:2204-2205.. Joosten. Stephan. B. and H... Evol. Cronobacter sakazakii subsp malonaticus subsp nov. Biotechnol. 15:152-157. Gorris. Cronobacter dublinensis sp nov and Cronobacter genomospecies I. The growth profile. C. Forsythe. Isolation of Enterobacter sakazakii and other Enterobacteriaceae from powdered infant formula milk and related products. J. nov. 21:771-777. comb. 38 . Iversen.Chapter 1 36. Fanning. R. . Syst. JAMA. Himelright. and descriptions of Cronobacter sakazakii comb. Int. Microbiol. Cronobacter dublinensis subsp. 17:87-92. N. J. 2004. and J. Technol. Cronobacter malonaticus sp. Van Puyvelde. 2007. C. Guillaume-Gentil.. nov. Mullane. SAJCN. 2004. J. R. and M. W. Appl. Cronobacter dublinensis subsp.. comb.. 40.. 2006. Lorch. Kandhai. Van Schothorst. and M. 2004.. 38:378-382... nov. 7:64. M. H. Springer Publishers. Reij. W. C. M. Enhanced microbiological safety of acidified infant formulas tested in vitro. nov. R. Iversen. nov. O. Occurrence of Enterobacter sakazakii in the food production environments and households. S. Cronobacter turicensis sp. M. Cronobacter genomospecies 1. Park. H. 2002. Bidlas. V. Enterobacter sakazakii infections associated with the use of powdered infant formulaTennessee. 46. Lardeau. M. Lane. 2001. 2002. Cronobacter sakazakii subsp sakazakii. thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant formula milk. M. Guillaume-Gentil. and S. I.. 39. and S. 2003. a new genus to accommodate the biogroups of Enterobacter sakazakii. an emerging pathogen associated with infant formula. Cleenwerck. Jung. nov. Fanning.

80:113-122. A. Appl. Food Microbiol. H. T. ILSI Europe Report Series ILSI. J. L. Medicine. Laboratory of Food Microbiology Wageningen University.). 67:12671270. and X. Liu. N. Enzymatic profiles of Enterobacter sakazakii and related species with special reference to the alpha-glucosidase reaction and reproducibility of the test system. H. X. J. Enterobacter sakazakii infections among neonates. 47. J. X. Cai. M. J. The Netherlands. infants. In: K. Lammerding. 55. J. H. Wageningen. 2007.. 20:684-686. P. 54. P. Microbiol. Syst. and H.Introduction Enterobacter sakazakii applied to environmental samples. Q. B. Minerva Pediatr. Quinn. 56. P. Zhang. Jaspar.. Y. 51. 57. 2007. Muytjens. and G.. G. Z. D. 116:73-81. Application of pulsed-field gel electrophoresis to characterise and trace the prevalence of Enterobacter sakazakii in an infant formula processing facility. and S. Gao. Behaviour and cross-contamination of pathogenic bacteria in household kitchens . A. 58:147-157. Real time PCR using TaqMan and SYBR Green for detection of Enterobacter sakazakii in infant formula. Fazil. J. 2003. 2006. Eberl. Microbiol. 50. C. 59:137-148. Huang. 1988. Wall. P. 39 . Methods. C.relevance to exposure assessment. K.. T. 53. 1997. Food Prot. 2000.. Fanning. Van Druten. In: PhD Thesis. R. L. Rattei. Zheng. 60:1420-1425. 2007. N. 65:21-31. Quinn. Using microbiological risk assessment (MRA) in food safety management. 48.. Whyte.Case reports and a review of the literature. B. L. Frishman. Brussel. A. Yang. H. and R. Microbiol. A. Lehner. Stephan. Lai. K. T. Microbiol. Healy. Mullane. Clin. 52. Fanning. Lammerding. K. and adults . Diep. Yates (ed. 29:609-625. Iversen. Molecular characterization of the alpha-glucosidase activity in Enterobacter sakazakii reveals the presence of a putative gene cluster for palatinose metabolism. P. X. Muytjens. D. A. T. and S. children. Lammerding. Food Microbiol. H. Ruepp. H. J. Kusumaningrum. 2006. Breeuwer. An overview of microbial food safety risk assessment. Luo. Riedel. 1984. Mullane. Wall. M. Clin. Hazard identification and exposure assessment for microbial food safety risk assessment Int. Food Prot. Walsh. 26:743-746. Int. Van der Ros-van de Repe. C. A. and A. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. J. G. J.. N. M. Whyte. M. 2001. Enterobacter sakazakii: an emerging bacterial pathogen with implications for infant health. 49. Roelofs-Willemse. L. J.

204. Genotypic characterization of Enterobacter sakazakii isolates by PFGE. M. Farber. 3rd. Appl. Nazarowec-White. Farber. R. New Jersey. 127-144. 2008. 40 . B. Farber.. 61. Haas. Enterobacter sakazakii infections in neonates associated with intrinsic contamination of a powdered infant formula. J. M. K.C. J. Gurtler. K. L. L. Appl. D. 60. Venkitanarayanan.. p. 1997. I. M. and J. 66. H. 2005. J. and S. J. BOXPCR and sequencing of the fliC gene. Food Prot. Nazarowec-White. Farber. 62. milk. J. S.). Farber. D. Coton. M. Nair. J. and M. Washington DC. and R. 1999.. Med. Infect. Epidemiol. M. A. M. and J. I. Vergnaud. Lett. specific detection of Enterobacter sakazakii in infant formula using a real-time PCR assay. In: J. B. 177 . 24:9-13. 1983.W. J. Incidence. Integrating concepts: a case study using Enterobacter sakazakii in infant formula. 63. Microbiol. Enterobacter sakazakii: infectivity and enterotoxin production in vitro and in vivo. Simmons. J. F.. J. Sonderkamp. 1989. and L. F. Forsythe (eds. Appl. J. S. M. 2008. 1997. survival. Microbiol.. M. J. Thermal resistance of Enterobacter sakazakii in reconstituted dried-infant formula. and R. Lenati. 68. and K. Beuchat. 72:2539-2546. Pagotto. Bougle. M. H. 99:844-850.. 2003. Leclercq. L. K. Enterobacter sakazakii (emerging issues in food safety) ASM Press.Chapter 1 58. Brackett.. H.. New Jersey. Richards. Metts. Zanen. Control Hosp. and J. Wachsmuth. C. 67. 66:370-375. M. and J. Muytjens. 60:226-230. Farber. liquid infant formula. J. Washington. or apple juice. 69. Proudy. p. Zwietering. Environ. 104:26-34. 48:559567. F. A. Clin. In: D. Gelfand. Analysis of eight cases of neonatal meningitis and sepsis due to Enterobacter sakazakii. 64. Microbiol. and M. 10:398-401. Survival and growth of Enterobacter sakazakii in infant rice cereal reconstituted with water. Schaffner (ed. Reij. R. M. and growth of Enterobacter sakazakii in infant formula. Nazarowec-White.. New Brunswick. H. Rapid..J.). P. 2005. 68:59-63. 18:115-120. Microbiol. 59. Cloning and Sequencing of the ompA gene of Enterobacter sakazakii and development of an ompA-targeted PCR for rapid detection of Enterobacter sakazakii in infant formula. H. Food Prot. Phenotypic and genotypic typing of food and clinical isolates of Enterobacter sakazakii. Food Prot. M. Ferguson. J. New Brunswick.. and J. E. Nazarowec White. M. Kollee. Pagotto. Appl. G. Microbiol. Pathogenicity of Enterobacter sakazakii. E. Bidawid. M. and J.M.. M. Microbiol. 2006. W. Farmer. Coton. Microbial Risk Analysis of Foods ASM Press. 65. M. Seo.. 2008.

Muyldermans.int/nutrition/publications/code_english. 73. J. Food Drug and Administration (FDA). W. de Smet. 1961. C. 71. 5:4161-4169. 39:293-297. J. 6:113-115. Enterobacter sakazakii is a rare cause of neonatal septicemia or meningitis in VLBW infants J. 1:313-315. R. M. Musser. F. G. E. Zervos.. Weir. 75. Lemons. Pediatr. Singh. Foley. S. 1981. Fanaroff. R. L. and S. and J. 2009. A. Van Acker. Williams. 144:821-823. N. A top-down proteomics approach for differentiating thermal resistant strains of Enterobacter sakazakii. Powdered infant formula and fatal infecton with Enterobacter sakazakii. International code of marketing of breast-milk substitutes. S.. Proteomics. B.who.Muytjens. 72. A. 2005. F. Med.pdf. Stoll. Neonatal death from pigmented coliform infection. Rev. Smeets. Isolation and enumeration of Enterobacter sakazakii from dehydrated powdered infant formula. Berne. Bougatef.. Accessed June 2009. A. A. and M. Ned. Assoc. Available at: http://www. D. T. Urmenyi. 74. 77. 2002. J. Buchanan. R. and M. Accessed 10 February. Clin.S. S. V. 2004. A.. A.Introduction 70. L. Genetisch karakterisatie van Enterobacter sakazakii isolaten van Nederlandse patiënten met neonatale meningitis.Voss. M. Tijdschr. Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula. A. 78. M. Microbiol.. U. Can. and S. Simjee. J. C.. L. Microbiol. H. Hansen. Edelson-Mammel. Meis. J. J. 76. S. World Health Organisation (WHO). G. 166:1570. 2001. Naessens. The Lancet. Monday. A. Med. 2006. J. Microbiol. Application of molecular techniques to the study of hospital infection. 41 . L. 2002. Clin. 19:512-530. Goering. 1998. Lauwers. G Melchers.

Chapter 1 42 .

43 .Protocol for the detection of Cronobacter spp.

Cronobacter spp. Guillaume-Gentil.R. is a motile. and there is thus a need to investigate whether and where Cronobacter spp. such as scrapings from dust. Using this screening method. have been isolated from milk powder–based formulas. has been used to indicate the taxonomic change. and M. strains were isolated from three individual factories from 18 of 152 environmental samples. which is detected in a 4-h colorimetric assay. necrotizing enterocolitis. K. Reij. 67:6. peritrichous. For this purpose.: its yellow pigmented colonies when grown on tryptone soy agar and its constitutive -glucosidase. 1267–1270. and spilled product near equipment. * Throughout this thesis the designation Cronobacter spp. Van Puyvelde. It is documented as a rare cause of outbreaks and sporadic cases of life-threatening neonatal meningitis.C. 44 . This Chapter has been published as “A new protocol for the detection of Enterobacter sakazakii* applied to environmental samples” Kandhai.. vacuum cleaner bags. The method is useful for routine screening of environmental samples for the presence of Cronobacter spp. Beumer. a simple detection method was developed based on two features of Cronobacter spp.. and sepsis. Journal of Food Protection. Van Schothorst. Cronobacter spp. Gram-negative rod that was previously known as a “yellow pigmented Enterobacter cloacae”. O. R. 2004. occurs in these manufacturing environments. M. M.W.Chapter 2 Abstract Cronobacter spp.

in milk powder is usually performed by enrichment in Enterobacteriaceae enrichment broth. biochemical reactions. is a member of the family Enterobacteriaceae. It is suspected that this micro-organism is present in the environment of the processing equipment (10. Most Cronobacter spp. Until 1980. cloacae in DNA relatedness. this organism was referred to as yellow pigmented Enterobacter cloacae. (7). the specific yellow pigment production. 45 . The isolation of Cronobacter spp. and antibiotic susceptibility (4). in which a variety of related coliforms may be present. 13). It was then reclassified as a unique species based on differences from E. Isolates were commonly obtained by streaking onto blood or chocolate agar. Since heat treatments such as pasteurization readily kill the micro-organism (11). in environmental samples in food production. followed by plating on violet red bile glucose agar and sub-culturing on sheep blood agar and eosin methylene blue agar (8) or on tryptone soy agar (TSA) (12). isolates were screened for the presence of DNase activity on toluidine blue agar after 2 and 7 days of incubation at 36 °C and for the formation of yellow pigmented colonies on TSA at 25 °C for 48 h (4. using the API 20E system for identification.Protocol for the detection of Cronobacter spp. strains described in the literature were isolated from cerebrospinal fluid of the patients (7. An additional characteristic feature is the production of Tween 80–esterase in 3 to 8 days (1). and its application to environmental samples collected from three milk powder production plants. Several outbreaks or sporadic cases of either severe neonatal meningitis in premature infants or necrotizing enterocolitis have been attributed to Cronobacter spp. (3. contaminated dry infant formulas have been identified as the source of Cronobacter spp. In some of these. 9). Introduction Cronobacter spp. This study describes a simple method for the detection of Cronobacter spp. 13). 11). In all. Often. contamination must have occurred after processing. A faster and elective protocol is preferred for the detection of Cronobacter spp. this full confirmation protocol takes more than a week.

streaked onto violet red bile agar (VRBL. Oxoid). The TSA plates were incubated for 48 h in daylight at room temperature. Preliminary identification Isolates were tested for oxidase. On many plates. In case of enrichment.0) phosphate buffer (Merck.. fewer than 10 colonies were present. as the type strain.Chapter 2 Materials and methods Detection A total of 152 dry samples were obtained from the environments of three milk powder production plants. whereas for 35% an enrichment step was used. Germany) in a final 46 . or vacuum cleaner bags. Other samples were streaked directly onto VRBL after homogenizing. France) and the corresponding identification software (API Lab Plus version 3. however. α-Glucosidase activity For the assessment of α-glucosidase activity. The yellow pigmentation on TSA is a characteristic feature of Cronobacter spp. Md.3. without enrichment. Buchs.). corresponding to ATCC 29544. Basingstoke. UK). scrapings. They were collected from floor sweepings. Darmstadt. KgaA.3). as shown in Figure 1. Approximately 65% of the samples were analyzed for the presence of Cronobacter spp. and incubated for 20 to 24 h at 37 °C. From each VRBL plate. 10 g of dry sample was homogenized in 90 ml of buffered peptone water (Oxoid Ltd. 10 colonies typical for coliforms were purified on TSA (Oxoid). using Cronobacter sakazakii DSM 4485.. and in that case all colonies were purified on plates.3 M (pH 7. Oxidase-negative isolates were further identified using the API 20E test system (bioMérieux SA. Hampshire. incubated for 18 to 20 h at 37 °C. Switzerland) was dissolved in distilled water at 50 °C and added to 0. Sparks. paranitrophenyl-α-D-glucopyranoside (Fluka Chemie Gmbh. Marcy l’Etoile. spilled dry products. using Oxidase DrySlide OXIDASE (Becton Dickinson and Company.

Landsmeer. UK) at 405 nm. Rosco.).1 h. cloacae 218 strain that is αglucosidase negative. was used to confirm the identity of presumptive Cronobacter spp. Each colony was suspended in glass tubes (10 by 100 mm. equivalent to 16 µM PNP.85% NaCl solution and a ready-to-use diagnostic tablet (Diatab 50421. Care was taken to transfer a whole colony to the test tube. tubes were incubated in a water bath at 37 °C for 4 ± 0.25 ml of a 0. concentration of 4 g/L of buffer. For some isolates. digitized. caused by the release of PNP.Protocol for the detection of Cronobacter spp. The formation of the yellow colored paranitrophenyl (PNP) hydrolysate was measured after 0. whereupon 2 ml of the paranitrophenyl-α-D-glucopyranoside solution was added.Wilmington. The DNA was then transferred to a membrane and probed with a region of the rRNA operon to reveal the pattern of rRNA genes.85% NaCl (Merck) in distilled water. and stored in a database containing 6 DuPont Cronobacter spp. was considered positive. isolates (2). and 24 h using a Novaspec II spectrophotometer (Pharmacia Biotech. A positive control was also included each time consisting of a suspended tablet with Cronobacter sakazakii DSM 4485. Negative controls were included each time consisting of a suspended tablet without bacteria and another one with an E. A yellow color in the supernatant.3 at 405 nm after 4 h. indicated the presence of α-glucosidase. Del. A minimal absorption of 0. reference fingerprints combined 47 . Taastrup. the RiboPrinter (DuPont Qualicon. The Netherlands) containing 0. the α-glucosidase activity assay was also performed with paranitrophenyl-α-D-glucopyranoside tablets as follows. Denmark). Five yellow. freshly grown colonies on TSA were individually tested for α-glucosidase activity. Individual colonies grown on TSA were suspended in 2 ml of physiological salt solution. Ribotyping An automated ribotyping system. The mixture was then incubated in a water bath at 37 °C. 4. 0. Emergo. After vigorous vortexing for a few seconds. In this system DNA was extracted from a colony and then digested with EcoRI into discrete-sized fragments. The use of a standardized inoculum level of at least number 1 on the McFarland scale is recommended. Cambridge. The pattern was recorded. Positive and negative controls were included as described below.

Chapter 2 with the 2. fingerprints (5). Schematic diagram for the enrichment and identification of Cronobacter spp.000 in-house Cronobacter spp. Figure 1. strains and provides further evidence of the identity of the isolates in addition to the biochemical identification. Comparison of patterns in the database allows for the assessment of relatedness between Cronobacter spp.. 48 .

coliforms were isolated from 33 samples. (9). on the basis of oxidase test and biochemical identification with API 20 E. Coliforms were detected with both procedures. 27 samples were analyzed using both methods. All 32 presumptive Cronobacter spp. From VRBL agar. namely. 152 dry samples were collected. it appears that Cronobacter spp. were tested for their α-glucosidase activity. a pre-enrichment step in buffered peptone water followed by streaking on VRBL agar as described in Figure 1 or a direct plating of diluted samples on VRBL agar. all colonies were streaked on TSA plates and were identified with API 20E. A total of 100 samples were not pre-enriched and yielded in total 30 coliforms. isolates and 18 other coliforms. The API 20E system was not able to identify isolates obtained from 38 samples.. Samples were analyzed with one of the two procedures. As shown in Table 1. Among the Enterobacter species. In another study. numerous Cronobacter spp. strains.. 7 of which were presumptive Cronobacter spp. which had been isolated from environmental samples and had been identified using the API 20E system. can be isolated from environmental samples with and without pre-enrichment.Protocol for the detection of Cronobacter spp. From samples positive for Cronobacter spp. Fifty-two samples were subjected to preenrichment. isolates were obtained in several cases that could be differentiated by their API 20E profiles and their ribotyping fingerprints. the presence of α-glucosidase is known to be specific for Cronobacter spp.. many different species and strains of coliforms were found. and nine of the coliforms were presumptively Cronobacter spp. The combination of the yellow colonies on TSA and positive α-glucosidase activity was therefore considered further evidence for the identity of Cronobacter spp. strains were isolated from 13 of 27 samples without enrichment and from 16 of the same 27 samples with enrichment (data not published). Some of these coliforms were subsequently identified as presumptive Cronobacter spp. Presumptive Cronobacter spp. Results and discussion From the line environment of three milk powder plants. 49 . From the data presented here.

Rahnella aquatilis Serratia ficaria Unidentified a Factory A 56 2 1 Factory B 49 Factory C 47 3 1 5 8 (16) 1 c 8 4 (6) 2 1 1 2 1 2 1 5 3 10 c 9 6 (10)c 2 1 3 b 1 1 1 6 1 2 21 1 2 1 7 Identification was based on oxidase test and API 20E. None of the other coliforms were positive in the 4-h assay. 6. One Pantoea isolate and one strain of Erwinia showed substantial α-glucosidase activity to be rated positive in the 24-h assay. Presumptive identification of coliforms isolated from environmental samples taken in three milk powder factoriesa Species No. The Table indicate the number of samples from which a certain species was isolated. and 10 different isolates were obtained from the samples based on ribotyping results. 29 Cronobacter spp. c Implies that 16. isolates hydrolyzed paranitrophenyl-α-Dglucopyranoside within 4 h. the difference between Cronobacter spp. Leclercia adecarboxylata Pantoea spp. As shown in Table 2. Erwinia nigrifluens Escherichia coli Escherichia hermanii Escherichia vulneris Klebsiella ornithinolytica Klebsiella oxytoca Klebsiella pneumoniae Kluyvera spp. by the API 20E system did not show α-glucosidase activity and were not confirmed to be Cronobacter spp. b Please note that 3 of the 32 presumptive isolates that were identified as Cronobacter spp.Chapter 2 Table 1. After 24 h. This indicates that the specificity of the α-glucosidase 50 . and the other coliforms was less clear. by their ribotype fingerprint. of samples Chryseomonas luteola Citrobacter diversus Enterobacter amnigenus Enterobacter cloacae Cronobacter spp.

isolates were α-glucosidase negative after 4 h. Presumptively identified as Cronobacter spp. it can be 51 . Using a database of reference fingerprints. The presumptive Cronobacter spp.Protocol for the detection of Cronobacter spp. Yellow-pigment formation and the -glucosidase assay after 4 and 24 h incubation by strains grown on TSA Alpha-glucosidasea Presumptive identification (no. by the API 20E system with a low percentage of match (80%) with the expected biochemical profile of Cronobacter spp. ribotyping or by α-glucosidase presence. only applies for a short-term assay. although they were identified as Cronobacter spp. Ribotyping is a method that generates a highly reproducible and precise fingerprint of bacterial rRNA genes. Table 2. (1) Pantoea spp. (3)b Enterobacter cloacae (4) Erwinia nigrifluens (1) Erwinia nigrifluens (1) Escherichia coli (1) Escherichia hermanii (1) Escherichia vulneris (1) Klebsiella ornithinolytica (1) Klebsiella oxytoca (1) Klebsiella pneumoniae (1) Pantoea spp. strains were also characterized using the ribotyping technique. Three presumptive Cronobacter spp.. No false-positive results were observed with the 4-h α-glucosidase assay. (1) Rahnella aquatilis (1) a b Yellow-pigment on TSA + + + + + + + + + + 4h + - 24 h + + + - + : paranitrophenol concentration (PNP) greater than 16 µM. These three isolates were clearly α-glucosidase negative after retesting. (29) Cronobacter spp. as measured by OD at 405 nm. activity of Cronobacter spp. by the API 20E system but not confirmed by. of strains) Citrobacter diversus (3) Cronobacter spp.

This method is useful for routine screening of environmental samples taken from milk powder factories and most likely also for their dry milk powder–based formulas (6). and also investigate relatedness among Cronobacter spp. were not Cronobacter spp. strains of the Riboprinter database. dr. next to yellow pigmentation. by ribotyping. Matz Peterz for critically reading and reviewing the manuscript. It should be noted that at least a 90% of biochemical reaction of the API 20E is needed for a reliable identification. it can be concluded that the three isolates identified by the API 20E system as Cronobacter spp. isolates. and the α-glucosidase assay. it can be concluded that the combination of yellow pigmented colonies on TSA and the presence of α-glucosidase activity detected in a 4-h assay are a good method to differentiate Cronobacter spp.. oxidase/API 20E. from other coliforms. Acknowledgments Nadine Braendlin is kindly acknowledged for her expert advice and her help in obtaining samples. The authors thank Wilma Hazeleger. to confirm the identity of Cronobacter spp. The results showed that 27 presumptive Cronobacter spp. work is currently ongoing to optimize the enrichment step. Based on the absence of α-glucosidase and the ribotyping results. The remaining three strains were the three α-glucosidase-negative isolates and indeed had fingerprints that were significantly different from the six reference Cronobacter spp. and Dr. In summary. Prof. Leon Gorris. To ensure the detection of low numbers of Cronobacter spp. Two other isolates gave unclear fingerprints due most likely to the lack of DNA digestion by the EcoRI enzyme. isolates were confirmed to be Cronobacter spp. in environmental samples. The technology can be used here as a further tool. 52 .Chapter 2 used to classify and identify bacteria. The results of the ribotyping also indicated that in none of the three factories was a specific strain predominant (results not shown).

Farmer III. 3. and M. 1990. Microbiol. 2. Automated system rapidly identifies and characterizes microorganisms in food. O. 80:113-122. 2002. 1980. J. J. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial outbreaks. References 1. Zannoni. Joosten. Brenner. and R. A. and G. Salvadori. B. P.. Clin. 13:467-472. 15:508-512. H. J. Accessed 10 February. Enzymatic profiles of Enterobacter sakazakii and related species with special reference to the alpha-glucosidase reaction and reproducibility of the test system. children. 11. and R. Microbiol. Bruce. De Boer. A. Lai. Hill. Enterobacter sakazakii: A new species of "Enterobacteriaceae" isolated from clinical specimens. and J. K. Cooksey. Microbiol. Scheldeman. Hausner. M. Hickman. Appl. Mascini. Farber. T. Diagn. H. and the Enterobacteriaceae study group. Postupa. Bacteriol. E. 26:743-746. F. Syst.. Food Technol. U. M. Zbl. H. Muytjens. C. A. 1997. J. P. 5. Ahmed. J. C. Van der Ros-Van de Repe. 2002. Microbiol. Infect. Clin. J. and H. N. C. 20:684-686. Isolation and enumeration of Enterobacter sakazakii from dehydrated powdered infant formula. Technol. Detection of bacterial contamination in sterile UHT milk using an L-lactate biosensor. Steingrimsson. O. 10. A 256:103-108.. Environ.Case reports and a review of the literature. Enterobacter sakazakii in melkpoeder. De Ware(n)chemicus. 50:77-81. Clark. 1983. J. 7. H.S. J. and G. Marugg. 1984. D. 1988. O.. J. C. 1993. Tween-Esterase activity in Enterobacter sakazakii.. Nazarowec-White. Aldova. Guillaume-Gentil. C. K. 12. 32:1730. Dis. Muytjens. Enzyme Microb. and E.. Jaspar. 24:9-13. M. 53 . Bakt. Int. Genetic heterogeneity in Bacilllus sporothermodurans as demonstrated by ribotyping and repetitive extragenic palindromic-PCR fingerprinting. J. Microbiol. Food and Drug Administration. Hyg.. 6. Roelofs-Willemse. 1996. 8. Asbury. Heyndrickx. M. M. 2001. L. and adults . 30:569-584. Lett. M. 68:4216-4224. Skladal. J. Enterobacter sakazakii infections among neonates. Thermal resistance of Enterobacter sakazakii in reconstituted dried-infant formula. 9. 2009. infants. Herman. L. O'Hara.. E. W. 4.Protocol for the detection of Cronobacter spp. L. Heuvelink. F. Zwartkruis-Nahuis. Appl. 2002.. M. Van Druten. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. Medicine. D. H. Kodde.

J. Naessens. G. Van Acker. and S.Chapter 2 13. Lauwers. A. J. Microbiol. Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula.. 39:293-297. F. Muyldermans. Clin. 2001. 54 . Bougatef. de Smet. A.

Enrichment procedure for Cronobacter spp. 55 .

. all belonging to the Enterobacteriaceae. isolation on violet red bile glucose agar. *Throughout this thesis the designation Cronobacter spp. The method is based on selective enrichment at 45 ± 0. were suppressed. When exposed to light during incubation at 37 °C. Crononacter spp. whereas the reference procedure (enrichment in buffered peptone water. 68:1. 56 . 2005. in environmental samples. Journal of Food Protection.5 °C.Chapter 3 Abstract A method was developed to detect and identify Cronobacter spp. Using this new protocol. J. 64–69. V. whereas 35 of 39 strains of potential competitors. Cronobacter spp.D. has been used to indicate the taxonomic change.5 °C in lauryl sulfate tryptose broth supplemented with 0. produces yellow colonies within 24 h. All of the Cronobacter spp. was isolated from almost 40% of the samples. strains tested (n=99) were able to grow in mLST at 45 ± 0. A survey was carried out with 192 environmental samples from four different milk powder factories. Sonnard. and H. This Chapter has been published as “A simple and rapid cultural method for detection of Enterobacter sakazakii* in environmental samples” Guillaume-Gentil. in environmental samples. identification was confirmed by testing for a-glucosidase activity and by using API 20E strips.C. M.5 M NaCl and 10 mg/L vancomycin (mLST) for 22 to 24 h followed by streaking on tryptone soy agar with bile salts. Kandhai. Joosten. and biochemical identification of randomly chosen colonies) only yielded 26% positive results. O. Marugg. This selective method can be very useful for the rapid and reliable detection of Cronobacter spp.

g. 15. and may thus be prone to false-negative results in the presence of other Enterobacteriaceae. Introduction Cronobacter spp. 25–27). 26). We also introduce a new procedure that produces yellow colonies on the isolation agar within 24 h by exposing the plates to visible light during incubation at 37 ºC. and isolation on violet red bile glucose agar). This approach is laborious and requires up to 7 days obtaining a confirmed positive result. this procedure is not selective for Cronobacter spp. we describe the development of a selective enrichment procedure for Cronobacter spp. and necrotizing enterocolitis. 3. Recently. a yellow member of the family Enterobacteriaceae. based on lauryl sulphate tryptose. In some of these cases. in the factory environment. followed by biochemical identification of typical colonies (1. sub-culturing in Enterobacteriaceae enrichment broth. was indeed isolated (21. 22). 24). pre-enrichment in buffered peptone water. 19. is often a decisive factor in causing the outbreaks (4. bactereamia. 9. which suggests that multiplication of Cronobacter spp. a commonly used enrichment broth for coliforms. 20. contaminated powdered milk infant formula was identified as the source of Cronobacter spp. However. The only methods described thus far are based on general methods for the enrichment and isolation of Enterobacteriaceae (e. confirmed positive results can be obtained after 5 days.Enrichment procedure for Cronobacter spp. because the presence of Cronobacter spp. 21.. 23. where it appears to survive better than other members of the Enterobacteriaceae (5). 25. 4. in infant formula is often attributed to contamination originating from dry factory environments. several cases have been associated with an exposure of prepared infant formula to improper holding temperatures. To evaluate the effect of different cleaning strategies on the prevalence of Cronobacter spp. 26). 57 .5 °C). Here. The dose-response relationship for this opportunistic pathogen is not known. has been associated with various cases of severe neonatal meningitis. (2. albeit at very low concentrations of less than 1 CFU/g.. sensitive and specific detection methods are required. With this method. Management of factory environment hygiene is critical to the control of the organism. Cronobacter spp.5 M NaCl and by the application of a relatively high incubation temperature (45 ± 0. which is based on the detection of glucosidase and production of a yellow pigment. In surveys of dry infant formulas obtained from various countries. Nevertheless. an improved identification method was proposed (13). particularly in premature babies (2–4. 8. However. it remains very important to reduce the prevalence of the micro-organism in infant formula. Selectivity was enhanced by the addition of 10 mg/L vancomycin and 0.

(13).. Inc. Wilmington. or were kindly provided by Dr.. S.. by API 20E and automated ribotyping (data not shown). Del. Oxoid Ltd.. strains were selected to represent the major species of other Enterobacteriaceae found in factory environments. Nijmegen.000 in-house Cronobacter spp. Marcy l’Etoile. 58 . UK) and by Dr. These strains were selected to represent the major ribotypes of this species. All strains were maintained at -30 ºC in bacterial strain storage medium (Bio-Rad Laboratories. The identity of Cronobacter spp. 70 of which had been isolated from the environment of milk powder factories in Europe and Asia using a non-selective enrichment and isolation method. Pantoea agglomerans. Forsythe (Nottingham Trent University. The strains either came from the Nestlé Food Safety Microbiology culture collection or had been isolated from milk powder or from the environment of milk powder factories. digitized.Chapter 3 Materials and methods Bacterial strains Ninety-nine strains of Cronobacter spp.. L. Hercules. Escherichia vulneris. Hampshire. strain was recorded. France) and by ribotyping with the RiboPrinter (DuPont Qualicon.). The identification of the Cronobacter spp. The remaining strains were isolated from private households (14). Calif. was further confirmed with API 20E strips (bioMe´rieux SA. The Netherlands). Nottingham. Escherichia hermanii. Thirty-nine non–Cronobacter spp. Muytjens (Radboud University.. and stored in a database containing 11 DuPont reference Cronobacter spp. and Serratia ficaria) that can easily be confused with yellow Cronobacter spp. fingerprints combined with 2. Pantoea spp. H. fingerprints (unpublished results). All strains were different from Cronobacter spp.) (6). Basingstoke. Inc. were used in this study (Table 1).. Working cultures were prepared by inoculation of brain heart infusion broth (BHI. UK) and incubation for 20 h at 30 ºC. strains (Citrobacter spp. strains was based upon the combination of yellow pigmentation and a positive -glucosidase reaction. The ribopattern of each presumptive Cronobacter spp. The selection focused on yellow non– Cronobacter spp. as described by Kandhai et al.

158. 441. 328.100. 284. 460. 151. 167. 298. Pantoea spp. 288. “++” OD620 < 0. 369 310. Cronobacter spp. 108 64 104 254 255 80 143 188 246 168 415 416 440. 457. 266. Table 1. Cronobacter spp. 166. 318. 145. MP: isolates from milk powder. c LST (Lauryl Sulphate Tryptose broth) supplemented with 0. 156. 295. 297. 281. 429. 294. 390. 306. 383. 303. Citrobacter spp. 459. 276. Klebsiella oxytoca Klebsiella pneumoniae Klebsiella pneumoniae Leclercia Pantoea agglomerans Pantoea agglomerans Pantoea agglomerans Pantoea spp. 321. 316. “-“: OD620 < 0.010. 271. 461. 262. 376e (=ATCC 29544). 468. 164. d Radbout University. Cronobacter spp.Enrichment procedure for Cronobacter spp. 308. and other Enterobacteriaceae strains used in this study Species Strain code Sourcea Growth inb LST LST mLST 30 ºC 45 ºC 45 ºC Cronobacter spp. 450 65 142 427 E Yellow pigment on TSBA + ++ ++ ++ Cronobacter spp. 378 454. 319. Cronobacter spp. Cronobacter spp. 279. 377. 443. 428. Citrobacter spp. 393 313. Hafnia spp. 307. 323. Citrobacter spp. 458. 359. 48. 293. “+” : OD620 = 0. 283. 292. 366. Serratia ficaria Serratia ficaria E MPd MP Cd Cd H Not knownf Not knownf NFSMg NFSM MP NFSM NFSM MP MP MP NFSM NFSM NFSM NFSM MP E NFSM NFSM NFSM NFSM NFSM NFSM NFSM NFSM NFSM MP MP MP E NFSM NFSM E ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + ++ + ++ + ++ ++ ++ + ++ + ++ + ++ ++ + ++ ++ ++ + ++ ++ + ++ ++ ++ ++ ++ + ++ ++ ++ + ++ + ++ ++ ++ ++ ++ - + + + + + + + + + + + + + + + + + + + a E: Isolates from the environment of milk powder factories. 282. 291.010. 275. 356. 301. 299. 472 473 375. Enterobacter cloacae Enterobacter cloacae Enterobacter cloacae Enterobacter cloacae E aerogenes Escherchia coli Escherchia hermanii Escherchia hermanii Escherchia hermanii Escherchia vulneris Escherchia vulneris Hafnia. 471. 315. 278. 265. 469. 435. 384. 263. 305. 304. 462. 368. 300. Cronobacter spp. 302. 455.spp. 62 84 86 418 149 186 40 73 75 417 422. 317. 274. 327. Nijmegen (The Netherlands). 314. Pantoea spp. 442 42. e Type strain f Nottingham Trent University (UK). 463 456 31 189 411 20. 320. 285. Cronobacter spp. 290. 322. Providencia spp. 270.010 – 0. 277. 354. 49. g NFSM: Nestlé Food Safety Microbiology Culture Collection. 59 .5 M NaCL and 10 mg/L vancomycin (mLST). C: clinical isolates. 155. 287. H: households. 391 365 309. 312. 272. 280. 273. 286. 470. b Growth was assessed by measuring the optical density at 620 nm after incubation. Cronobacter spp. Pantoea spp.

5 M NaCl and 10 mg/L vancomycin (Fluka Chemie. and 48 h using a microtiter plate reader (Multiskan/MCC 340.100. 6. 46. strains (Table 2).25 M NaCL LST supplemented with 0. 0. b Nestlé Food Safety Microbiology (NFSM) 262.J. N. 45. Oxoid) supplemented with different concentrations (0. 0.5. 24.75. 328. The microtiter plates were incubated at 44.1 ml aliquots of freshly grown BHI cultures diluted 1:100 in physiological saline were inoculated into 10 ml volumes of standard LST and mLST to give an initial count of ca. 60 . 359. Based on the results of these experiments. and 47 °C.50 M NaCL LST supplemented with 0. strainsb Standard LST LST supplemented with 0. and 1.5 °C was selected for further evaluation with 99 Cronobacter spp. “-”: OD620 < 0. 0.25.). 369. 263. 391. Buchs. 384. 310. Helsinki. Growth of 15 Cronobacter strains in LST supplemented with various NaCl-concentrations Temperature 37 ºC 44 ºC 45 ºC 46 ºC 47 ºC Growtha of 15 Cronobacter spp. Table 2. “++”:OD620 > 0. 266.0.Chapter 3 Development of a selective enrichment broth for Cronobacter spp. The conditions for selective enrichment culture of Cronobacter spp. 383.00 M NaCL a ++ ++ ++ ++ + ++ ++ ++ ++ - ++ ++ ++ + - ++ ++ + - + + - Growth was assessed by measuring the optical density at 620 nm after 24 h incubation. 280. 0. 393.010. Whitehouse Station. 104 CFU/ml and then incubated at 30 or 45 °C for 24 h. Growth was assessed by determining the optical density at 620 nm after 0. 302.0 M) of NaCl (Merck. “+” : OD620 0. For this part of the study. Finland).010.010 . LST broth supplemented with 0. 270. Aliquots (5 µl) of freshly grown BHI cultures were inoculated into wells containing 200 µl lauryl sulfate tryptose broth (LST.75 M NaCL LST supplemented with 1. 362. Labsystems. strains and 39 strains of other Enterobacteriaceae. The optical density was then measured at 620 nm. were established in microtiter plate growth assays using 15 Cronobacter spp. Switzerland) (mLST) and incubated at 45 ± 0.

A loopful of pre-enrichment broth (ca. Oxoid) and examined after 24 h of incubation at 37 °C for yellow pigment production. spilled dry products.. scrapings of processing areas. the volume of BPW was adapted proportionally (1:10). Non-selective enrichment and isolation For non-selective enrichment (BPW-VRBL) samples (10 g each) were suspended in 90 ml of buffered peptone water (BPW. moist environmental sponge samples (8 by 4 by 1 cm) were obtained. 100 ml of BPW. 61 . Sponges were soaked with ca. The enrichment broth was incubated for 16 to 20 h at 37 °C. oxidase.10 µl) was transferred to 10 ml of mLST and incubated for 22 to 24 h at 45 ± 0.Enrichment procedure for Cronobacter spp. 600 lx) to induce production of a yellow pigment by Cronobacter spp. When quantities smaller than 10 g were analyzed. From each plate. contents of vacuum cleaner bags. From the same factories. A loopful of pre-enrichment broth (ca. Oxoid). Oxoid) and incubated at 37 °C for 24 h.5 g/L bile salts 3 (TSBA. From each plate. 10 µl) was streaked onto violet red bile lactose agar (VRBL. (13) and identified with API 20E strips. five colonies demonstrating typical coliform characteristics were transferred to tryptone soy agar (TSA. During incubation (24 h at 37 ºC). samples were first preenriched in BPW as described above.5 °C followed by streaking of a loopful onto a 14-cm petri dish containing TSA supplemented with 1. plates were exposed to artificial white light (ca.. and α-glucosidase activity as described by Kandhai et al. Oxoid). five yellow colonies were selected for confirmation as described above. etc. Two-stage selective enrichment For the two-stage selective enrichment (BPW-mLST-TSBA). Samples Dry environmental samples were obtained from different areas of four milk powder factories and included floor sweepings.

Schematic diagram for the enrichment and identification of Cronobacter spp. 10 g of sample was suspended in 90 ml of mLST and incubated for 22 to 24 h at 45 ± 0. BPW-mLST-TSBA and BPW-VRBL.Chapter 3 Figure 1.5 °C. When quantities smaller than 62 . in environmental samples with mLST-TSBA. Single-stage selective enrichment For the single-stage selective enrichment (mLST-TSBA).

All 15 Cronobacter spp.5 °C and that the addition of NaCl was essential for selective enrichment of Cronobacter spp. However. the volume of mLST was adapted proportionally (1:10). The NaCl was added because Breeuwer et al. in particular Bacillus spp. whereas 35 out of the 39 non–Cronobacter spp. is able to grow at temperatures up to 46 or 47 °C. 63 . Preliminary tests with LST + 0. After incubation. 24) indicating that Cronobacter spp.Enrichment procedure for Cronobacter spp. Klebsiella oxytoca. 12)... was modified by adding extra NaCl and the antibiotic vancomycin. The extended growth study with pure cultures (Table 1) revealed that many of the selected strains of Enterobacteriaceae were good competitors in LST at 45 ± 0. The remaining competitors belonged to a variety of species: Escherichia hermanii. 17). The vancomycin was added as an additional safeguard against the growth of competing gram-positive micro-organisms. Enterobacteriaceae strains were inhibited. Sponges were soaked with ca. and lactic acid bacteria (10. LST broth. Enterobacter cloacae. none of these strains produces a yellow pigment and could be easily distinguished from Cronobacter spp.5 M NaCl and an incubation temperature of 45 ± 0. 16. A relatively high incubation temperature was chosen on the basis of results of different studies (5.5 °C was selected for further evaluation.. widely used as semi-selective enrichment broth for coliforms (11. Results and discussion To enhance the selectivity of Cronobacter spp. has a higher tolerance for osmotic stress than do other members of the Enterobacteriaceae such as Salmonella spp. 7. on the isolation medium TSBA. 100 ml of mLST. (5) demonstrated that Cronobacter spp. Escherichia coli. 10 g were analyzed. strains evaluated grew well in standard LST up to 46 °C and were able to grow in LST at 37 °C at increasing osmolarities up to 1 M NaCl (Table 2). Klebsiella pneumoniae and Citrobacter freundii. suspensions were transferred to TSBA for evaluation. strains were able to grow. all of the tested Cronobacter spp. and Klebsiella pneumoniae.5 M NaCl had indicated that these gram-positive bacteria were often found in environmental samples and were apparently not sufficiently inhibited by the presence of lauryl sulfate (results not shown). The combination of LST with 0.

Similar numbers of Cronobacter spp. in the samples. Table 3. positive samples were obtained with BPW-mLSTTSBA and mLST-TSBA (73 and 75.Chapter 3 The performance of mLST broth was assessed in a comparative study involving 192 environmental samples collected from four milk powder factories. This difference is likely due to an uneven distribution of Cronobacter spp. Thus. Both the one and twostage enrichment procedures (mLST-TSBA and BPW-mLST-TSBA. Pantoea agglomerans. Nevertheless. not all the yellow colonies growing on TSBA could be confirmed as Cronobacter spp..5) The total number of positive samples found in this comparative study was still considerably higher (88) than that obtained with either of the two mLST-based methods (Table 3). whereas the reference method yielded only 51 positive results (Table 3). respectively) were used. 64 . in environmental samples from four milk powder factories. (%) samples Confirmed positive with: mLSTBPW-mLSTTSBA TSBA 18 18 9 10 23 19 25 26 75 (39) 73 (38) BPW-VRBL 9 5 19 18 51 (26.. additional tests such α-glucosidase activity and identification with API 20E strips are necessary to eliminate false-positive results. Detection of Cronobacter spp. which had been divided to allow comparison of the different procedures as outlined in Figure 1. using three different protocols Source Factory A Factory B Factory C Factory D Total Tested 48 46 44 54 192 (100) Confirmed positive 20 10 24 34 88 (46) No. respectively). or Escherichia vulneris. Some strains were identified as Pantoea spp. The difference in efficiency between the mLST-based procedures and the reference procedure is probably due to both the selectivity of mLST and the utilization of TSBA as isolation agar rather than randomly picking coliform colonies from VRBL plates. and the non-selective enrichment method (BPW-VRBL) (14) was used as a reference.

7. we also used an accelerated procedure to obtain yellow colonies on the isolation agar medium (TSBA) by exposing the plates to artificial white light during their incubation at 37 °C. Prior to implementing this method. E. Bidlas for determination of ribopatterns and Dr. The direct mLST-TSBA method has been used in the development of a Cronobacter spp. the latter method may be preferred for routine purposes because it can be combined with Salmonella testing. Acknowledgments The authors thank Mrs. 65 . Jackson for his critical review of the manuscript. This modification represents a clear advantage to the methods used previously. specific PCR assay on the BAX system (Dupont Qualicon) that provides results within 30 to 32 h (18). T. Such information is essential for developing control and prevention strategies for this emerging opportunistic pathogen. In this study. Effective environmental analysis can contribute to a better understanding of the route(s) by which Cronobacter spp. from environmental samples. 21). and in all cases the presence of yellow colonies could already be observed within 24 h (data not shown).Enrichment procedure for Cronobacter spp. However. Although the direct mLST-TSBA method is 1 day faster than the BPW-mLST-TSBA route. enters factories producing powdered infant formulas and will help managers identify in-factory ecological niches where this organism can persist. strains from our collection. the first stage of this method also consists of a non-selective preenrichment. which are based on incubation at 25 °C for 48 to 72 h (3. we tested more than 100 different Cronobacter spp. both methods may be prone to false-negative results. and further improvements of the sensitivity may be necessary. The procedures outlined in this study will enable efficient and rapid isolation of Cronobacter spp.

Asbury. F. Anonymous. Susceptibility to vancomycin of Enterococci isolated from dairy products. Pediatr. S. O. Switzerland. 11.most probable number technique. Geneva.gov/~comm/mmesakaz. Dessication and heat tolerance of Enterobacter sakazakii. J. Jonsdottir. C. Clin.. 12. J. Clark. J. 21:135-136. Hickman. M. 2. 2004a. and R. 3. Lardeau.. Food and Agricultural Organization of the United Nations. 1996. Wkly Rep. Enterobacter sakazakii infections associated with the use of powdered infant formula Tennessee. and H. 10. M.fda. Ball. E. Enterobacter sakazakii: a new species of "Enterobacteriaceae" isolated from clinical specimens. 95:967-973. Mortal. Van Schothorst. Appl. Bruce. 6. R. 67: 12671270. Peterz. Switzerland. Gallagher. D. CX/FH 03/13. 2002.cfsan. Three cases of neonatal meningitis caused by Enterobacter sakazakii in powdered milk. M. 5. Ludvigsson. W. S. Isolation and enumeration of Enterobacter sakazakii from dehydrated powdered infant formula. P. Steingrimsson. ISO 4831:1991: Microbiology . 2003.html. Bakt. Appl. 10 March 2003.General guidance for the enumeration of coliforms . Hyg. 13. and O.. 1989. Sisto. Kandhai. A new protocol for the detection of Enterobacter sakazakii applied to environmental samples. G.. Farmer. 1983. Radiol. 51 (14):297-300. Karlsson. 1980. O. 9. 4. International Organization for Standardization. 2002. Postuva. Italy. IDF 73B:1998: Enumeration of coliforms colony count technique and most probable number technique at 30 °C.. L. 7. A. Breeuwer. International Dairy Federation. Giraffa. 25:335-338. P. Rome. Int. Hausner. Van Puyvelde. Tween-esterase activity in Enterobacter sakazakii. Microbiol. K. Available at: http://www. Codex Alimentarius: risk profile of Enterobacter sakazakii in powdered infant formula. Geneva. 2001. W. M. International Organization for Standardization. J. Joosten. Food Prot. J. 2003.Chapter 3 References 1. 1991. K. Zbl. J. 27:2054-2056. 66 . Aldova. Center for Disease Control and Prevention. Beumer. C. J. N. Automated system rapidly identifies and characterizes microorganisms in food. 50:77-81. 1997. Brenner and the Enterobacteriaceae study group. International Dairy Federation 1998. Guillaume-Gentil. and F. Cerebral infarctions due to CNS infection with Enterobacter sakazakii. Food and Agriculture Organization. 30:569-584. Microbiol. Reij. Syst. G. Food Technol. United States Food and Drug Administration.. J. and W. Microbiol. 1991.. Biering. A. P. R. A256: 103-108. G. Anonymous. 8. Bacteriol. E. MMWR Morb. and M.

and P. Kollee. A. Sonderkamp. Lai. Medicine 80:113-122. 1983.. Analysis of eight cases of neonatal meningitis and sepsis due to Enterobacter sakazakii. B. F. 39:293-297... J. Manie. Nazarowec-White. 22. Reij. Guillaume-Gentil. Tift.Enrichment procedure for Cronobacter spp. 67 . Infect. 26. Microbiol. Simmons. T. 1979. Enterobacter sakazakii infections among neonates. L. Marugg. Comparison of fecal coliform agar and violet red bile lactose agar for fecal coliform enumeration in foods. M. Muytjens. 16. P. A.. 7:196-199. V. sakazakii in milk powders and environmental samples (manuscript in preparation). W.J. and H. and J. H. 2001. Clin. Gorris. Van Acker. Antimicrobial resistance of bacteria isolated from slaughtered and retail chickens in South Africa. L. Microbiol.L. and L. Microbiol. W. C. 1997. Microbiol. 19. 9:372-373. J. L. J. Barbour. O. Lauwers. L. Leclercq. J. S. and M. J. Haas. 1988. Appl. 2001. 68: 1631-1638. M. H. 60:226-230. Control Hosp. Farmer 3rd. Kollee. S. Enterobacter sakazakii meningitis in neonates: causative role of formula? Pediatr.L. M. J. Food Prot.. O. Jackson. 1998. Bacteremia associated with Enterobacter sakazakii (yellow-pigmented Enterobacter cloacae). Environ. K. S. G. 20:684-686. H. Kandhai. A. 10:850851. J. P. Mrozinski. C. Clin. Gouws. J. Roelofs-Willemse. 10:398-401. M. W. Development of a PCR based method for the detection of E. J. J.. and J. H. Naessens. Epidemiol. Infect. M. Muyldermans.. M. Clin. H.. 14.. J. M. Clin Microbiol. Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula. Veith. H. L. Incidence. T. J. Monroe. H. and S. 27. Khan.. Enzymatic profiles of Enterobacter sakazakii and related species with special reference to the alpha-glucosidase reaction and reproducibility of the test system. Dis.. Occurrence of Enterobacter sakazakii in food production environments and households. infants. and adults. Muytjens. K. de Smet. 1989. Infect. Brözel. Enterobacter sakazakii meningitis in neonates. 23. K. 1984. Van der Ros-Van de Ripe. I. Van Druten. Wanegue. Zanen. Wachsmuth. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. M. and P. Muytjens. E. 15. and P. Ferguson. A. 2002. Willis. H. Jaspar. Van Schothorst. J. Pediatr. Farber. G. Bougatef. and growth of Enterobacter sakazakii in infant formula. M. Microbiol. Joosten. 21. 26: 253-258. children. 1988. survival. 26: 743-746. 17. H. A. Dis. Muytjens. Guillaume-Gentil. L. Clin. The Lancet 363:39-40. and W. 18. and J. J. A. 1990. Robinson. and J. 20. 2004b. and G. A. 24. Metts. Baylac. Enterobacter sakazakii infections in neonates associated with intrinsic contamination of a powdered infant formula. Appl. 18: 115-120. Gelfand. L.. Microbiol. 25. C..

Chapter 3 68 .

Occurrence of Cronobacter spp.

69

Chapter 4

Abstract
Cronobacter spp. occasionally causes illness in premature babies and neonates. Contamination of infant formulae during factory production or bottle preparation is implicated. Advice to health-care professionals focuses on bottle preparation, but the effectiveness of prevention depends on the degree of contamination and contamination sites, which are generally unknown. To keep contamination to a minimum in the finished product depends on knowledge of the occurrence of Cronobacter spp.. We used a refined isolation and detection method to investigate the presence of this micro-organism in various food factories and households. Environmental samples from eight of nine food factories and from five of 16 households contained Cronobacter spp.. The widespread nature of this micro-organism needs to be taken into account when designing preventive control measures.

This Chapter has been published as “Occurrence of Enterobacter sakazakii* in food production environments and households” Kandhai, M.C., M.W. Reij, L.G.M. Gorris, O. Guillaume-Gentil, and M. Van Schothorst. The Lancet, 2004; 363:1, 39–40. * Throughout this thesis the designation Cronobacter spp. has been used to indicate the taxonomic change.

70

Occurrence of Cronobacter spp.

Introduction
Several species within the genus Enterobacter have been recognised as important causative agents of hospital acquired infections. Cronobacter spp. in particular has caused several outbreaks or sporadic cases of severe neonatal meningitis in premature babies, or necrotising enterocolitis. Cronobacter spp. has been isolated from various clinical materials, such as cerebrospinal fluid, blood, skin, wounds, respiratory tract (sputum, throat, and nose), digestive tract, and urine (1). Cronobacter spp. was isolated from powdered substitutes for breast milk (4). Infant formulae are pasteurised during manufacturing, and Cronobacter spp. does not survive such heat treatment (5). The microorganism’s presence in the finished product, therefore, probably originates from the factory environment, from heat-sensitive micronutrients added after pasteurisation or from bottle preparation. In several investigations into outbreaks or cases of Cronobacter spp. infection in premature babies and neonates, the micro-organism was isolated from blenders, bottle cleaning brushes, etc. (3). Muytjens and Kollee investigated the occurrence of Cronobacter spp. more widely, but could not isolate the micro-organism from any environment they examined, which included surface water, soil, mud, rotting wood, grain, bird dung, rodents, domestic environments, cattle, and untreated cow’s milk (3). Cronobacter spp. has, however, been isolated from cheese, minced beef, sausage meat, and vegetables (2), but we are unaware of any investigation into the natural habitat of Cronobacter spp.. Why Cronobacter spp. infection occurs only in neonates and premature babies, and why it is associated only with the consumption of infant formulae, is unclear. The US Food and Drug Administration issued an alert in April, 2002, to health-care professionals about the risks associated with Cronobacter spp. infections among neonates fed with milk-based infant formulae, and recommended certain bottle-preparation practices. The alert stated that a major contribution to the prevention of Cronobacter spp. infections in premature babies and neonates is the prevention of contamination of infant formulae during production and bottle preparation. However, knowledge of the aetiological and ecological characteristics of Cronobacter spp. is sparse, and its occurrence in factories that

71

Chapter 4

produce infant formulae, hospital kitchens, and households has hitherto not been studied in depth. We have started a systematic investigation of the habitat of Cronobacter spp..

Materials and methods
Environmental samples were obtained from nine factories and 16 households (Table1). In factories, samples were obtained by scraping or sweeping surfaces in the production-line environment or by sampling vacuum cleaner bags. Samples from households were taken from vacuum cleaner bags. From each sample, we analysed 10 g for the presence of Cronobacter spp. by selective enrichment in lauryl sulphate tryptose broth, supplemented with 0.5 mol/L salt, incubated at 45 ºC for 22 to 24 h. A loopful was streaked on violet red bile lactose agar to select for coliforms during the subsequent incubation at 37 ºC for 24 h. All coliform colonies on violet red bile lactose agar were streaked on to tryptone soy agar plates and incubated for 48 h in daylight at room temperature. For the identification of Cronobacter spp., we tested all yellow-coloured oxidase-negative colonies for the presence of -glucosidase. Presumptive Cronobacter spp. isolates were identified with the API 20 E test system (bioMerieux, Marcy L’Etoile, France) and were also characterised as Cronobacter spp. with the ribotyping technique.

Results and discussion
Cronobacter spp. was isolated with varying frequency from nearly all environments examined. Although the proportion of positive samples differed between environments, the differences were not significantly different (Table 1). The presence of Cronobacter spp. in factories producing milk powder, cereals, chocolate, potato flour, and pasta, as well as in domestic environments, strongly indicates that it is a widespread micro-organism. Since measures to prevent neonatal cases of Cronobacter spp. infections are being discussed in various countries by industries, public-health workers, and consumers, our findings may contribute to the current knowledge. Most importantly, the widespread nature of Cronobacter spp. should be taken into account in the design of effective control measures.

72

Medicine. M. Lai. and adults . 2002.Case reports and a review of the literature. J. Lett. Wanegue.Occurrence of Cronobacter spp. Environmental samples of various origins tested for the presence of Cronobacter spp. Farber. L.0..64 0. children. Environ. J. Roelofs-Willemse. K. 2001. Appl.0.0. A. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. Muytjens. Muytjens. Clin. J. Thermal resistance of Enterobacter sakazakii in reconstituted dried infant formula. Nazarowec-White. L. Enterobacter sakazakii infections among neonates.0.27 0. Appl. M. 4. Comparison of fecal coliform agar and violet red bile lactose agar for fecal coliform enumeration in foods. H. C. Microbiol. and data interpretation. infants. H. Enterobacter sakazakii meningitis in neonates: causative role of formula? Pediatr. or in writing the report. 5.64 0.42 0 . which is supported by the Nestlé Research Centre. Baylac. 80:113-122. and P.53 0.404 0.01 . Table 1.12 – 0. K.0. 1997.10 .03 . 1990. 73 . and L.78 0. 3.53 0.52 0.0. Lausanne.08 . H. 1988. 2. and G.0.19 . Origin Number of samples analysed 23 26 11 8 8 9 15 26 5 16 Number of samples positive for Cronobacter spp.. 24:9-13. The sponsor had no role in study design.. Infect. H.002 . References 1.03 .57 Milk powder factory Milk powder factory Milk powder factory Milk powder factory Chocolate factory Cereal factory Potato flour factory Pasta factory Spice factory Households Acknowledgments This research is part of PhD research project of Chantal Kandhai. Microbiol.. Switzerland. Leclercq. Jaspar. data analysis. Dis. 26:743-746. 68:1631-1638. A. Kollee.15 .0.0. (%) 2 (9) 9 (35) 1 (9) 2 (25) 2 (25) 4 (44) 4 (27) 6 (23) 0 5 (31) 95% confidential interval 0. data collection. Microbiol. and J. 9:372-373.

then what is? Maaike Arts UNICEF. C. when my eyes fell on the last paragraph of the article: “This research is part of a PhD research project . . 2004. from various factories. M. Nestlé produces infant formula and other breast-milk substitutes. Hanoi. 72 Ly Thuong Kiet Street. and the related contamination of infant formulae with Cronobacter spp. M Gorris. I was surprised to see only one recommendation from the authors at the end of the article: “The widespread nature of this micro-organism needs to be taken into account when designing preventive control measures. The Lancet. Van Schothorst. Guillaume-Gentil.” A few lines higher I read that no conflicts of interest were declared. in factories and households The isolation of Cronobacter spp. 74 . G. I was wondering why the authors did not make such recommendations. and other micro-organisms. M. Kandhai M. References 1. L. If this is not a conflict of interest. which is supported by the Nestlé Research Centre. W.Chapter 4 Appendix: Correspondence Cronobacter spp. Occurrence of Enterobacter sakazakii in food production environments and households. including infant formula factories (1). Vietnam. and (2) inclusion of a warning on infant formulae and other breast-milk substitutes that the product might be contaminated by Cronobacter spp. Reij.” Two other recommendations quickly come to mind: (1) enhancement of the promotion of and support for breastfeeding. 363: 39–40. .. are causes for concern. O.

6. We believe that our conclusion “The widespread nature of this micro-organism needs to be taken into account when designing preventive control measures” is the only one justified on the basis of the results we reported in the paper. Laboratory of Food Microbiology. MWR. Chantal. Olivier Guillaume-Gentil. and Nestlé Research Centre. we are convinced that we clearly disclosed our affiliations and financial support as required by the journal.Switzerland (OGG) References 1. as a starting point for designing effective control strategies. 6700 EV Wageningen. we cannot substantiate the recommendations suggested by Arts. Martine W. FAO/WHO. Switzerland. or in writing the report.Occurrence of Cronobacter spp. data collection. and data interpretation. and Mike van Schothorst Wageningen University. Leon G. As stated in the original publication. M. Gorris. PO Box 8129. Lausanne. data analysis. “The sponsor had no role in the study design. Italy. 2004) (1) as a potential source of information. Microbiological Risk Assessment Series No. M. Feb 2–5. 75 . Rome. we would like to draw readers’ attentions to the joint FAO/WHO Expert Workshop on E. LGMG. However. Netherlands (MCK. Quality and Safety Assurance Department. in infant formulae.” Kandhai. Reij. On the basis of the research we reported. Available at: http://who. The purpose of our study was to gain a better understanding of the ecology of Cronobacter spp. Enterobacter sakazakii and other micro-organisms in powdered infant formula: meeting report. sakazakii and other micro-organisms in powdered infant formula (held in Geneva.int/foodsafety/micro/ meetings/feb2004. Authors’ reply Maaike Arts asks why we did not include in our Research letter other recommendations about Cronobacter spp. With respect to the suggested conflict of interest. MvS). 2004.

Chapter 4 76 .

77 .Prevalence of Cronobacter spp.

J. fresh human milk (0/7). dry cereals (6/123). Part of this Chapter has been accepted for publication as “A study into the occurence of Cronobacter spp. fresh raw minced meats (7/222). possibly occurring in many different foods and environments causing illnesses predominantly through powdered infant formula. None of the samples taken from fresh raw bovine milk (0/51). follow-up formula for consumers > 1 year (1/5). M. as well as on and in humans. vegetables (2/47). is an opportunistic pathogen. can occur in a wide range of dry and wet foods. a broad survey of foods manufactured or marketed in The Netherlands.C. A.G. and L.01.1016/j. R. This widespread occurrence is important to consider when manufacturing foods for consumers at potential risk for Cronobacter spp.E. van Schothorst. not limited to infant formulae or milk products.M. M. R. was isolated from milk powders (7/175).2010.foodcont. This study confirms that Cronobacter spp. in The Netherlands between 2001 and 2005” Kandhai. Beumer. Cronobacter spp. was conducted over a 5-years period (2001 – 2005). To add to the existing knowledge on the occurrence of Cronobacter spp. infection. Food Control.J. and human skin (1/116) samples. spices (1/28). Gorris.H.Chapter 5 Abstract Cronobacter spp. M. including relevant non-food environments.. Heuvelink. 78 . and to help identify habitats that could be potential sources of this organism. powdered formulae for consumers < 1 year (8 /395).007. doi:10. van Tilburg.R.C. samples of fresh human faeces (1/98). Dijk. Using a specifically designed realtime polymerase chain reaction method for definite confirmation. Reij. baby bottles (0/86) and bottle-nipples (0/95) were found to be positive for Cronobacter spp.. 2010.W. other powdered instant products (1/182).

” is used consistently in the subsequent part of this paper. disease have been detected most commonly among newborn and very young infants in hospital nurseries and neonatal intensive care units. being rare in the frequency of involvement in food-borne disease but with a lethality rate up to 70% among neonates (30). 63. is particularly due to the fact that it has been recognised as an opportunistic pathogen for infants. Cronobacter muytjensii and Cronobacter dublinensis. The sixth species is preliminarily indicated as genomospecies I. facultative anaerobic bacteria. infection and at least 29 deaths 79 . To indicate the taxonomic change. within the family of Enterobacteriaceae. Although other Enterobacter species have been involved in nosocomial diseases. pre-term or low birth-weight infants and those immuno-compromised are at greatest risk (16). The International Commission on Microbiological Specifications for Foods (ICMSF) placed Cronobacter spp. 29). The Scientific Panel on Biological Hazards of the European Food Safety Authority (EFSA) noted that although Cronobacter spp. neonates up to ca 4 to 6 weeks of age. Introduction Cronobacter spp. 72). 11. Cronobacter malonaticus. where the micro-organism is referred to in general terms or where historic information that is not specific to the newly recognized particular species is cited. 50. 19. and has also been isolated from infants in association with necrotizing enterocolitis (18. non-spore forming. around 156 documented cases of Cronobacter spp. The genus was originally defined as a Enterobacter sakazakii species in 1980 (21). 55). 40) as six species in a new genus.. Cronobacter gen. 61. Cases of Cronobacter spp. motile. The micro-organism has been linked with serious infections in infants. All Cronobacter species have been linked retrospectively to clinical cases of infection in either infants or adults. in the category of severe hazards for restricted populations. Five of the new species are Cronobacter sakazakii. the importance of Cronobacter spp. the designation “Cronobacter spp. causing bacteraemia and meningitis. Outbreaks of infant meningitis. especially neonates (1.Prevalence of Cronobacter spp. has caused illness in infants up to 1 year. (Enterobacter sakazakii) are Gram-negative. and therefore all species should be considered pathogenic (20). Cronobacter turicensis. The first cases attributed to this organism occurred in 1958 in England (73). nov. but very recently it has been reclassified (39. Since then and up to July 2008. necrotizing enterocolitis and or bacteraemia have been reported (2.

including milk powder factories and other factories (43). an average lethality rate of 19% was calculated. insects (27. However. Specific international attention has been given to the safety of food for infants and young children as it relates to the possible presence of Cronobacter spp. 49. (18-20). sausage meat. dust from households (43). The Food and Agriculture Organization of the United Nations (FAO) and the World Health Organization (WHO) have established two international risk assessments on the topic (18. beef. only powdered infant formulae have been linked to outbreaks of infections as an unequivocal source (1. respectively (23). Notably. < 10% and 19%. Considering the recent review of outbreaks caused by Cronobacter. 61). 36. has been isolated from a variety of environmental samples. the overall lethality of 67 invasive cases was estimated to be 26. 23). for drafting an new Code of Hygienic Practice (6) aimed at providing risk management guidance at the international level.. More recently. 19) that have provided a scientific basis for the Codex Alimentarius Committee on Food Hygiene (CCFH). In the Code of Hygienic Practice products for infants up to 12 months of age are considered. in powdered formulae intended for this consumer group. although several expert meetings have established that those younger than 2 months and infants of HIV positive mothers are at greatest risk for infection caused by Cronobacter spp. The lethality of Cronobacter meningitis. bacteraemia and necrotizing enterocolitis (NEC) was estimated to be 41. In order to design appropriate and effective control measures for powdered infant formula. 62). 9. 59.Chapter 5 from all parts of the world appeared in the published literature and in reports submitted by public health organizations and laboratories (20). powdered fruit flakes) and other environmental sources. as well as blenders and cleaning brushes used to prepare infant feeds in hospital kitchens (4.7%. vegetables. including water (22). at low concentrations in powdered infant formulae has been frequently reported (8. The micro-organism has been isolated from several food products (i.e. This would include identification of its natural habitats and/or specific niches. need to be well understood. The occurrence of Cronobacter spp. as well as its occurrence 80 . the various possible sources of Cronobacter spp.9%. 53) and rats (24). 28. Cronobacter spp. cheese.

To increase our knowledge on sources and occurrence of the microorganism. fresh human faeces (n=98). with the exception of 27 samples imported from the Ukraine that were intended for animal feed.and EasternEuropean countries: Ukraine (43). Lithuania (5). Of the 175 milk powders tested. and bottle-nipples (n=95). fresh raw minced meats (n=222). 42 were produced in The Netherlands while the remainder (133) were taken from cargos and bulk goods as part of systematic examinations of food imports into The Netherlands. Though methods differed. other powdered instant foods (n=182). The milk powders originated from Middle. Latvia (4) and Slovakia (2). These investigations considered food products marketed in the country in retail and wholesale. fresh human breast milk (n=7). A number of different methods were used in these investigations. namely between 2001 and 2005.. dry cereals (n=123). several investigations were conducted in The Netherlands over a five years period. For part of the investigation period. in addition to analysis for Cronobacter spp. in factory environments and as a contaminant of food ingredients and finished food (or feed) products. baby bottles (n=86). human skin (n=116). a specifically designed real-time Polymerase Chain Reaction (PCR) protocol was used for definite confirmation of presumptive Cronobacter spp. reflecting the rapid evolution of detection and identification methodologies during the time of the investigation. possibly related to the occurrence of the pathogen. 81 . imported food and feed products. powdered formulae (PF) for young children below 1 year of age (n=395). in all cases. Estonia (18). follow-up formulae (FUF) for consumers > 1 year of age (n=15). dried vegetables (n=47). isolates. Material and methods Samples The occurrence of Cronobacter spp. was determined in milk powders (n=175). the level of Enterobacteriaceae was also assessed as an indicator of product hygiene. the Czech Republic (8). Poland (53).Prevalence of Cronobacter spp. as well as food products. dry spices (n=28). fresh raw cow’s milk (n=51). All milk powders were intended for human consumption. manufactured in The Netherlands.

and poultry (n=5). as well as raw meats. For this reason. The Netherlands).V. including 112 powdered formulae for consumers below 1 year of age and 3 follow-up formulae. dips (n=19). isolation and identification methodology for this particular micro-organism experienced rapid improvement and innovation. Fresh cow’s milk (n=51) was collected from bulk tanks from 26 farms at two time points. were asked to scrub off as much skin area as possible while taking a shower. two months apart. Nijmegen. whereas 12 instant pudding powders had been collected directly in factories in The Netherlands. 20-40 years (n=96) and 40-60 years (n=12).. including beef (n=141). Additionally. and 48 from persons older than 14 years. During the period of investigation. and vegetables (n=47) and spices (n=28). comprising of dry powdered foods. mixed beef and pork (n=50). Samples of human faeces were collected from volunteers. Several other food products were obtained from retail. 9 from children between 9 and 14 years. Human breast milk samples (n=7) were obtained from nurturing mothers. Used baby bottles and bottle-nipples were collected at day care facilities and households on the day of use. 82 . Volunteers (n=116). Enrichment and isolation of Cronobacter spp. flavoured instant drinks (n=33).Chapter 5 A total of 295 powdered formulae obtained from retailers across the country were sampled. including ice cream mixtures (n=62). belonging to the age categoried 0-20 years (n=8). and 12 follow-up formulae. enjoying a general healthy status and belonging to different age categories: 41 samples from children younger than 5 years of age. 27 samples of ice cream mixtures were obtained from wholesale. and dry cereals (n=123). An additional 115 powdered formulae were obtained directly from five manufacturers in The Netherlands. Samples of human skin particles were obtained by using scrub gloves (Parfumerie Douglas Nederland B. coffee creamers (n=4). calf (n=10). intended for young children older than 12 months. Most volunteers scrubbed only their arms and legs. pork (n=16). including 283 powdered formulae intended for use for infants up to the age of 12 months. instant pudding powders (n=25).

After overnight incubation at 37 °C. the selective enrichments were directly streaked onto TSA plates (Method B). which is further described in the following. further referred to as mLST (25). Typical coliform-colonies were picked and tested for the production of yellow-pigmented colonies on Tryptone Soya agar (TSA. the EE cultures were streaked onto VRBL agar plates and incubated overnight at 37 °C (Method A2). different methods were employed in the course of the study. the pre-enrichments were streaked onto violet red bile lactose (VRBL. Marcy l’Etoile. Oxoid Ltd.5 ºC. England) agar (Method A1) and 10 ml portions were mixed with 90 ml of Enterobacteriaceae enrichment (EE. sakazakii specific 4-h α-glucosidase assay (45).Prevalence of Cronobacter spp. Colonies that produced yellow pigment after overnight incubation at 37 °C were tested with the E. Bruz cedex. Following overnight incubation at 37 °C. as described previously (45). Basingstoke. whereupon positive isolates were further identified and confirmed. Irrespective of the method employed. except that “Enterobacter sakazakii isolation agar” (ESIA. Positive isolates were further identified and confirmed as mentioned above. with reference to the various abbreviated identification codes used to identify the specific methods.) broth. colonies were further identified using the API 20E system (bioMérieux. France) was used instead of TSA (Method C1).5 ºC 83 . The samples of raw meats were examined essentially according to procedure B. the powders were dissolved by swirling slowly. Hampshire. 0.1 ml was added to 10 ml of lauryl sulphate tryptose broth (Oxoid Ltd.). 25 g or 50 g portions were carefully sprinkled onto a nine-fold volume of buffered peptone water (BPW) at a temperature of 45 °C.. Oxoid Ltd. After overnight incubation at 44  0.) and a positive αglucosidase reaction within 4 h. AES Chemunex.) supplemented with 0. France) and subsequently confirmed by a specific PCR assay (see below for PCR procedure details). Oxoid Ltd. The ESIA plates were incubated overnight at 44  0. all presumptive Cronobacter spp. After soaking for 30 to 60 min on the work bench. From the powdered food products collected in 2001-2003. From the BPW pre-enrichments of the dry powdered products and dry cereals collected in 2004 and 2005. Table 1 gives an overview of the methods.5 M sodium chloride (NaCl) and 1 mg/ml vancomycin (Oxoid Ltd.

d i EE. Turquoise coloured-colonies were streaked onto TSA for further identification and confirmation. ESIA. violet red bile lactose agar.d.. colonies were further confirmed and identified as described above. the ESSB enrichments were streaked onto ESIA plates (Method D).d. “Enterobacter sakazakii selective broth”. c h VRBL. were enriched in “Enterobacter sakazakii selective broth” (ESSB.d. 84 . n. Samples of used baby bottles were obtained by rinsing the inside of the bottles with 10 ml sterile peptone physiological salt solution containing 0. DFI. Table 1. not included in this method. Isolation VRBLc VRBL TSAf ESIAg ESIA ESIA ESIA+TSA+VRBL+DFIi f BPW.b EEd mLSTe mLST mLST ESSBh n. “Enterobacter sakazakii isolation agar”.d. Germany) and adding the rinsing solution to 90 ml ESSB. tryptone soy agar.Chapter 5 and turquoise-coloured colonies were picked off and streaked onto TSA for further identification and confirmation as described above. The raw milk samples (10 ml) and the entire scrub gloves containing human skin particles were directly selectively enriched. used in this study Method A1 A2 B C1 C2 D E a b Pre-enrichment BPWa BPW BPW BPW n. g n. Methods for the isolation of Cronobacter spp. Bottle-nipples were individually placed directly into 100 ml ESSB. Ten grams of the human faeces samples were added to 90 ml ESSB. Selective enrichment n.1% Tween 80 (Van Merck-Schuchardt. used baby bottles and bottle-nipples. Hohenbrunn. AES Chemunex). in mLST (90 and 100 ml.d. n. respectively) without preenrichment in BPW. e mLST. TSA. The mLST enrichments were streaked onto ESIA plates (Method C2). Enterobacteriaceae enrichment broth. buffered peptone water. Druggan-Forsythe-Iversen agar. Samples of human faeces. Following 24 h incubation at 37 °C. Presumptive Cronobacter spp. ESSB.d. modified lauryl sulphate tryptose broth.

The new real-time PCR method targets a specific sequence within the gene encoding the major outer membrane protein (OmpA) of Cronobacter spp. Foster City. USA). in violet red bile glucose (VRBG. Presumptive Cronobacter spp. Colonies of presumptive Enterobacteriaceae were. (GenBank accession number DQ000206.C. confirmed by means of tests for fermentation of glucose and the presence of oxidase (31). VRBL and Druggan-Forsythe-Iversen (DFI.2. Oxoid Ltd. TSA. PCR confirmation of Cronobacter spp. Enterotube II (Becton Dickinson an Company) and the presence of α-glucosidase (45) and subsequently confirmed with the PCR assay. USA). National Centre for Biotechnology Information). were examined for the number of Enterobacteriaceae (CFU/g) by using the pour plating technique. colonies were confirmed using a newly developed PCR assay. The vegetables and spices were homogenized in peptone physiological salt solution (1:10 (w:v)) and dilutions were not enriched but directly streaked onto the following four media: ESIA.) agar plates (Method E).). The specificity of the sequences was tested by a blast search in GenBank (BLASTn version 2. Enumeration of Enterobacteriaceae In order to assess the general hygienic status of all products. BBL Crystal (Becton Dickinson and Company. The Cronobacter spp. all samples collected in 2005.10. Microbact (Oxoid Ltd. Washington. which is specific for all Cronobacter species tested. The set of oligonucleotides was designed using Primer Express software (Applied Biosystems. Typical colonies were confirmed with the following biochemical identification methods: API 20E (bioMérieux). Oxoid Ltd. USA) (Table 2). except for two samples of flavored instant drinks and one sample of coffee creamer.) agar (32).Prevalence of Cronobacter spp. Sparks Maryland. The oligonucleotides were purchased from Eurogentec (Liege.. Belgium). 85 . in accordance to the ISO 21528-1 method. probe was labelled at the 5’ end with the reporter dye 6-carboxy-fluorescein (FAM) and at the 3’ end with the black-hole quencher (BHQ). National Centre for Biotechnology Information. D.

Germany) system and the MagNA Pure Compact® Nucleic Acid Isolation kit I (Roche Diagnostics). Subtyping of Cronobacter spp. Mannheim. The Pulsed Field Gel Electrophoresis (PFGE) technique of contour-clamped homogeneous electric fields (CHEF) was used for genomic typing of isolates confirmed as Cronobacter spp. Samples were amplified with an initial denaturation step at 95 °C for 10 min to activate the FastStart Taq DNA polymerase and 40 cycles of denaturation at 95 °C for 10 s and annealing and extension at 60 °C for 15 s. by PCR. Primers and probe used for the real-time PCR assay specific for Cronobacter spp. Those scored as positive by the instrument were confirmed by visual inspection of the graphical plot (cycle number versus fluorescence value) generated by the instrument. 35 isolates of Cronobacter spp. DNA was extracted using the automated MagNA Pure Compact® (Roche Diagnostics. Genomic DNA was digested in agarose plugs with XbaI (10 U) (Roche Diagnostics). Roche Diagnostics) in a 20 μl PCR mixture containing 2. DQ000206 For selectivity tests. Designation Esak ompA-F Esak ompA-R Esak ompA-probe a Sequence (5’ → 3’) ggt gaa gga ttt aac cgt gaa ctt gcg cct cgt tat cat cca aa ccc gga aaa gcg cat ggc c Positiona 319-342 388-369 348-366 Positions correspond to GenBank accession no. and 44 isolates representing a range of other genera and species were grown overnight at 37 °C on TSA. Amplification and detection were carried out in a LightCycler Instrument (version 2.5 μl of the sample DNA. 500 nM each of the primers. whereupon 100 μl was used for DNA extraction. Samples positive for the target gene were identified by the instrument at the cycle number where the fluorescence attributable to the target sequences exceeded that measured for background. 250 nM of the probe.0.Chapter 5 Table 2. The resulting fragments were resolved by CHEF-PFGE with a CHEF DR-III apparatus (Bio- 86 . The temperature transition rate was 20 °C/s. One loop-full of bacterial culture was suspended in 200 μl sterile distilled water. and 4 μl LightCycler TaqMan Master mix (Roche Diagnostics).

Isolates that had more than 95% fragments in common were designated “closely related”. Pseudomonas aeruginosa (n=3). Enterobacter amnigenus. Enterobacter aerogenes. flavoured instant drinks. All of the following 44 strains gave a fluorescence endpoint signal below the threshold line (n=1 unless indicated otherwise): Aeromonas hydrophila. instant pudding powders. 87 . The micro-organism was isolated from the following dry powdered food sources tested: 8 out of a total of 395 powdered formulae. Escherichia coli O157. USA) at a constant voltage of 200 V for 24 h at 13 °C and a linearly ramped pulse time of 5 to 60 s. Isolation of Cronobacter spp. all 35 Cronobacter spp. 1 out of 15 follow-up formulae. Shigella flexneri (n=2). Escherichia coli. Shigella sonnei (n=3). gave a positive signal in the real-time PCR assay. Listeria innocua. 7 out of 175 milk powders. Richmond.. Proteus mirabilis (n=2). Isolates were considered “indistinguishable” if 100% of fragments were identical. Pseudomonas putida. Citrobacter freundii (n=2). strains tested. Klebsiella pneumoniae (n=3). California.Prevalence of Cronobacter spp. from the different samples investigated in this study. Belgium). Table 4 summarizes the results of the isolation of Cronobacter spp. Enterococcus faecium.. Klebsiella oxytoca (n=2). Salmonella Enteritidis. Proteus vulgaris. which included 13 official reference strains and 22 isolates from various sources obtained in previous investigations at the Laboratory of Food Microbiology of theWageningen University (43). and Yersinia enterocolitica (n=3). Rad Laboratories. Listeria monocytogenes (n=2). Citrobacter spp. Hafnia alveii. Serratia marcescens. The fingerprints generated were processed using BioNumerics software (Applied Maths. Salmonella Typhimurium (n=2). Enterobacter cloacae (n=6). Kortrijk. Results Selectivity of the real-time PCR As shown in Table 3. Klebsiella spp. 1 out of 182 of the other powdered instant products (including ice cream mixtures.

were all confirmed using a newly developed real-time PCR assay. with the exception of the human faecal isolate.Chapter 5 dips and coffee creamers). raw minced meats (7 out of 222). lactardi Cronobacter spp. Citrobacter freundii Citrobacter freundii Strain code/Origin Enterobacter sakazakii NCTC 8155 Enterobacter sakazakii NCTC 9238 Enterobacter sakazakii NCTC 11467 Enterobacter sakazakii DSM 4485 Enterobacter sakazakii NCTC 9844 Enterobacter sakazakii NCTC 9846 Enterobacter sakazakii NCTC 9529 Enterobacter sakazakii DSM 18705 Enterobacter sakazakii ATCC 51329 Enterobacter sakazakii DSM 18703 Enterobacter sakazakii DSM 18702T Enterobacter sakazakii DSM 18706T Enterobacter sakazakii DSM 18707T Enterobacter sakazakii Isolates from Wageningen University culture collection M800 PHLS 073-177 ATCC 8090 NCTC 6272 Results realtime PCRa + + + + + + + + + + + + + - 88 . lausannensis Cronobacter dublinensis subsp.. In addition. was not isolated from the fresh raw cow’s milk samples nor from the samples of fresh human breast milk. which proved to be highly specific for Cronobacter spp. Table 3. used baby bottles or bottle-nipples. dublinensis Cronobacter muytjensii Cronobacter turicensis Cronobacter malonaticus Cronobacter dublinensis subsp. Cronobacter spp. using the probes as described in Table 2 Strain Cronobacter sakazakiib Cronobacter sakazakii Cronobacter sakazakii Cronobacter sakazakii Cronobacter dublinensis Cronobacter dublinensis Cronobacter genomospecies 1 Cronobacter dublinensis subsp. isolates obtained throughout the study period. Other food products yielding Cronobacter spp. The presumptive Cronobacter spp. isolates were: dry cereals (6 out of 123). as shown in Table 3 and shown below. (n=22)c Aeromonas hydrophila Citrobacter spp. the micro-organism was isolated twice from human sources. including faeces (1 out of 98 fresh human faeces) and skin scrubbings (1 out of 116 human skin scrub samples). vegetables (2 out of 47) and spices (1 out of 28). Real-time PCR results of 79 test strains.

Klebsiella oxytoca Klebsiella oxytoca Klebsiella pneumoniae Klebsiella pneumoniae Klebsiella pneumoniae Listeria innocua Listeria monocytogenes Listeria monocytogenes Pseudomonas aeruginosa Pseudomonas aeruginosa Pseudomonas aeruginosa Pseudomonas putida Proteus mirabilis Proteus mirabilis Proteus vulgaris Salmonella Enteritidis Salmonella Typhimurium Salmonella Typhimurium Serratia marcescens Shigella flexneri Shigella flexneri Shigella sonnei Shigella sonnei Shigella sonnei Yersinia enterocolitica Yersinia enterocolitica O3 Yersinia enterocolitica O9 a b Strain code/Origin Isolate VWA Isolate VWA ATCC 23355 Isolates VWA ATCC 29212 ATCC 25922 NCTC 12900 NCTC 8105 Isolate VWA ATCC 49131 Isolate VWA ATCC 33495 ATCC 13883 Isolate VWA NCTC 11288 SLCC 2479 NCTC 5348 ALM32 ATCC 27853 NCTC 10662 Isolate VWA ALM29 NCTC 11938 ATCC 13315 ALM36 ALM40 ATCC 13311 ATCC 13880 ALM60 Isolate VWA ATCC 11060 Isolate GG&GD Haarlem Isolate VWA ALM5 Isolate UMC St. 89 . 2007 (39). -.Prevalence of Cronobacter spp. Continued Strain Enterobacter amnigenus Enterobacter aerogenes Enterobacter cloacae Enterobacter cloacae (n=5) Enterococcus faecium Escherichia coli Escherichia coli O157 Hafnia alveii Klebsiella spp.. c Presumptively identified by the α-glucosidase assay (45) and the API 20E system (bioMérieux). positive. Nijmegen CCUG8239A Results realtime PCRa - +. Table 3. Radboud. negative The Cronobacter species identification is according to Iversen et al.

5 0 – 20.0) Mixed beef / pork Pork 0 / 16 Calf 0 / 10 Poultry 0/5 2 / 47 (4.6 50 g 50 g 50 g 50 g 50 g 25 g 10 g 10 ml 10 ml 10 g Scrub glove C1 C1 C1 C1 C1 E E C2 C2 D C2 90 .8 – 7.1 – 17.0 0 – 41.5) 2 / 101 (2.1 – 19.3 – 7.3 – 6. >1 yr 0 / 4 Ice cream 0 / 87 mixtures 0/2 Flavored instant 0 / 33 drink Instant pudding 0 / 12 powder 1 / 25 (4.5 0.0 – 9.5 – 13.2 – 9.6 3.4 0 – 60.8 0.6) 0 / 14 Follow-up 1 / 11 (9.1 0 – 10.4) 5 / 50 (10. Prevalence of Cronobacter spp.9) Human skin 2001 2005 2001 2002 2003 2004 2005 2004 2005 2003 2005 2005 2003 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 2005 25 g 50 g 25 g 25 g 50 g 50 g 50 g 50 g 50 g 50 g 50 g 50 g 50 g 50 g 50 g 50 g 50 g Ae B A A A B B B B A B B A B B B B 0. Confidence # positive/total (%) interval (%)a 1.2 2.6 0 – 3.0) Human faeces 1 / 116 (0.8 0 – 7.2 0.6 0 – 30.0 0 – 5.0) Dip 0 / 19 Coffee creamer 0 / 4 6 / 123 (4.6 0 – 60.2 0 – 4.1 0.9 0 – 52.1 – 12.2 0.2 – 40.1) 0/4 PF < 1 yr 1 / 40 (2.6) Spices Raw cow’s milk 0 / 51 0/7 Human milk 1 / 98 (1.4 0 – 4.6 0 – 26.8 0 – 17.0) 0 / 101 5 / 139 (3.1 0 – 84.1) formulae.6 – 20.2) Dry vegetables 1 / 28 (3.6 0 – 23. in samples of various origins.6 Year of Sample sampling size Methodb 7c / 171d (4. using methods as described in Table 1 Sample Powdered foods Milk powder Prevalence of 95% Cronobacter spp.Chapter 5 Table 4.9 0.6 1.9) Dry cereals Raw minced meats Beef 2 / 141 (1.7 0 – 60.

2 0 – 3. 91 . Because of degradation of genomic DNA. Continued Sample Baby bottles Bottle-nipples a Prevalence of Cronobacter spp. e A. In total. four of the 34 isolates subjected to the subtyping procedure (one of the six isolates from dry cereals. d Samples collected from cargos and bulk goods and included 27 products intended for animal feed. Table 4. assuming the drawing process is a series of independent draws from a binomial distribution with the probability p. Subtyping of Cronobacter spp. and 2 in Poland. whereas two isolates generated patterns that could not be distinguished (Figure 1). Of the remaining 30 isolates that had been successfully subtyped. visited one month after each other. # positive / total (%) 0 / 86 0 / 95 95% Confidence interval (%)a 0 – 4. the isolate from the pudding powder and one of the two beef isolates and the human skin isolate) did not yield a result. 35 isolates were obtained of which 34 were positively confirmed by PCR. Of the total of 35 Cronobacter spp. b The methods used are summarized in Table 1. includes use of either method A1 or method A2. which had been purchased from two different outlets of the same supermarket chain. These two isolates were cultured from two samples of minced mixed beef and pork. isolates. 2 in Ukraine. the remaining isolate (from human faeces) had not been kept for molecular confirmation and PFGE subtyping.Prevalence of Cronobacter spp. 34 were subjected to PFGE subtyping. c 3 samples were produced in The Netherlands.8 Year of Sample sampling size 2005 2005 10 ml rinse each Method D D The 95% confidence interval was calculated. 28 isolates generated a unique XbaI restriction pattern each.

Cluster analysis of PFGE fingerprints of the 30 Cronobacter spp. Note that the PFGE fingerprints of the faeces isolate is not included in this Figure.Chapter 5 Figure 1. * indicates the two generated patterns that could not be distinguished. 92 . as it was not analysed. isolates.

b Detection limit. 680 CFU/g.000 – 10.100 100 – 1. A wheat bran sample containing over 10. 43 (19%) between 1. 91 CFU/g. The sixth dry cereal sample (a type of muesli) that had been found positive for Cronobacter spp. was below the detection limit.000 a Dry productsa 223 1 0 0 1c Raw minced meats 28 47 80 43 24 b All dry samples collected in 2005. 80 (36%) between 100 and 1.Prevalence of Cronobacter spp.. Out of 222 samples. However.000 > 10.000 CFU/g.000 CFU/g. A total of 223 dry powder samples had Enterobacteriaceae counts below the detection limit (≤ 10 CFU/g). and these included 5 of the 6 dry cereals samples that had been found positive for Cronobacter spp.000 1.000 CFU/g.000 and 10.000 CFU/g Enterbacteriaceae. 110 CFU/g. c Wheat bran. was negative for Cronobacter spp.000 CFU/g. Enumeration of Enterobacteriaceae in products collected in 2005 Enterobacteriaceae (CFU/g) 10 10 .. Enterobacteriaceae counts were found to be quite variable in the various minced meats. including dry powders and cereals as well as raw minced meats. and 2. The number of Enterobacteriaceae in the instant pudding powder that yielded Cronobacter spp. contained Enterobacteriaceae at 15 CFU/g. for three samples Enterobacteriaceae counts were not done. 93 . Table 5. 47 (21%) between 10 and 100 CFU/g. the Enterobacteriaceae counts were:  10 CFU/g. 55 CFU/g. and 24 (11%) over 10.. except for two samples of flavored instant drinks and one sample of coffee creamer. (Table 5). 28 (13%) had counts below detection. 250 CFU/g.. For the seven meat samples positive for Cronobacter spp. Enterobacteriaceae counts Enterobacteriaceae counts were investigated in products sampled in 2005.

Chapter 5 Discussion Isolation and confirmation methods An investigation of the occurrence of Cronobacter spp. 35. as in our Methods B. C1. In the selective enrichment step. 25. mLST broth is used (25) in combination with a relatively high incubation temperature to suppress potential competitors. It may proof useful for tracing infections to the source in sporadic cases or outbreak investigations. This is a challenge considering the relative low levels of this micro-organism that are found in foods and the fact that it is present among other micro-organisms. However. as compared to other Enterobacter species (60). 65. and its prevalence in a variety of samples of different origin was carried out in The Netherlands between the years 2001 and 2005. indicated that it represents a reliable confirmation test (Table 3) which is able to differentiate all Cronobacter spp. 51. The PFGE subtyping method also showed to be a reliable method to subtype Cronobacter spp. following in a timely way the ongoing developments and improvements in this field. and C2 (Table 2). The results obtained with the specific real-time PCR method. During this study. isolates. or for the identification of contamination sources in food manufacturing operations. ideally providing a specific and sensitive enrichment procedure. Several isolation and (molecular) detection methods have been developed over the past years aiming to detect and type Cronobacter spp. from many other Enterobacteriaceae. 37. it has been demonstrated that some 94 . 41. which we developed to further improve detection and definite confirmation. 52. While most of the recent studies focused on the development of molecular detection/confirmation of Cronobacter spp. 33. several chromogenic and fluorogenic agar media have been described for Cronobacter spp. 46. Regarding detection. The ISO procedure for milk and milk products published in 2006 (33) uses a selective enrichment procedure part of which was applied in the current study. (34. mainly based on its unique feature. 34. 68).. various traditional and innovative isolation and (molecular) detection/confirmation methods were used. 45. further research should focus on improving the enrichment step. in particular in powdered infant formulae (13. 52. 70). 64).. being the presence of α-glucosidase. as this so far remains an important bottle-neck (13.

on human skin and in fecal samples was rather low (~1%). with the micro-organism being isolated once from a faecal sample of a female adult. Cronobacter screening broth (CSB). and substantiates humans as potential sources for contamination. 22. 59) as well as from a variety of other foods (3. including soil (15) and insects (27. isolates do not grow well in mLST (37). (1980) previously reported isolating the micro-organism from various clinical sources. 71). food factory environments (43. therefore recently.Prevalence of Cronobacter spp. 49. 58. 18-20. previously. 10.. could be isolated from the stool of a child with diarrhea (66) and from stool samples of hospitalised infants (0. 95 . consumer homes (45). the microorganism was not detected in the seven human breast milk samples investigated.. These observations highlight that the microorganism can be present on or in humans. Cronobacter spp. 26).5%) (48). Our study of Cronobacter spp. While the occurrence of Cronobacter spp. its presence in skin scrapings is. making them a potential source of contamination both during product manufacture as well as food preparation. 36. The prevalence of Cronobacter spp. has been isolated from powdered infant formulae and milk powders (8. either being produced within the country or reaching the national market via import from other countries. in a variety of food and feed products. More recent studies also reported that Cronobacter spp. confirms and further extend the results of previous studies on the occurrence of the micro-organism. this opportunistic pathogen may be present in a variety of foods and feeds. Notably. including a child’s stool (21). The results obtained (Table 4) show that also in The Netherlands. has been developed (35). 57. Though prevalences have not been systematically reported. 37. Occurrence and prevalence of Cronobacter spp. as well as once from skin particles collected from another female adult.5%) and patients in the age group 61-70 years (6. a new enrichment broth for Cronobacter spp. and environmental samples. Farmer et al. Cronobacter spp. with prevalences ranging from 1.4 to 10. in human faeces in the current investigation is not a new observation.0% of samples where positives were observed.

Citrobacter spp. in various sources. legume products. So far only powdered infant formulae have been definitely attributed to cases of Cronobacter spp. it only investigated its prevalence. 42.. while levels in contaminated products may be of the order of 10-3 CFU/g (12.7% (74) and less recently 14% (59). it is not apparent that there is one single. Enterobacter agglomerans (Pantoea agglomerans) and Serratia spp. fruits and vegetables. (45). Based on the habitat profile of these possibly coinhabiting micro-organisms and their physiological features. could be isolated from vegetable products (22. meat and fish and products made from these foods. actual data on prevalence and concentration are scarce. as well as water suitable for food preparation. 6. 2008) (20). a range of other Enterobacteriaceae were isolated. 48. it is likely that these microorganisms (and may be thus also Cronobacter spp. such as cereals. ready-to-eat. In a recent review (22). Concerning prevalence in powdered formulae. 69) and the vegetable drawer of domestic refrigerators (2. The spectrum of food contaminated with Cronobacter spp. Enterobacter cloacae. and most of the investigative attention has been focused on these products. The current study did not determine the actual levels of Cronobacter spp. most important natural habitat for this micro-organism. identified to-date. Relating to concentration.) are associated with the phytic flora (56. fermented and cooked food products). Following prevalence data were obtained (with prevalence 96 . operations with good practices in place should be able to achieve levels of Cronobacter spp. of 10-4 CFU/g or better (6). some reported prevalence figures in positive batches are about 1% and 3% (data quoted in FAO/WHO. the widespread occurrence of Cronobacter spp. 19). in dry environments. Recent research indeed showed that Cronobacter spp. In the course of previous research into the occurrence of Cronobacter spp. was isolated from dried vegetables (4. noting that the presence of the microorganism had been confirmed in several foods and food ingredients from plant origin. Escherichia coli. as Cronobacter spp.6%) (Table 4). infection.. seems to cover both raw/fresh and processed food (being dried. in food and beverages other than infant formula was documented.2% out of 137 refrigerators) (47). Nevertheless.. as well as in foods of animal origin such as milk. Erwinia spp. Overseeing the variety of sources of Cronobacter spp. herbs and spices. Our results extend these observations.2%) and spices (3. 69). frozen.Chapter 5 53). such as Klebsiella spp.

2.9% for dry cereals. and raw minced meats tested negative for Cronobacter spp.1% for follow up formula intended for children > 1 year.Prevalence of Cronobacter spp. in these products is relatively infrequent or absent and that levels of Enterobacteriaceae in general were found to be below the detection limit. and special diet formula) thus are key to the health and well-being of infants and young children and our finding that the prevalence of Cronobacter spp.1% for milk powders. The apparently sound hygiene status of the products investigated is important because milk powders represent a basic ingredient in the majority of dry powdered foods. powdered formulae. are positive indicators of the product safety and hygiene. 1. as did dry powder products such as ice-cream mixtures. expressed per specific investigation. instant pudding powders. 4.0% for instant pudding powders. It is apparent that prevalence rates of Cronobacter spp. instant drinks. All these products (being powdered infant formula. Of the two cereal samples showing levels of Enterobacteriaceae above detection (Table 5). follow-up formula..6% for powdered formulae intended for consumers < 1 year. in powdered formulae and other foods/feeds are in the similar low range as reported for PIF as mentioned above and concur with prevalence estimates of Jung and Park of 5 to 10% for agricultural food components such as brown rice and tomatoes (42). Table 5 clearly shows that most of the dry products (including milk powders) contained low levels of Enterobacteriaceae (< 10 CFU/g). including drymixed powdered infant formulae and follow-up formulae. were detected in our study and others may be related to limitations of (selective) enrichment and isolation methods used. The relatively low frequency that Cronobacter spp. as indicated in Table 4): 4. 4. in our specific survey some samples from milk powders. the one with the lowest level of contamination (a muesli product) was positive for 97 . 4.03. the Enterobacteriaceae levels were determined to possibly link product hygiene to the occurrence of Cronobacter spp. Notably.4 to 10.0% for raw minced meats.2% for dried vegetables and 3. 9. but may as well reflect that that the level at which Cronobacter spp. In powdered products and cereal samples taken in 2005 and raw minced meat samples.6% for spices. actually is present in various products/samples indeed is very low. dips and coffee creamers.

a picture has emerged that the pathogen is widespread in nature. low-birth weight (< 2500 g). Also other infants below the age of 1 year are at particular risk.. in/on humans and utensils. Indeed. confirmed and documented cases are extremely sparse at the moment. those at greatest risk of infection are neonates (< 28 days). To consider all such sources and vectors may not be practical nor necessary in assuring food safety for the relevant consumers. occurring in foods and food ingredients. particularly pre-term. 19) and of the European Food Safety Authority (EFSA) (16) have shown that Cronobacter spp. Thus. as a hazard in food manufacture This publication is one of a series of studies over the last several years by many investigators all dedicated to the elucidation of the occurrence and prevalence of Cronobacter spp. etc. tightest control is needed for products intended for the consumers at 98 . (17).Chapter 5 Cronobacter spp. though solid case reports are fewer. Cases occurring in this age-group are relatively well confirmed and documented. it has been reported before that although Enterobacteriaceae may be a good general hygiene indicator and are promoted as such by Codex Alimentarius for powdered infant formulae (6). Considering Cronobacter spp. considering the actual consumer risk the micro-organism poses when present in a food. poses different levels of risk of infection for different consumer age groups.. soil. insects. there may be many sources and vectors that can play a role in the spread and ultimate contamination of foods by the opportunistic pathogen. and immunocompromised infants and those less than two months of age and also infants of HIV-positive mothers (20). factory environments and people’s homes. Above that age limit. Overall. including sensitivity of the consumer and the actual exposure of consumers to the pathogen through consumption. There may be relatively fewer relevant sources and vectors to control. Consumer risk is determined by a combination of factors. while the sample with the highest level of contamination had been found negative. they do not necessarily provide any insights in the occurrence of Cronobacter spp. To limit consumer risk. Among infants. Risk assessments or comparable studies developed under the auspices of the Food and Agricultural Organization of the United Nations (FAO) and the World Health Organization (WHO) (18.

formula for special medical purposes and human milk fortifiers. probably due to their resistance to desiccation 99 . Whilst control of contamination of ingredients to some extent is possible by applying some form of thermal treatment during their production and assuring prevention of post treatment re-contamination. in powdered infant formulae as it impacts on the ultimate consumer exposure. 58). adherence to strict microbiological requirements and thorough selection and purchase procedures (12). These results suggest that also the hygiene awareness of personnel need to be taken into consideration to prevent or reduce the prevalence of Cronobacter spp.. in the factory environments. there are various routes by which Cronobacter spp. Improvements in processing conditions such as very strict management of water in high hygiene zones.Prevalence of Cronobacter spp. The current study once more highlighted the possibility for humans as vectors for the microorganism. for powdered formulae intended for very young infants. are present in the milk powder processing environment (43. Considering consumer exposure and risk. Therefore. including Cronobacter spp. i. 54. two important ones relate to food manufacture: 1) through ingredients added in dry mixing operations during the manufacturing of powdered formula. Cronobacter spp. highest risk. A recent Microbiological Risk Assessment conducted under the auspices of the Food and Agricultural Organization of the United Nation and the World Health Organization (19) has shown that factory hygiene may be a factor contributing to the risk posed by Cronobacter spp. in powdered infant formula. in high hygiene processing areas and in powdered formula preparation units (6). 2) through contamination of the formula from the processing environment in the steps during or following the drying.e. Recent guidance developed by Codex Alimentarius (6) therefore contains microbiological criteria for product safety related to Cronobacter spp. Codex provided guidance for the establishment of monitoring programs for Cronobacter spp. may help further reduce the level Enterobacteriaceae. current technology cannot completely eliminate this microorganism from the manufacturing environment and prevent it gaining access to the processing line and final food product. can end up in powdered formulae (6).

For relevant foods and sub-populations. have relatively short lag times and high specific growth rates at temperatures commonly found in those environments (38. but rather extend already available evidence that Cronobacter spp. The concentration and prevalence of Cronobacter spp. 14). during manufacture and processing as well as final preparation is paramount to limit the consumer risk. Our results do not point to a specific niche of Cronobacter spp. in prepared formula prepared in hospitals and homes has been recognised as a key risk factor for sensitive consumers as well (19. strict control of contamination from food ingredients. can be found in many different food products. making all these potential contamination sources of the micro-organism for foods during manufacture and preparation. such as dust samples in factories and households (43). Nevertheless. proper bottle preparation and handling practices are critical in order to sufficiently control the risk. Therefore and because contamination and proliferation of Cronobacter spp. to achieve. since strains of Cronobacter spp. One should consider using sterile liquid formula. Conclusion Cronobacter spp. 7.. once the powder has been reconstituted. which may indicate an appropriate level of hygiene at the day care facilities and homes sampled from. caregivers should be mindful of these risk factors in preparing and handling infant formulae since Cronobacter spp. 44). for those groups at greatest risk. 14) and can grow rapidly (44). 67). if not impossible. Due to its prevalence in dry environments. cells are able to survive over a long period in dry powdered infant formulae (7.Chapter 5 (5. is not equal in all the potential sources and that not all sources are linked to the production and preparation of food for the 100 . is not a pathogen for the majority of the population. total elimination may be difficult. but rather a concern for specific sub-populations in the public. Measures for prevention should consider the actual target sub-population of a food as well as the consumer exposure due to the presence of the micro-organism in food. food environments as well as on and in humans. It is noteworthy that used baby bottles and bottle-nipples were found to be negative for Cronobacter spp.. Presence of water could lead to outgrowth.

Ludvigsson. Microbiol. F de Smet.. These foods are milk powder and infant formulae and the consumers at most risk are neonates (< 28 days). G. Baumgartner. (Enterobacter sakazakii) in different categories of ready-to-eat foods other than infant formula. G. as well as the contamination routes and growth opportunities that are still presented to the micro-organism in some of current practice. Food Microbiol. References 1. J. consumers that are at most risk. J. N. 2009. Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula. Switzerland. C. M. P. particular pre-term. A. low-birth weight (< 2500 g). Detection and frequency of Cronobacter spp. Clin. Bougatef. The authors would like to thank Ellen Brinkman and Ornella Vázquez for their contribution in collecting and analyzing samples. Preminger. Biering.Prevalence of Cronobacter spp.. 2001. Naessens. 2001. 1989. and those less than two months of age and also infants of HIV-positive mothers. Jonsdottir. Muyldermans. Microbiol. and O. it is a key that for these relevant foods and for these particular consumers. 90:356-358. adequate control measures are being implemented that take into account the widespread occurrence of Cronobacter spp. Lausanne. S. Bar Oz. Liniger. K. C. and C. Acknowledgements This research has been financially supported by Nestlé Research Center. Enterobacter sakazakii infection in the newborn. Thus. Peleg. E. Block. M. Acta Paediatr. Marcel Zwietering is gratefully acknowledged for critically reviewing the manuscript. Three cases of neonatal meningitis caused by Enterobacter sakazakii in powdered milk. 136:189-192. 101 . and immuno-compromised infants.. A. Lauwers. Clin. Int. Over the past years several improvements have been made to the (pre)-enrichment step and microbiological culture methods for the recovery of Cronobacter spp. Grand. and S. Karlsson. J. 3. Iversen. 39:293-297. O. and I. A. 27:2054-2056. and also the volunteers who donated faecal and skin particle samples. Steingrimsson. and this has helped to better investigate the prevalence in various locations. 2. Arad. Acker Van J.. A. 4. Clark. B.

. C. New Brunswick. Hill. Enterobacter sakazakii infections associated with the use of powdered infant formula . Hartantyo. W. Caubilla-Barron. G. Estuningsih. Jaradat. 17. Cawthorn. Lardeau. EFSA. A. 2007. Gaspar. Path. D. Appl. Food Microbiol. Diagn. Desiccation and heat tolerance of Enterobacter sakazakii. 2008. J. 145-185. 13:467-472. R. 7. D. O'Hara. J. Int. M. R. Production of powdered infant formulae and microbiological control measures. and S. O. M... Washington. Farber. 2002. In: J. International survey of Cronobacter sakazakii and other Cronobacter spp. 6. E. C. EFSA. Callaway.Tennessee. Forsythe (eds. C. 2009. C. and D. 95:967-973. 2009. J. 2001. N. L. and H. J. 1990. 15. CAC/RCP 66-2008. P. P. C. Breeuwer. Foodb. The EFSA Journal. J.. J. 2008. J. J. Opinion of the scientific panel on biological hazards on the request from the commission for review of the opinion on microbiological risks in infant 8. Food Prot. Druggan.C. Microbiol. Witthuhn. Culture media for the isolation of Cronobacter spp. 13. J. Buchanan. Forsythe. Int. 16.. 2007. J.). Forsythe. Dry stress and survival time of Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered infant formula. Cooksey. 127:129-138. Osaili. 51:297-300. R.. J. and S. Park. H. P. p. Enterobacter sakazakii (emerging issues in food safety) ASM Press. 14. Edrington. Hoffman. J. Fox. M. T. 68:1900-1902. T. Food Microbiol. Dis. Opinion of the scientific panel on biological hazards on the request from the commission related to the microbiological risks in infant formulae and follow-on formulae. Food Microbiol. L. Shaker. N.Chapter 5 5. Microbiol. Anderson. and C. Chap. R. S. 12. Joosten. 70:2111-2117. C. J. Sani. Int. D. MMWR. 9. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial outbreaks. P. Dis. A. and R. S. J. S. S.. 2005. CDC. S. Pathogen prevalence and influence of composted dairy manure application on antimicrobial resistance profiles of commensal soil bacteria. and R. 6:217-224. Quintas. Z. Code of hygienic practice for powdered formulae for infants and young children... N. Peterz. 2004. Evaluation of different methods for the detection and identification of Enterobacter sakazakii isolated from South African infant formula milks and the processing environment. Survival of Enterobacter sakazakii in a dehydrated powdered infant formula. C. Cordier. M. Steingrimsson. 136:169-178. 102 . 113:1-35. Food Prot. Clark. New Jersey. 2009. W. and R. CAC. Siqueira. Iversen. 2003. in follow up formulas and infant foods. Infect. 10. 2008. B. Nisbet. Botha. T. 11. and S. 136:185-188. Edelson-Mammel. Jackson.

FAO/WHO. and L. 15 Rome.) in powdered followup formulae: meeting report. L.. 2005. T... D. E. Food Microbiol. Micro-organisms in Foods 7: Microbiological Testing in Food Safety Management. East Afr. Enterobacter sakazakii infections associated with the use of powdered infant formulaTennessee. 18. and E. 103 . Hickman. 2002. M. N. R. 2002. Himelright. Int. V. 2005. Med. 28:1297– 1304. M. J. O. Harris. Rome. Gathura. De Ware(n)chemicus. Food Prot. 30. Enterobacter sakazakii in food and beverages (other than infant formula and milk powder). 20. 2003. J. 9:1355-1356. M. B. 32:1730. B. 2004. Kornacki. Food Microbiol. 22.. Italy. 1980. 27. 2008. 25. and H.. V. A simple an rapid cultural method for detection of Enterobacter sakazakii in environmental samples. 2007. Microbiological Risk Assessment Series No. R. V. Zwartkruis-Nahuis. and the Enterobacteriaceae study group. P. Brenner. J. Syst. Italy. Infect. FAO/WHO. M. FAO/WHO. D. Friedemann. 2006. Farmer III. Infect. J. Kyule. Antimicrobial resistance of bacterial organisms isolated from rats. 23. 29. F. Marugg. J. Kandhai. 28. Int. Gakuya. de Boer. Lorch. Isolation of Enterobacter sakazakii from midgut of Stomoxys calcitrans. Beuchat. and S. Kodde. Microbiol. J. J. I. Anderson. New York. Enterobacter sakazakii: A coliform of increased concern to infant health. formulae and follow-on formulae with regard to Enterobacteriaceae as indicators. JAMA. 19. Dis. J. 21. Microbiological Risk Assessment Series No. Emerg. J. Asbury. Int. Springer Publishers. F. The EFSA Journal. 10 Rome. F. 6. 2001. E.. Enterobacter sakazakii (Cronobacter spp. J. Enterobacter sakazakii: A new species of "Enterobacteriaceae" isolated from clinical specimens. W. Heuvelink. 287:2204-2205. Braig. Epidemiology of invasive neonatal Cronobacter (Enterobacter sakazakii) infections. 104:1-34. J. M. and M. M. Lehane. J. Gurtler. M. A. 2002. Eur. 2001. Enterobacter sakazakii and other Micro-organisms in powdered infant formula: meeting report. 26. Joosten. Bacteriol. 116:1-10. Hamilton. ICMSF. Microbiological Risk Assessment Series No. Ahmed. and H. A. J. 444:1-14. 68:64-69. Sonnard. 2009. 30:569-584. Enterobacter sakazakii and Salmonella in powdered infant formula: meeting report. Guillaume-Gentil. Friedemann. J. Italy. Clin. 78:646-649. J. Dis. M. M. Enterobacter sakazakii in melkpoeder.. C.Prevalence of Cronobacter spp. 24. Kariuki. D.

Environ. 40. Appl. 2008. Mullane. C. 2004. nov Cronobacter sakazakii subsp sakazakii. 74:2550-2553. M. Forsythe. N. H. S. Cronobacter dublinensis sp nov and Cronobacter genomospecies I.. ISO 21528-1 Microbiology of food and animal feeding stuffs Horizontal methods for the detection and enumeration of Enterobacteriaceae Part 1: Detection and enumeration by MPN technique with pre-enrichment.. 2006. P.. Cronobacter muytjensii sp nov. . Stephan. Fanning. nov. 96:133139. C. Joosten.Detection of Enterobacter sakazakii. Joosten. Cronobacter turicensis sp nov. Forsythe. C. I. C. and H. Appl. nov.. Tal.... Biol.. The taxonomy of Enterobacter sakazakii: proposal of a new genus Cronobacter gen. Int. S. C. E. Cronobacter muytjensii sp. Food Microbiol. R.. ISO. 2007. Microbiol. Druggan. S.. R. J. Cronobacter dublinensis subsp. comb. ISO. 2004. Dublinensis subsp. A. BMC Evol. 2005. 36. 39. Iversen. D. and S. nov. 58:1442-1447. Schumacher. Cronobacter turicensis sp. B. 7:64. Cronobacter malonaticus sp. lausannensis subsp. and Cronobacter dublinensis subsp.. Bidlas. Development of a novel screening method for the isolation of "Cronobacter" spp. 38:378-382. Evol. Iversen. B. Forsythe. ISO/TS 22964 and IDF/RM 210-first edition 2006-02-01: Milk and milk product. C. J. thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant formula milk. McCardel. Cronobacter sakazakii subsp malonaticus subsp nov. Druggan. J.. Cronobacter gen. Cronobacter genomospecies 1. Fanning. N. A. 37. J. ISO 21528-2 Microbiology of food and animal feeding stuffs Horizontal methods for the detection and enumeration of Enterobacteriaceae Part 2: Colony-count method. Iversen. and S. Food Microbiol. Joosten. nov. a new genus to accommodate the biogroups of Enterobacter sakazakii. Syst. and S. Mullane. Gschwend. Stephan. Microbiol. and S. Comparison of media for the isolation of Enterobacter sakazakii. Int. 73:48-52. J. Environ. 33. Lett. ISO. nov. C. J. Iversen. nov. nov. nov. 104 . 38. nov. P. Iversen. and of three subspecies. Microbiol... Forsythe. comb. C.. 32. nov. 2004. Feer. Lehner. 34. and H. a preliminary study. 2008. 2004. The growth profile. and R. Iversen. Cronobacter dublinensis sp. Lane. Cleenwerck. nov. 35. J. Iversen.. lactaridi subsp. 21:771-777.. Cronobacter dublinensis subsp. A.Chapter 5 31.. Lehner. and proposal of Cronobacter sakazakii gen. Lehner. Appl. Isolation of Enterobacter sakazakii and other Enterobacteriaceae from powdered infant formula milk and related products. Stephan.. Microbiol. Marugg. nov and descriptions of Cronobacter sakazakii comb. 2007. A selective differential medium for Enterobacter sakazakii. (Enterobacter sakazakii). K.

2008. 2008. Kuzina. Jung. S. K. J. 42. Curr. A. Reij. R. infants. S.. Chen. P. K. J. L. 17:516519. 122:196-203. 52. A rapid and reliable alternative to ISO 21528-1 : 2004 for detection of Enterobacteriaceae. L. 46. The Lancet. Grognou. and L. Prevalence and genetic diversity of Enterobacter sakazakii in ingredients of infant foods. Stephan. Medicine. Van Puyvelde. 2007... L. and K. J. J. 2006. W. A new protocol for the detection of Enterobacter sakazakii applied to environmental samples. Microbiol. S. Anastrepha ludens (Diptera : Tephritidae). M. and M. Stephan. Leuschner. and R.. Jang. 2008.. Zwietering. V. 67:12671270. Kandhai. Diep. Food Prot. 49. Thelen. 2004. S. G.. R. Jackson. K. Peloquin. Cox. Biotechnol.. 45. and T. Kang. Kilonzo-Nthenge. Microbiol. Y. 50. Hong. L. Prevalence and thermal stability of Enterobacter sakazakii from unprocessed ready-to-eat agricultural products and powdered infant formulas. 125:344-346. C. 48. Kandhai. 2001. Environ. C. F. Beumer. R. Van Schothorst. C. 2006. Enterobacter sakazakii infections among neonates. Lai. Int. H. M. C. H. 44. Rapid detection of Enterobacter sakazakii using TaqMan real-time PCR assay. F. K. K. A. M. J. Occurrence of Listeria and Enterobacteriaceae in domestic refrigerators. and J. Park. Miler. J. Kandhai. Guillaume-Gentil. Vacek. S. W. M. and C. T. Ryu. Appl. B. Marugg. Kim. 2006. M. Kim. 23:409-409.. A. W. Iversen. J. 2004. and adults . Int. E. Klijn. D. Effects of preculturing conditions on lag time and specific growth rate of Enterobacter sakazakii in reconstituted powdered infant formula. Nam. Reij. Microbiol. and S. O. W. Park. Godwin. 42:290-294. Nitzsche. 71:608-612. K. H. A. Gorris. G. M. and M. BMC Microbiol. Sil. M. Food Microbiol. K. A medium for the presumptive detection of Enterobacter sakazakii in infant formula (vol 21. C. Occurrence of Enterobacter sakazakii in the food production environments and households. Comparison of two chromogenic media and evaluation of two molecular based identification systems for Enterobacter sakazakii detection. Lehner. Reij. 363:39-40. Gorris. Isolation and identification of bacteria associated with adult laboratory Mexican fruit flies. J. Donald. Van Schothorst. Baird. K. Heu. 72:2721-2729. J.Case reports and a review of the literature. Food Microbiol. 2006. 2001. Food Prot. G. 105 . B. M. 51. Joosten. R. Food Sci. H. M. J. Food Microbiol. children.. Biotechnol. 41. Van Schothorst. Breeuwer. M. 80:113-122. and S. and M. O. pg 527. 43. S. 2004). 15:152-157.Prevalence of Cronobacter spp. 47. Guillaume-Gentil. C. 6:15..

Chapter 5

53.

Mramba, F., A. Broce, and L. Zurek. 2006. Isolation of Enterobacter sakazakii from stable flies, Stomoxys calcitrans L. (Diptera : Muscidae). J. Food Prot. 69:671-673. Mullane, N., B. Healy, J. Meade, P. Whyte, P. G. Wall, and S. Fanning. 2008. Dissemination of Cronobacter spp. (Enterobacter sakazakii) in a powdered milk protein manufacturing facility. Appl. Environ. Microbiol. 74:5913-5917. Mullane, N., C. Iversen, B. Healy, C. Walsh, P. Whyte, P. G. Wall, T. Quinn, and S. Fanning. 2007. Enterobacter sakazakii: an emerging bacterial pathogen with implications for infant health. Minerva Pediatr. 59:137-148. Mullane, N. R., D. Drudy, P. Whyte, M. O'Mahony, A. G. M. Scannell, P. G. Wall, and S. Fanning. 2006. Enterobacter sakazakii: biological properties and significance in dried infant milk formula (IMF) powder. Int. J. Dairy Technol. 59:102-111. Mullane, N. R., J. Murray, D. Drudy, N. Prentice, P. Whyte, P. G. Wall, A. Parton, and S. Fanning. 2006. Detection of Enterobacter sakazakii in dried infant milk formula by cationic-magnetic-bead capture. Appl. Environ. Microbiol. 72:63256330. Mullane, N. R., P. Whyte, P. G. Wall, T. Quinn, and S. Fanning. 2007. Application of pulsed-field gel electrophoresis to characterise and trace the prevalence of Enterobacter sakazakii in an infant formula processing facility. Int. J. Food Microbiol. 116:73-81. Muytjens, H. L., H. Roelofs-Willemse, and G. H. J. Jaspar. 1988. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. J. Clin. Microbiol. 26:743-746. Muytjens, H. L., J. Van der Ros-van de Repe, and H. A. Van Druten. 1984. Enzymatic profiles of Enterobacter sakazakii and related species with special reference to the alpha-glucosidase reaction and reproducibility of the test system. J. Clin. Microbiol. 20:684-686. Muytjens, H. L., H. C. Zanen, H. J. Sonderkamp, L. A. Kollee, I. K. Wachsmuth, and J. J. Farmer, 3rd. 1983. Analysis of eight cases of neonatal meningitis and sepsis due to Enterobacter sakazakii. J. Clin. Microbiol. 18:115-120. Nazarowec-White, M., and J. M. Farber. 1997. Incidence, survival, and growth of Enterobacter sakazakii in infant formula. J. Food Prot. 60:226-230. Noriega, F. R., Kotloff, K.L., Martin, M.A., Schwalbe, R.S. 1990. Nosocomial bacteremia caused by Enterobacter sakazakii and Leuconostoc mesenteroides resulting from extrinsic contamination of infant formula. Pediatr. Infect. Dis. J. 9:447-449. O'Brien, S., B. Healy, C. Negredo, S. Fanning, and C. Iversen. 2009. Evaluation of a new one step enrichment in conjunction with a chromogenic medium for the

54.

55.

56.

57.

58.

59.

60.

61.

62. 63.

64.

106

Prevalence of Cronobacter spp.

detection of Cronobacter spp. (Enterobacter sakazakii) in powdered infant formula. J. Food Prot. 72:1472 -1475. 65. Oh, S. W., H. J. Chung, and D. H. Kang. 2006. Comparison of enrichment broths for detection and isolation of Enterobacter sakazakii from infant formula milk. J. Rapid Meth. Autom. Microbiol. 14:325-336. Raghav, M., and P. K. Aggarwal. 2007. Isolation and characterization of Enterobacter sakazakii from milk foods and environment. Milchw. Milk Sci. Int. 62:266-269. Reij, M. W., and M. H. Zwietering. 2008. Integrating concepts: a case study using Enterobacter sakazakii in infant formula. p. 177 - 204. In: D.W. Schaffner (ed.), Microbial Risk Analysis of Foods ASM Press, Washington DC, New Brunswick, New Jersey. Restaino, L., E. W. Frampton, W. C. Lionberg, and R. J. Becker. 2006. A chromogenic plating medium for the isolation and identification of Enterobacter sakazakii from foods, food ingredients, and environmental sources. J. Food Prot. 69:315-322. Schmid, M., C. Iversen, I. Gontia, R. Stephan, A. Hofmann, A. Hartmann, B. Jha, L. Eberl, K. Riedel, and A. Lehner. 2009. Evidence for a plant-associated natural habitat for Cronobacter spp.. Res. Microbiol. 160:608-614. Seo, K. H., and R. E. Brackett. 2005. Rapid, specific detection of Enterobacter sakazakii in infant formula using a real-time PCR assay. J. Food Prot. 68:59-63. Shaker, R., T. Osaili, W. Al-Omary, Z. Jaradat, and M. Al-Zuby. 2007. Isolation of Enterobacter sakazakii and other Enterobacter spp. from food and food production environments. Food Control. 18:1241-1245. Simmons, B. P., M. S. Gelfand, M. Haas, L. Metts, and J. Ferguson. 1989. Enterobacter sakazakii infections in neonates associated with intrinsic contamination of a powdered infant formula. Infect. Control Hosp. Epidemiol. 10:398-401. Urmenyi, A. M. C., and M. D. Berne. 1961. Neonatal death from pigmented coliform infection. The Lancet. 1:313-315. Yoo, M. K., S. S. Kim, and S. Oh. 2005. Isolation and genotyping of Enterobacter sakazakii from powdered infant formula manufactured in Korea. J. Food Sci. Biotechnol. 14:875-877.

66.

67.

68.

69.

70. 71.

72.

73. 74.

107

Chapter 5

108

Lag time and specific growth rate

109

In this publication several determinants of the growth of Cronobacter spp. M. product contamination at manufacture. can be present. 72:4.7 h at 10 ºC to 1. and also growth after reconstitution has to be minimised using appropriate control measures.73 ± 0. whereas the optimal specific growth rate was 2.W.Chapter 6 Abstract Cronobacter spp.6 ºC and 47. This Chapter has been published as “Effect of pre-culturing conditions on lag time and specific growth rate of Enterobacter sakazakii* in reconstituted powdered infant formula” Kandhai. The following key growth parameters were determined: lag time. was able to grow at temperatures between 8 and 47 ºC.H. In order to prevent illness. Applied and Environmental Microbiology. The resulting minimum – and maximum temperature estimated with the Secondary Rosso equation were 3. specific growth rate and maximum population density. in dry powdered infant formulae and it has been linked to cases of meningitis in neonates. and spiked into powdered infant formula. during preparation. Zwietering. Strains did not have different lag times and lag times were short taking into account that the cells had spent several (3-10) days in dry powdered infant formula. 110 ..M. M.6 ºC. respectively. The estimated optimal growth temperature was 39. Reij.31 h-1. Cells were harvested at different phases of growth. in reconstituted infant formula are reported.G. 2006. 2721-2729.C. Van Schothorst. L. * Throughout this thesis the designation Cronobacter spp. has been used to indicate the taxonomic change. C. The effect of temperature on the specific growth rate was described using two secondary growth models. and M. especially those born prematurely. Cronobacter spp. Grognou. M. although in low levels. The estimated lag time varied from 83. Cells harvested at different phases of growth did not exhibit any significant difference in either specific growth rate or lag time.3 ± 18.4 ºC. After reconstitution in sterile water. The growth rates and lag times at various temperatures obtained in this study may help in calculation of the period for which reconstituted infant formula can be stored at a specific temperature without detrimental impact on health.43 h at 37 ºC and could be described with the hyperbolic model and reciprocal square root relation. Gorris.

from samples of commercially available dry powdered infant formulae has been reported (5. is a motile peritrichous. When present in dry formula Cronobacter spp.e. Should lag times be apparent before growth. contamination may also occur during reconstitution of the dried infant formula in hospitals or at home. Growth models can simulate growth 111 . Although the levels of Cronobacter spp. cells. with mortality rates of 40 to 80% (4). reconstituted infant formulae offer rich growth environments that allow immediate proliferation provided that the cells are in a sound physiological state. The recovery of Cronobacter spp. sepsis and necrotizing enterocolitis in prematures and full term infants. reconstituted infant formula is a good medium for growth. cells from which they may gradually recover. 10). may grow during preparation. occurring in dry powdered infant formulae are generally very low. has been detected in various other dry environments (8). However. in infant formulae have been associated with outbreaks of meningitis. In the case of Cronobacter spp. 9. Occasional contamination of the dried infant formula during manufacture is a source of the micro-organism occurring in reconstituted product. i. cooling. product contamination at manufacture and/or during preparation and growth after reconstitution has to be minimised using appropriate control measures. the product’s (intrinsic) conditions and the environmental (extrinsic) conditions. microbial cells will either start to grow immediately or show a distinct phase of no apparent growth (the lag phase). which influence their physiological state. and the suitability of the product to sustain their growth. that the external conditions (mainly temperature) are favourable and that there is sufficient time for growth. increasing the probability of illness. then this may be a result of an injured state of the Cronobacter spp. particularly those with predisposing medical conditions (3). as is evidenced by the start of cell proliferation (17). as Cronobacter spp. Gram-negative rod that occasionally causes neonatal meningitis and sepsis. Mathematical models are useful tools to evaluate the effectiveness of control measures. Depending on the source and the history of contaminating bacterial cells. Baranyi & Roberts (1) emphasized that the lag time is a period of adjustments to the new environment during which only intracellular conditions change. Cronobacter spp. storage and holding of the bottles. In order to prevent illness.Lag time and specific growth rate Introduction Cronobacter spp.

Both isolates were obtained from Dr. the lag time data were fitted. Stock cultures were maintained at –20 ºC in cryo vials (Greiner Bio-one GmbH. Strain 94 (S94) was isolated from a sample of a vacuum cleaner bag obtained from a domestic environment (8). is required. 112 . To develop growth models. Radboud. insight into parameters describing growth of the microorganism. in reconstituted infant formula and aid in the design of effective control measures to reduce exposure of susceptible consumers in both hospital and household settings. Frickenhausen.. The Netherlands. During the secondary modelling step the Square root Ratkowsky model (13) and the Secondary Rosso model (14) were fitted to the estimates of the specific growth rates at various temperatures. Nijmegen.Chapter 6 after reconstitution and the effect of key intrinsic or extrinsic conditions can be determined. containing 0. Germany). This study describes the effects of a number of pre-culturing conditions on key growth parameters for Cronobacter spp. Cronobacter sakazakii ATCC 29544 (4) was the reference strain. Furthermore. Likewise. growing in reconstituted (with sterile water) powdered infant formula. GmbH CH9471 Buchs. Becton Dickinson and Co.7 ml of a stationary-phase culture suspension in Brain Heart Infusion (BHI. H. France) broth with 0. the effects of temperature on specific growth rate and lag time were quantified and compared with literature values. Strain MM9 was isolated from milk powder and strain MC10 was isolated from a patient.L. University Medical Centre (UMC) St. The resulting parameters permit useful predictions of the growth of Cronobacter spp. such as the lag time and specific growth rate. Viable counts were used to construct growth curves that were used to derive the key growth parameters by curve fitting using the Modified Gompertz equation as primary growth model (20).3 ml 87% glycerol (Fluka-Chemica. Muytjens. 38800 Le Pont de Claix. using the logarithm of the inverse of the Ratkowsky model and the hyperbolic equation (19). Materials and methods Organisms Four Cronobacter isolates were used in this study.

6. Spiking of the infant formula Infant formula was bought in local shops.000 x g. Characterization of these isolates together with 70 other Cronobacter isolates. MM9 and S94 were incubated at 37 ºC for 18 hours to obtain middle stationary cells. these BHI grown cultures were centrifuged for 5 minutes at 20 C at 2958 x g (MSE. The obtained bacterial suspension was sprayed on commercial dry infant formula (1:50 (w/w)). Cells harvested were transferred into 30 ml BHI and incubated for 1. Specific groups could not be distinguished from the dendrograms (results not shown).Lag time and specific growth rate Switzerland). using a perfume sprayer (Designed by Gérard Brinard. The Netherlands). 1 ml of a stock culture of ATCC 29544 from the –20 ºC was diluted in 2 ml glycerol. In order to obtain enough cells in the lag phase. resulting in a cell suspension of about 104 CFU/ml for lag phase grown cells and between 105 and 107 CFU/ml for the other growth phases. Leusden.5 hour at 37 ºC. Hermle GmbH & Co. Mistral 3000i.. D-7209 Gosheim. stationary phase and late stationary phase cells. In order to obtain cells for spiking dry infant formula. The final bacterial concentration 3-4 days after 113 . Strains MC10. early stationary phase. Cells were incubated for 3. Preparation bacterial suspension Strain ATCC 29544 was cultured by transferring 250 µl of the stock culture into 250 ml BHI followed by incubation at 37 ºC. middle stationary phase. whereupon the suspension was centrifuged (Hermle Z 231M. 24 and 72 hours to obtain respectively the exponential phase. using the pulsed field electrophoresis technique (15) revealed a large variability in molecular fingerprints. 16. Leicester. Germany) for 10 minutes at 10. the numbers of the viable bacteria in the infant formula were sufficiently low to prevent them from influencing the growth of the spiked cells (data not shown). United Kingdom). DA Drogisterij. Cells were washed twice in 1% physiological salt solution and subsequently suspended in 30 ml 1% physiological salt solution for the lag phase cells or 250 ml 1% physiological salt solution. B.

5 ºC 47 ºC 37 ºC 2. in reconstituted infant formula: initial estimates based on published data Parameter Tmin Tmax Topt µopt k Initial estimate 5. Table 1. Here Nt is the number of micro-organisms [CFU/ml] at time t [h]. N0 is the number of micro-organisms at time of inoculation. highest temperature supporting growth (Tmax) and the specific growth rate (µopt) at optimal growth temperature (Topt). The lag time at each temperature was initially estimated by the k value.Chapter 6 spiking was 102-104 CFU/g of dry powder and the infant formula maintained a water activity of aw < 0.5 h 3.22.75 -1 Reference (s) (10) (4. Parameters for growth of Cronobacter spp. t is the time [h]. which is the product of the specific growth rate (µm) and lag time () and which is known to have a large variability. but can be considered constant over a wide range of temperatures (18). Design The initial estimates of the lowest temperature supporting growth (Tmin). the number of micro-organisms in each sample and at every sampling time was roughly predicted using the exponential growth function (N(t) = N(0) e (µm*t) and the secondary Rosso equation (Equation 3). µm is the specific growth rate [h-1]. 6) (6) (6) (7) In order to determine appropriate sampling times and dilutions for plate counting. All growth experiments were performed within 10 days after spiking the infant formula. were used to predict the specific growth rates at various temperatures. shown in Table 1. 114 . using the design values as shown in Table 1.

[ h-1]) and the asymptotic value (A. England) with a spiral plater (Eddy Jet. specific growth rate (µm. Hampshire. ten-gram portions of spiked infant formula were reconstituted in 100 ml sterilised tap water. England). 49. The three kinetic parameters. Leerdam. namely lag time. Data analysis For describing the evolution of the microbial count with time. [-]) (20) at the tested temperature for each growth curve. In the experiments to determine the effect of various growth phases. depending on the temperature of incubation. IUL instruments. 39. Basingstoke. 47. Middle stationary phase grown cells of strain ATCC 29544 were additionally used to assess (in duplicate) growth in reconstituted infant formula at the following temperatures (ºC): 8. specific growth rate and maximum population density were estimated by fitting with the Modified Gompertz equation (Equation 1). 48. 37 ºC (n=14).K. 50. 29 ºC (n=14).. a primary model was used. Basingstoke. MC10 and S94 were incubated only at 29 and 37 ºC. Samples were surface plated onto tryptone soy agar (TSA. I. 46.Lag time and specific growth rate Growth experiments To prepare samples for growth experiments. This resulted in estimates for the lag time (. 45. 14. Inoculated plates were incubated for 20 to 24 h at 37 ºC before manual counting. was measured at various time intervals. 43. appropriate dilutions were made in Peptone Saline Solution (NaCl 8. bottles with strain ATCC 29544 were incubated as follows (the number of bottles incubated is given in parentheses: 10 ºC (n=9). The Netherlands).S.5 g/L and neutralised Bacteriological Peptone 1g/L from Oxoid. 21 ºC (n=11). [ h]).  Nt    me   Ln  Aexp  exp  A   t  1        N0    (1) 115 . Middle stationary phase grown cells of strains MM9. 38. Ltd.. 41. Growth of Cronobacter spp. bv.

and c [ºC-1] are so-called Ratkowsky parameters (13). The secondary growth model of Rosso et al. This model contains four parameters of which two are easily interpretable.5].Chapter 6 In order to obtain reliable estimates for the growth parameters. and b [ºC-1 h-0. 116 . Tmax and Topt. there should be at least 3 data points over a range of 3-hours and 3-logs. Growth curves that did not meet these requirements were excluded. A Bélehrádek type model (Equation 2). Tmin. 2) for data points within the exponential phase. If Tmin  T  Tmax then  m T (2)   bT  Tmin 1  expcT  Tmax 2 T  Tmin T  Tmax then  m  0 and if where Tmin is the extrapolated minimum temperature [ºC] at which the specific growth rate (µm.[h-1]) is zero. experimental growth curves had to meet the following requirements: 1) at least 2 data points should fall within the lag-time.. was used to describe the relation between the specific growth rate and the temperature. Tmin and Tmax. unless the lag time was shorter than 2 hours. also known as the (expanded) square root model of Ratkowsky (13). Tmax is the extrapolated maximum temperature at which µm = 0. (14) (Equation 3). 3) 3 data points should be in the stationary phase at least 1-hour apart. This model contains all four interpretable parameters µopt. was used as well to describe the effect of temperature on growth rate.

describing the specific growth rate. The logarithm of the inverse of the secondary Ratkowsky model (Equation 4) and the hyperbolic equation (Equation 5) were used (19) to describe the lag time temperature relation.5]. The Tmin and Tmax values were assumed to be equal to the Tmin and Tmax of equation 3 (Secondary Rosso growth model). 117 . The q value is comparable to Tmin.Lag time and specific growth rate If Tmin  T  Tmax then (3)   m T      Topt  Tmin T  Topt  Topt  Tmax Topt  Tmin  2T   T  Tmax T  Tmin  2 Topt  Tmin     opt  and if T  Tmin T  Tmax then  m  0 where Tmin and Tmax are defined in equation 2. Topt [ºC] is the temperature at which the specific growth rate µm [h-1] is optimal. ln    p T q (5) where p is a measure for the decrease of the lag time when the temperature is increased and q is the temperature at which the lag time in infinite (no growth). ln  T     lnbT-Tmin 1  expcT  Tmax 2  (4) where b [ºC-1 h-0. and c [ºC-1] are the so-called Ratkowsky parameters. and µopt is the µm at optimal temperature.

29 and 37 °C were subjected to univariate analysis of variance. and 37 °C was predicted based on published growth parameters (see Table 1. 118 . A significance level of 5% was used. All data analyses were performed using SPSS (SPSS release 11. furthermore. Results Determination of growth parameters. specific growth rate. using pre-cultured middle stationary ATCC 29544 cells In order to design the experiments optimally and to obtain growth curves meeting the specified requirements. respectively. at 10. Standard deviations were calculated with Excel and are reported as plus or minus the mean. the course of each individual growth curve of Cronobacter spp.. contaminated infant formula. Lag time data were log transformed and specific growth rate data were square root transformed in order to obtain homogeneity of variance. The modified Gompertz equation (Equation 1) was fitted to the observed number of micro-organisms in time.03. for verification. the data obtained with strain ATCC 29544 at 10. the resulting experimental count data of growth in reconstituted. These figures show.S. 29. for initial estimates). For the growth curves at 10 ºC and 29 ºC comparable results were obtained (data not shown). Variations in the initial number of micro-organisms at time zero may have been due to a gradual decline of the cell number in the dry infant formula. SPSS Inc. 21.5 for Microsoft Windows 95/98/NT/2000. 21. windows 2. Examples of the design growth curve at 21 ºC and 37 ºC are shown in Figures 1A and 1B. resulting in estimates for the lag time. Chicago. asymptote and the initial number of organisms (19).A). Fitting was done by minimizing the residual sum of squares (RSS) with both the solver function in Excel and Table curve 2D. U.Chapter 6 Statistical analysis In order to determine whether pre-culturing conditions have a significant effect on lag times and/ or specific growth rates.

Dotted lines represent fits of the Modified Gompertz equation to each single experiment.Lag time and specific growth rate Figure 1. The straight line represents the design growth curve at 21 ºC. The different symbols indicate different replicate experiments. 119 . pre-cultured to middle stationary state. using the cardinal values as shown in Table 1 under “initial estimate”. Predicted and measured growth curves at 21 ºC (A) and 37 ºC (B) of Cronobacter sakazakii ATCC 29544.

Figure 2A. 29 and 37 ºC to represent a temperature range relevant for reconstituted infant formulae.04 h-1 at 10 ºC to 2. Growth was observed in all cases and at every temperature. and no apparent effect of the physiological state was found. : early stationary phase cells.Chapter 6 Effects of the physiological growth phase on µm and  Six different pre-cultures of the reference strain. ATCC 29544 were prepared to yield a variety of physiological growth phases at the moment of spiking of the powdered infant formula. After reconstitution with sterile tap water.12 ± 0.  OJ SK V J U Q FHO D D H RZ OV exponential phase cells. 120 . 21. μm strongly increased. Square root of specific growth rate data for various physiological growth states of strain ATCC 29544 as function of temperature. Values for specific growth rates varied from 0. all six types were incubated with various replications at 10. o: stationary phase cells.29 ± 0. With increasing temperature. Temperature had a marked effect on both the specific growth rate (μm) and the lag phase (λ) (Figures 2A and 2B). : middle stationary phase cells.45 h-1 at 37 ºC. : late stationary phase cells. In Figure 2B it is shown that the lag time decreased with increasing temperature and that there was also no apparent effect of the various physiological growth phases on the lag time.

Log lag time for various physiological growth states of ATCC 29544 as a function of the temperature. △: exponential phase cells. statistical analysis showed that the k-value (product of λ and μm ) was also not significantly influenced by the cell history (Table 2).Lag time and specific growth rate Figure 2B. As depicted in Table 2. o: stationary phase cells. middle stationary phase and late stationary phase) of strain ATCC 29544 as shown in Figures 2A and 2B were analysed with the univariate analysis of variance test. this value can be considered as borderline significance. : middle stationary phase cells. However. : early stationary phase cells. exponential phase. this resulted in p-values > 0. : late stationary phase cells. This test corroborates the visual observation (Figures 2A and 2B) that the cell history had no significant effect on either the specific growth rate or on the lag time during subsequent cultivation in reconstituted infant formula. 121 . : lag phase cells. except for the lag time at 21 ºC (p=0.05. Furthermore. early stationary.048). Estimates of the specific growth rates and lag times for the various pre-culturing conditions (lag phase.

295 0. It is shown in Figures 3 and 4 that neither the lag time nor the maximum specific growth rate was significantly different for strains of different origin. 122 .629 0. 29 and 37 °C.254 0. 21 29 and 37 °C did not significantly influence the specific growth rates and lag times. Statistical evaluation (p-values) of the univariate analysis of variance for the effects of the physiological growth phase of Cronobacter sakazakii ATCC 29544 cells on lag time. at 29 and 37 °C.408 Value at: 29 °C 0.461 0. where the square root of the specific growth rate (√µm) is plotted as function of the temperature. no reliable estimates of the specific growth rate and lag time could be derived from the models at that temperature as the growth curves did not meet the requirements as formulated in the materials and method section. pre-cultured to middle stationary phase.314 0.05 meaning a significant effect of the physiological state on the variable Parameter 10 °C Lag time () [h] Specific growth rate (µm) [h-1] k-value ( x m) [-] 0.048 0. MM9 and S94.166 0. Although growth was observed up to 47 ºC.Bold value indicates p-value < 0.539 37 °C 0. from 8 up to 50 ºC with middle stationary pre-cultured cells of strain ATCC 29544 only.Chapter 6 Effects of the strain variability on µm and  The effect of strain variability on the growth parameters was studied by growing three other strains MC10.512 Estimations of growth parameters Since statistical analysis showed that the physiological growth phases of strain ATCC 29544 at 10.407 21 °C 0. 21. All of the estimated specific growth rates were combined and modelled with equation 2 and equation 3 as shown in Figure 3. The secondary growth parameters derived from optimal fits are shown in Table 3 and 4. further experiments were performed over a wide range of temperatures.092 0. Table 2. specific growth rate and the product of both (k) at 10.

13 .1.28 95% Confidence interval -0.2.7 2.31 95% Confidence interval 1. in reconstituted infant formula resulting from fits by the Rosso secondary growth (Equation 3) Parameter Tmin Tmax Topt µopt RSS Estimate 3.9 0. in reconstituted infant formula resulting from fits by the Ratkowsky Secondary growth model (Equation 2) Parameter Tmin Tmax b c RSS Estimate 2. For comparison the experimental specific growth data 123 .5) at 37 ºC.0407 .48.9 .115 h-1 ( μm =0.0521 0.31 1.19 48.28. it can be seen that the specific growth rate of Cronobacter spp. and 2.1 .4.5) at 46 oC.40.4 2. Although the RSS was slightly higher.113 h-1 ( μm = 1.047 0.60 47.5 ºC-1 Table 4. The residual sum of squares (RSS) for the secondary Rosso model (Equation 3) was 1. the secondary Rosso model was used for further evaluation since it consists of four parameters that all have a biological meaning and can be interpreted as such. Parameter values for the effect of the temperature on the specific growth rate (µm) for Cronobacter spp.5) at 10 oC to 1.79 47.0.242 h-1 ( μm = 1.498 h-0.144 .339 h-0.239 1.9 0. varied from 0. Transforming the square root μm data from Figure 3 back to μm.2 38. Parameter values for the effect of the temperature on the specific growth rate (µm) for Cronobacter spp.5.0.49.31 and for the square root Ratkowsky model (Equation 2) was 1.063 h-0.Lag time and specific growth rate Table 3.64 47.26 .42 .335 Dimensions ºC ºC ºC-1 h-0.6 39.48 Dimensions ºC ºC ºC h-1 From Figure 3 it is apparent that the differences between the fits of these models are smaller than the experimental variability.

The effect of the temperature on the lag time was described by fitting the reciprocal square root relation (Equation 4) and the hyperbolic model (Equation 5) to the logarithmic transformation of the data.. Parameters estimated with both models are given in Table 5. +: growth rates published by Nazarowec-White et al. (6) and Nazarowec-White et al. ○: growth rates of ATCC 29544 pre-cultures to various growth states as estimated by the fit of the Modified Gompertz model to each individual growth curve. Square root of measured and fitted specific growth rates as function of the temperature. (9).. 124 . The estimated lag times as resulting from the fit by the modified Gompertz model were plotted against the temperature and presented in Figure 4. (10) have been transformed and included in Figure 3.3 ± 18. ♦= S94 and ▲= MC10.Ratkowsky and : Rosso (fitted to square root transformed data of the current study only)... with an estimated minimal lag time of 1.. Figure 3. Also for middle stationary phase grown cells of ● = MM9.73 ± 0.Chapter 6 measured by Iversen et al. (7). fits by secondary growth model of . The temperature at which the lag time was minimal was between 37 and 39 ºC..7 h at temperatures around 10 ºC.. The longest lag time observed was 83.44 h. ◊: growth rates published by Iversen et al. It is assumed that the Tmin and Tmax value are equal to the Tmin and Tmax value of Equation 3 describing the specific growth rate.

♦ = S94 and ▲= MC10. modelled with the hyperbola model (----------) and reciprocal Ratkowsky model (_________). Log lag time as function of temperature.11 95% Confidence interval -a -a 0. + indicates the lag times at 10 and 23 oC as published by Nazarowec-White et al.62 Hyperbolic (Equation 5) a – fixed (from Table 4) by using the values estimated with the secondary Rosso model for growth (Equation 3). Logarithmic transformed lag time data of current study.60 47.29.275 .Lag time and specific growth rate Table 5.8 3. 125 . ○ = ATCC 29544 pre-cultures to various growth states.29 25. Parameter values for the effects of temperature on the lag time () for Cronobacter spp.75 2.88 .6 0.024 0.1.4.3 2. in reconstituted infant formula Model Inverse Ratkowsky (Equation 4) Parameter Tmin Tmax b c RSS p q RSS Estimate 3.645 2..(10). Figure 4.0.3 .176 22. Also for middle stationary phase grown cells of ● = MM9.021 .023 0.

126 . : middle stationary phase grown cells.37 [-].7.11 and for the reciprocal square root relation 2.  OJ SK V J U Q FHO D D H RZ OV H xponential phase grown cells. At temperatures ranging from 20 to 46 ºC values for k were between 0.06  4. as function of temperature is shown. The dotted line represents the average value for the data from 20 up to 46 ºC. Below 20 ºC it increased to an average value of 10. The residual sum of squares (RSS) for the hyperbolic model was 2.08 to predict the lag time at any temperature and equation 3 to determine the specific growth rate resulted in a RSS value of 16. The k-value between 8 and 47 °C was 5.49 [-].05  1.Chapter 6 The logarithmic transformed lag times as fitted by the reciprocal square root relation and the hyperbolic model are represented in Figure 4. the parameter k.82 and 11. the product of the lag time and the specific growth rate (µm x ).29.92 [-]. with an average of 4. Using the k-value of 5. o: stationary phase grown cells.6 [-]. : early stationary phase grown cells. Based on the RSS values for the models describing the lag time it can be concluded that the hyperbolic model and reciprocal square root model best convey the experimental lag times. : late stationary phase grown cells. From this graph it can be concluded that both models fits the data reasonably well. Figure 5.08  3. In Figure 5. Parameter k (product of  and m) as function of the temperature for each growth experiment of strain ATCC 29544 pre-cultured to various growth phases and grown at temperatures from 8 up to 47 °C.

which varied between ± 102 and 105 CFU/ml did not seem to affect the maximum population density. ∆: maximum population density. pre-cultured to various physiological states and its maximum population density at various temperatures. since it has the ability of increasing the lag time at higher temperatures and it contains more interpretable parameters. Figure 6. Discussion Cronobacter spp. the maximum number of cells reached was ±106 CFU/ml. The initial inoculum level. may grow in wet spots and survive in dust containing residues of infant formulae and contaminate the product after the drying process. may contaminate infant formulae either during production or during bottle preparation. At temperatures tested above 45 ºC. In all situations 127 .Lag time and specific growth rate The reciprocal square root model is recommended. In factory environments Cronobacter spp.2 CFU/ml (Figure 6). o: initial number of cells. The maximum number of cells of strain ATCC 29544 reached at the various incubation temperatures between 8 and 46 ºC varied between approximately 107 and 109 CFU/ml. The initial inoculum of strain ATCC 29544. In hospital and household kitchens infant formulae may be contaminated likewise with dry or wet residues by utensils and via environmental vectors. with an average of 108. The dotted line represents the average value for the maximum population density data from 8 up to 46 ºC.

. (2004) (6. the plate count technique was used and growth was measured in a more representative range and over multiple logs of bacterial counts.Chapter 6 the physiological state of the contaminating cells is not known. A possible explanation might be that the cells were spiked into dry powdered infant formula prior to the growth experiments. Specific growth rates reported by Iversen et al.. In our study. (1997) who reported similar specific growth rates. 10) as shown in Figure 3.e.71 ± 0. In our study the specific growth rates of Cronobacter spp. 16). In this medium Cronobacter spp. Cells that had been in the powdered infant formula for up to 4 weeks did not show a longer lag phase either (data not shown). no effect on either parameter was found. cells had a rather short lag time of 1. One cause for the difference with the latter study may be due to differences in growth conditions: in our experiments spiked dry powdered infant formula samples reconstituted with sterile tap water were used as growth medium. whereas Iversen et al. In the current study. spiked into dry infant formula is comparable to Nazarowec-White et al. measured growth in other media by inoculating with an overnight TSB culture (6). which measures growth only at much higher levels (> 106 CFU/ml) than the plate count method used in our study (16). Another explanation for not measuring differences in resuscitation times might be the rich growth medium (reconstituted powdered infant formula) used. It was envisaged that the time of pre-culturing. But it is known that the duration of the lag time may depend not only on the growth environment but also on the previous history of cells (12. Another difference is the use of the rapid automated bacterial impedance technique.. and their metabolic activity had possibly changed in all cases to a comparable level. i. thus the estimates for the lag time and specific growth rate can be considered more accurate (2).. Our specific growth rate data were compared to specific growth rates published by Nazarowec-White et al. the physiological growth phase of the inoculum. (1997) and Iversen et al. The period (3 to 10 days) the cells were in the powdered infant formulae before carrying out a growth experiment was found not to influence the growth parameters. based on visual observation of growth curves and on statistical analysis of growth data. could have an effect on the lag time but probably not on the specific growth rate (16). 128 .50 h at 37 °C and small differences in lag phase may not become apparent. (2004) were consistently over 10% lower..

but not all strains appeared to be pathogenic (11). the secondary Rosso model (Equation 3) and the reciprocal square root model (Equation 4) successfully estimated the lag time and specific growth rate for the whole growth temperature range (Table 4 and 5). 10) in combination with the relatively short lag time and high specific growth rates found in this study..5 ºC and 47 ºC (4). respectively. The modified Gompertz model (Equation 1). The theoretical minimal-. growth of Cronobacter spp. allowed prediction of the growth with enough accuracy to determine the appropriate sampling times and dilutions to measure growth curves.L. Acknowledgements The authors wish to thank Ingrid Maas for skilful technical assistance.6 ºC (Equation 3).Lag time and specific growth rate The experimental design. not all the Cronobacter spp. Muytjens is thanked for providing the Cronobacter spp. such that our quality requirements were met for 95% of the experimental growth curves. Nevertheless the results described in this paper (Table 4 and 5) allow to predict growth of Cronobacter spp. 9. strains MM9 and MC10. The fact that low numbers of Cronobacter spp. showed that a number of 105 CFU/ml of certain Cronobacter spp. is gratefully acknowledged.9 ºC (Equation 2) and 3. (2003). and maximum growth temperatures as fitted with both the secondary growth models are 2. Switzerland. Dr. in reconstituted infant formulae under conditions that closely mimic the conditions in the actual food product in both hospital and household settings and will be useful in designing effective control measures. The authors thank Ingrid Hombergen and Etienne Vidal for carrying out part of the experiments. strains necessarily need to be regarded as potential pathogens.19 ºC and 48. 129 . strains could be lethal to suckling mice after ingestion. strains has been observed between 5. however. however. underscores the need for careful preparation and use. Lausanne. A study by Pagotto et al. cells have occurred in dry infant formulae (5. H. The financial support by Nestlé Research Center. based on initial estimates of specific growth rate and lag time. As there is a lack of information about virulence factors.60 ºC and 47. In practice.

Marks. Recommendations for calculating growth parameters by optical density measurements. J. H. J. 6. 2002. Bacteriol. Occurrence of Enterobacter sakazakii in the food production environments and households. Microbiol. J. Begot. J. F. M. Food Prot. A. 73:197-201. J. J. Bacteriol. Food Microbiol.. and J.. 2004. Methods. The Lancet. and A. 1995. 26:199-218. 5. 9. and L. Microbiol. D. M. and M. K. V. J. M. 2. Bidawid. Syst. J. Roelofs-Willemse. Rome. and S. 2003. Relationship between temperature and growth rate of bacterial cultures. and G. J.. Lane. Farmer III. Iversen. Labadie. Guillaume-Gentil.. Juneja. 1982. 30:569-584. and the Enterobacteriaceae study group. Enterobacter sakazakii: infectivity and enterotoxin production in vitro and in vivo. L. M. Ordonez. 38:378-382. Kandhai. Pin. J. 130 . Baranyi. thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant formula milk. Muytjens. Garcia de Fernando. J. H. O. 12. A. A. F.. 26:743-746. Farber. 2004. Roberts. H. J. Microbiol. J. M. Pagotto. Microbiological Risk Assessment Series No. M. 6. J. Reij. 23:199-213.. J. Jaspar. M. Lebert. Gorris. Enterobacter sakazakii: A new species of "Enterobacteriaceae" isolated from clinical specimens. FAO/WHO. 1997. Ratkowsky. J. D.. Van Schothorst. C.de Boer. A. Food Prot. Enterobacter sakazakii in melkpoeder. Mathematics of predictive food microbiology. 66:370-375. W. and growth of Enterobacter sakazakii in infant formula. J. 10. and A. De Ware(n)chemicus. Forsythe. M. The growth profile. 1996. 1980. and J. 11. I. Food Microbiol. 4. Int. G. and J. C. J. Clin.. Baranyi. Ahmed. T. F. M. A. Huang.. A. Int. Hickman. A. M.Chapter 6 References 1. M. D. and T. Farber. Incidence. J. C. Appl. Enterobacter sakazakii and Other Micro-organisms in Powdered Infant Formula: meeting report. survival. Desnier. 3. Italy. W. Lett. E. D. 8. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. Olley. 13. C. D. M. T. H. Asbury. 2003.. J. Int. Ball. Heuvelink. Risk Analysis. Nazarowec White. Growth and heat resistance kinetic variation among various isolates of Salmonella and its application to risk assessment. L. 60:226-230. Kodde.. 2004. 7. 2002. 32:1730.. Brenner. 25:225-232. McMeekin. G. 363:39-40. Daudin. 1988. 149:1-5. Zwartkruis-Nahuis. and E. Nazarowec-White. Analysing the lag-growth rate relationship of Yersinia enterocolitica. C.

A. and W. P. Environ. J. Voss. Food Microbiol. Swinnen. H.Lag time and specific growth rate 14.. M. J. F. and J. M. Academic Press. J. 20.. C. 15. Rosso. F. H. H. C. Zwietering.. Zwietering. Microbiol. Muytjens. 17. G. Microbiol. Rombouts. Microbiol. A. 6:113-115. Van Schothorst. Cuppers. Smeets. A. S. Predictive modelling of microbial lag phase: a review. E. E. 131 . Lobry. 61:610-616. G. M. J. Environ. Ltd. J. K.. C. London. p. Modeling of the bacterial growth curve. and K. B. 1998. 56:1875-1881.). M. 18. Van 't Riet. Microbiol. 1995. and J. Inc. Appl. T. and K. L. J. F. Genetische karakterisatie van Enterobacter sakazakii isolaten van Nederlandse patiënten met neonatale meningitis. Dens. L. I. Zwietering. 317-328. Van 't Riet. 60:195-203. H. de Koos. Melchers. 1990. 19. H. Modeling of bacterial growth as a function of temperature. M. M. Hasenack. Appl.. Jongenburger. and K. Appl. Flandrois. 1991. 2004. 16. A. G. Int. H. 57:1094-1101. R. J. Ned. de Wit.. L. Microbiol. de Wit.. Van Impe. Tijdschr. Convenient model to describe the combined effects of temperature and pH on microbial growth. I. Inhibition and inactivation of vegatative microbes. Geeraerd. Environ. Van 't Riet. 94:137-159. 1976. Med. J. Bajard. M. Environ. Appl. Resuscitation of injured bacteria in foods. In Skinner and Hugo (eds. J. 1994. Bernaerts. (London). M. Meis. Evalution of data transformations and validation of models for the effect of temperature on the bacterial growth.

Chapter 6 132 .

Growth during cooling 133 .

. Placing a container filled with reconstituted liquid formula in the refrigerator. Gorris. A mathematical model was built to predict the growth of Cronobacter spp. under dynamic temperature conditions using three models for the lag time.Chapter 7 Abstract Reconstituted infant formulae are excellent growth media for Cronobacter spp. Breeuwer. Predictions taking into account the large variability in heat transfer showed that proliferation of Cronobacter spp. The overall heat transfer coefficients (α) were determined and were shown to have a very large variability in both household refrigerators and an air-ventilated refrigerator equipped with a fan. This Chapter has been published as “Growth of Cronobacter spp. during the cooling process in three types of containers. This study describes the temperature profiles and methods to predict lag time and possible growth of Cronobacter spp. 134 . under dynamic temperature conditions occurring during cooling of reconstituted powdered infant formula” Kandhai.M. Immediate consumption or rapid cooling and storage at low temperature are therefore recommended as control measures to prevent microbial growth.H. L. or by reconstituting at temperatures of 25 °C or lower. does not mean that the temperature of the liquid is directly the same as the set-point of the refrigerator.G. (formerly Enterobacter sakazakii) and other micro-organisms that may be present in such products. P. The assumption of a constant k-value (k = lag time x specific growth rate = λ x μ = 2.88) fitted the experimental data best. during cooling may be prevented by limiting the volume to be cooled to portion-size only. Zwietering.C. however. and M. 2009. 72: 2489-2498.W. The various estimations for the lag time had a remarkably strong impact on the predicted growth. Reij Journal of Food Protection. M. M. The model may also be used to predict growth in other situations where dynamic temperature conditions exist.

however.Growth during cooling Introduction Cronobacter spp.6 and 47. Although only recently brought to wider attention of infant formulae manufacturers and the public through several outbreaks and public recalls. is a group of opportunistic pathogenic micro-organisms belonging to the family of Enterobacteriaceae (9. (2003). Reported concentrations in dry powdered formulae ranged from 0. A study by Pagotto et al. caregivers sometimes prefer to prepare all bottles needed for one day in advance (26). 16. suggests that a number of 105 CFU/ml of certain Cronobacter spp.59 h (17). reconstituted infant formula is a good medium for growth with only the barriers of short storage and low temperature to prevent bacterial growth. Several investigations showed that the microorganism can be found in a variety of environments and that contamination may occur during manufacturing and during preparation of the bottles (12. are not able to grow. 17. 19.92 CFU/g (7). Cronobacter spp. For practical reasons. 20. However.06.. even if the temperature of the refrigerator is controlled. 13). have been associated with neonatal deaths as early as 1958 (32).002 to 0. To prevent multiplication of micro-organisms most manufacturers prescribe consumption of infant formula directly after preparation.6 ºC (12. After reconstitution Cronobacter spp. strains could cause illness in suckling mice after ingestion (24). until recently known as the species Enterobacter sakazakii. Specifically in hospitals it is not uncommon that infant formulae are prepared once per day and placed in a refrigerator to be used within 24 h (27) or within 30 h with the refrigerator set at 4 °C (1). the filling 135 . the temperature of the reconstituted formula itself is not controlled yet. The cooling rate of the liquid depends on the temperature. At room temperature (25 ºC) the generation time was determined to be 0. Many manufacturers instruct that their powdered infant formulae should be reconstituted with moderately warm (lukewarm) water to assure that the powder dissolves well. The liquid formula then needs time to cool down in the refrigerator. 23). Cases of neonatal meningitis or necrotizing enterocolitis due to Cronobacter infection have in certain cases been associated with powdered infant formulae (4. In dry powdered infant formula Cronobacter spp. but after the addition of water. can multiply at temperatures between 3. 22)..029).

Chapter 7 rate. In the report of the technical meeting on Enterobacter sakazakii and Salmonella in powdered infant formula (8) a risk assessment model has been discussed that estimates the fate of Cronobacter spp. during cooling of reconstituted infant formula in refrigerators and to use the findings to propose control measures that may reduce and prevent the growth of Cronobacter spp.. and the volume of the container to be cooled. used in hospitals and nurseries. The aim of our study was to make a detailed investigation. also referred to as static refrigerators. was bought locally and had a bacterial count of less than 102 CFU/g. but not always. Rosset et al. of the growth opportunities of Cronobacter spp. These variations are mainly related to differences in temperature. tap water was brought to a full boil in a covered water boiler. cooled down on the bench to approximately 45 ºC and mixed with the powdered infant formula to obtain reconstituted formula with a temperature of 40 ± 3 °C. (27) studied temperature profiles in 25 neonatal care units and concluded that cold storage was the operation with the largest impact on potential growth of Cronobacter spp. heat transfer at the container surface is principally due to natural convection by a very limited airflow caused by variations in air density. In accordance with the instructions on the label. To our knowledge. which did not influence the experiments (data not shown). filling of the fridge. and the air velocities in the refrigerator. during the cooling process. investigations into the cooling process of baby bottles are scarce. In air-ventilated refrigerators mechanic ventilation forces air convection which improves heat transfer (18). the geometry of the food container. both experimentally and by modeling. In household refrigerators. Airventilated refrigerators are often.. during various scenarios of preparation and storage. and humidity gradients. 136 . the thermal properties of food and bottle. Materials and methods Preparation of reconstituted infant formula for cooling experiments Powdered infant formula suitable for infants less than 6 months of age.

The three containers were placed.10 m.02 0.21 Dimension kJ kg °C kg m m m m m m m W m °C -1 -1 -3 -1 -1 Reference (3) (30) (30) (2) W m-1 °C-1 Reconstituted infant formula was divided among the measuring beaker and the two types of baby bottles.Growth during cooling Cooling experiments A total of 25 cooling experiments were performed using an experimental set up whereby three different container types with reconstituted infant formulae were placed in the refrigerator and the temperature during cooling was recorded.032 0.115 0.025 0. While the bottles had different geometries.52 0. and was 137 . Eight out of nine household refrigerators were located in households. The ninth household type refrigerator was located in the laboratory. polycarbonate (kcontainerwall) Value 3.05 m.05 0. Table 1. (baby) bottle Type 1 (maximum 260 ml) with a diameter of 0. and (baby) bottle Type 2 (maximum 120 ml) with a diameter of 0.93 1. Parameters used in Equations 2 and 3 Parameter Specific heat capacity of infant formula (cp) Density of infant formula (ρ) Radius of 1-liter measuring beaker (r) Radius of bottle Type 1 (r) Radius of bottle Type 2 (r) Height of 1-liter measuring beaker (ht) Height of bottle Type 1 (ht) Height of bottle Type 2 (ht) Thermal conductivity of infant formula (kliquid) Thermal conductivity of bottle wall. The following three polycarbonate containers were used: a measuring beaker equipped with a lid with a volume of 1. both were filled with 120 ml of formula.000 ml and a diameter of 0.14 0.000 ml. and contained different amounts of other food materials.04 m (Table 1). either an incubator ventilated with air or in one of the 9 different static air refrigerators (further referred to as household refrigerators). in the center of. The measuring beaker held 1.95 0.

cp [J kg-1 °C-1] is the specific heat capacity of the reconstituted infant formula. The internal air temperature in the refrigerator was measured with a thermistor as well.c p dt   (1) where V [m3] is the volume of the reconstituted infant formula in the measuring beaker/bottle.t     2   t  r c p    e (2) where TC. During cooling the temperature was recorded in the center and near the wall of each container every 5 minutes for 24 h using thermistor metal oxide sensors (CM-UU) and a squirrel data logger (both: Eltek Ltd. A TC  TR . The temperature of the household refrigerators varied. t [s] is the time. which was recorded as a given parameter and not set as a controlled experimental parameter.Chapter 7 empty except for the three containers. UK). Values of the parameters are shown in Table 1. TC [ºC] is the temperature of the prepared liquid formula and TR [ºC] is the air temperature in the refrigerator. The temperature measured in the center of the bottles was slightly higher than the temperature measured near the wall and the difference was 0.   .. Cambridge. for a more fail safe approach.0  TR TC . After integration and rewriting this equation transforms into:  TR  TC .5 ºC at most. α[ J m-2 °C-1 s-1 = W m-2 °C-1] is the overall heat transfer coefficient. dTC V . The temperature measured in the center of the bottle was chosen for use in the model predictions.0 [ºC] is the temperature in the center of formula in the measuring beaker or bottle at the beginning of the cooling process and r [m] is the radius of the measuring beaker or bottle. 138 . ρ [kg m-3] is the density of the reconstituted infant formula. A [m2] is the surface area of the object. Cooling model for reconstituted infant formula Equation 1 describes the change in temperature of a bottle or beaker (TC) assuming that the temperature in the liquid is homogeneous and that heat is transferred from the liquid via the wall of the bottle to the air.

the internal heat transfer coefficient can be estimated by the ratio of the thermal conductivity of the reconstituted infant formula (kliquid) and the radius of the container (r). If free convection does not occur. When a fan is present. 139 .Growth during cooling Here. The internal heat transfer coefficient depends on the movement of liquid in the bottle. As shown in equation 3. not the top or the bottom of the container. this coefficient can be expected to be higher than in household refrigerators that are not equipped with a fan. we assume that both the bottles and the measuring beaker can be described as a cylinder and that only the side surface area of the liquid (A = 2·π·r·h) contributes to heat transfer. the coefficient of the wall. Test organism Stock cultures of Cronobacter sakazakii (formerly named Enterobacter sakazakii) ATCC 29544 were maintained at -80 ºC in cryo vials (Greiner Bio-one GmbH.A) broth with 30% (vol/vol) glycerol (Flukachemica. Maryland. Buchs. Frickenhausen. αcontainerwall [the ratio of thermal conductivity of the container material (kcontainerwall) and the thickness of the wall]. U. αproduct. 1  1  1  1 (3)   exernal  containerwall  product The external heat transfer coefficient largely depends on the movement of air in the refrigerator. and the internal heat transfer coefficient. The ratio of surface area and volume (A/V) can then be described as 2 r -1. During cooling of the bottles the liquid may circulate slowly due to free convection. Switzerland). Germany). the overall heat transfer coefficient α consists of the external heat transfer coefficient in air. Difco. containing a stationary-phase culture suspension in Brain Heart Infusion (BHI.S. Becton Dickson. αexternal.

to obtain reconstituted formula with 103 to 105 CFU/ml at a temperature of 37 ± 3 °C. inoculation and sampling The inoculum of strain ATCC 29544 was prepared by transferring 100 µl of the stock culture into 100 ml BHI broth. 21. by inserting a 1-ml disposable pipette with balloon. Leerdam. England) and then plated in duplicate onto Tryptone Soy Agar (TSA. under dynamic temperature conditions A growth model was built that uses the lag time and the specific growth rate of Cronobacter spp. This formula was distributed between the Type 1 bottle and the measuring beaker and placed in the laboratory refrigerators immediately. 23). sprayed onto dry infant formula to obtain a final concentration of 104 to 106 CFU/g and stored for 3 to 10 days as described elsewhere (17). similar to the procedure described at “Preparation of reconstituted infant formula for cooling experiments”.. is 6 ×102 CFU/ml. Samples were taken every 30 min during the first 4 h and subsequently every 60 min. Predicting growth of Cronobacter spp. Oxoid LTD. NaCl 8. Cells were harvested. Basingstoke. Basingstoke. A minimum level required to accurately count the organisms plated by the spiral plater. The specific growth rate (µm) for Cronobacter spp. however. during the constantly changing temperature conditions that occur during cooling. Plates were incubated for 20 to 24 h at 37 ºC and counted manually. leaving the bottles in the refrigerator. Hampshire. samples were first diluted in Peptone Saline Solution (PPS.Chapter 7 Growth medium. The inoculation levels chosen are higher than the very low numbers found in infant formulae (7. If appropriate. followed by incubation for 18 to 20 h at 37 ºC.5 g/L supplemented with 1 g/L neutralized Bacteriological Peptone from Oxoid. England) using a spiral plater device (Eddy Jet. No other products were present in the refrigerators. washed. The powder containing Cronobacter spp. at each temperature (Ti) was 140 . The Netherlands). was then mixed with water. and inoculum levels were chosen to be slightly higher than this minimum. 11.

and µopt [h-1] is the µm [h-1] at this optimal temperature (28). Tmax [C] is the extrapolated maximum temperature. Topt [C] is the temperature at which the specific growth rate is maximal.. Ti [ºC] used is the average temperature of the simulation interval. 141 . 1  exp c Ti  Tmax ln  Ti        2  (5) where b [ºC-1 h-0.  ln b Ti  Tmin . The Tmin and Tmax parameters were determined from previous growth experiments at temperatures ranging from 8 up to 47 ºC (17).Growth during cooling described with the secondary growth model (Equation 4) of Rosso et al. Three models were used to estimate the lag time. lag times were estimated using the inverse of the secondary Ratkowsky model (Equation 5) with parameters previously determined as shown in Table 2 (17). and c [ºC-1] are the so-called Ratkowsky parameters (34). Tmin [C] is the extrapolated minimum temperature. First. If Tmin  T  Tmax then (4)  m Ti        Topt  Tmin T  Topt   Topt  Tmax Topt  Tmin  2T Topt  Tmin  i i T  Tmax Ti  Tmin 2   opt   and if Ti  Tmin Ti  Tmax then  m Ti   0 where Ti is the temperature [ºC] at each moment in the center of the volume of the infant formula in the measuring beaker or the bottle.5]. (28) using cardinal parameters previously established (Table 2) (17).

the logarithmic model was used to estimate lag times (Equation 7) (25).6 47.5 5 5 7 7 Logaritmic model bl cl 142 .6 39. Earlier determined cardinal growth parameters taken from literature. λ [h] is the lag time.141 Dimension °C °C °C h -1 Equation 4 4 4 4 5 5 Reference (17) (17) (17) (17) (17) (17) (17) (17) (25) (25) Inverse of secondary Ratkowsky model Tmin Tmax b c °C h °C-1 - -1 -0.4 2. It is known that the k-value has a large variability but it can be considered constant over a wide range of temperatures (34).309 -1.org/). the lag times were estimated using a constant k-value. k   Ti . log10  Ti     cl lnTi   bl (7) where.Chapter 7 Second. to calculate the specific growth rate (µ [h-1]) and the duration of the lag phase ( [h]) as a function of temperature Model Secondary Rosso growth model Parameter Tmin Tmax Topt µopt Value 3.mramodels. Table 2.9 0.645 4.31 2. which is the product of the specific growth rate and the lag time (Equation 6). bl and cl are parameters as used in the microbial risk assessment model (25) (available at http://www.19 48. Ti     (6) Third.023 0.

5. Each fraction fi was added to the value of the previous time step. any further steps can be described by Equation 9A. the k-value (Equation 6) was determined from experiments taking into account the effect of a dynamically changing temperature on a lag time value. 7) are used with earlier estimated parameters (17) and are therefore predictions. In each time step the average temperature (and consequently the specific growth rate) was taken as constant. the fraction of the lag time elapsed was 143 . the fraction of the lag time elapsed (ft) in every time step was calculated numerically according to Equation 8. the dynamic first order model for bacterial growth was used and was solved numerically. providing the exponential function as outcome (Equations 9A to C). growth was predicted assuming =0. thus assuming that Cronobacter spp. As a conservative estimate. the lag time was assumed to be completed and growth was assumed to be starting with a specific growth rate corresponding to the temperature prevailing in the bottle at that moment. To specifically estimate lag time effects in this study. started growing instantaneously after addition of water. assuming that during the full time step (Δt) there is exponential growth.Growth during cooling It should be noted that all specific growth rate and the lag time models (Equation 4.  fi       t 60  Ti 1.i       f  i 1   (8) where t is the time step of the simulation [5 min] and ft [-] is the fraction of the lag time elapsed at step i and  (Ti 1. When there was a lag time. To determine the length of the lag time at constantly changing temperature. Once fi ≥ 1. using Equation 9A only. i ) [h] is the average of the lag time of the current and the previous time step. To include changing temperature also during the growth stage. So.

(9C) where Ni [CFU/ml] is the number of micro-organisms present at time step i. Data analysis Temperature profiles during cooling were fitted into Equation 2 using the solver function in Microsoft Office Excel 2003 by minimizing the residual sum of squares (RSS) to estimate the overall heat transfer coefficient (α) and the air temperature (TR). µm(Ti) [h-1] is the specific growth rate at the temperature prevailing in the current time step i and fi [-] is the fraction of the lag time elapsed at step i. The increase in numbers of Cronobacter spp.t (9A) For fi  1 : Ni  N 0 (9B) For the first value of N i  N 0e fi  1:  m Ti f i 1. To exactly determine the commencement of growth after the lag phase.1) of this time step. After the first fi  1 : N i  N i 1e  m Ti . in the step that ft has become greater than 1. the number of micro-organisms was stable (Equation 9B). When the fraction of the lag time elapsed was less than 1.Chapter 7 calculated using Equation 8. Experimental growth curves were fitted to the dynamic growth model to estimate the k-value using the same software. Standard deviations were calculated with Execl and are reported as plus or minus the mean. Confidence intervals were calculated in Excel.t . 144 . Equation 9C describes growth in the remaining fraction (fi . Equation 9C is valid only in the first iteration after the lag phase has been completed. cells assessed in the duplicate growth experiments and the α estimated for the different type of refrigerators were analyzed by analysis of variance (ANOVA).

Growth during cooling

univariate analyses using SPSS (SPSS release 12.01 for Microsoft Windows; SPPS Inc., Chicago, U.S.A.). Significant differences are presented at a 95%-confidence level (P  0.05).

Results and discussion Temperature variability within refrigerator
In our experiments large variations in air temperatures were observed both in time and between different locations in the same empty household refrigerator (Figure 1A). Starting one hour after closing the door with the temperature set at 6.3 °C, the temperature near the cooling element in the back of the upper department varied between -8.4 °C and +10.4 °C. This variation is due to switching on and off the refrigerator’s engine. The lower, middle and upper compartments of the refrigerator showed more constant temperatures and were on average of 5.5 ± 0.8, 6.5 ± 1.3 and 9.9 ± 0.7 °C respectively. These results are in line with data reported elsewhere. James and Evans (14) surveyed 252 household refrigerators without a fan and found that the coldest spot was on average 2.9 °C colder than the warmest spot within the same refrigerator, whereas the greatest range of average temperatures observed in a refrigerator was 12 ºC. Individual (not averaged) temperature differences ranged from 4.5 up to 30.5 °C in the same refrigerator. Door openings resulted in an increase in average temperature and in an even larger spatial variability (15). A variation in local air temperature does not immediately change the temperature of a bottle placed in that refrigerator, but varying temperature profiles may change the air flow and thus affect the heat transfer coefficient (α). Additionally, the speed of cooling as described in Equation 1, will be affected by variations in air temperature. In air-ventilated refrigerators equipped with a fan, the air is continuously in motion and the variation in air temperature is limited (Figure 1B). Near the air inlet the temperature varied from 5.4 to 7.3 ºC (average 6.3 ± 0.5 °C) and near the door the temperature was 7.7 ± 0.1 °C, but all other locations had temperatures of 6.7 ± 0.1 °C.

145

Chapter 7

As a result of the variations in air temperature it was not possible to establish one value for the air temperature required to fit the data to equation 2. Therefore, both the heat transfer coefficient (α) and the air temperature (TR) were fitted simultaneously to the experimental data according to Equation 2. The differences between the measured TR and the fitted TR were on average less than 1 ºC (data not shown).

Figure 1. Air temperatures in an empty household set at 6.3 °C (A) and an air-ventilated refrigerator set at 6.3 °C (B). Measurements started one hour after closing the door. Locations were (grey dashed line), door of upper shelf; (grey solid line), center of the upper shelf; (x), back wall of upper shelf, close to the cooling element (A) or the air inlet (B); (black dashed line), center of the lower shelf.

146

Growth during cooling

Overall heat transfer coefficient (α)
Figure 2 shows typical cooling profiles of the 1-liter beaker and both bottles placed in either a household refrigerator (2A) and an air-ventilated refrigerator (2B). During the cooling process the temperature fluctuated slightly due to variations in the air temperature. In a household refrigerator set at 6.3 °C the time required for cooling the prepared formula to 10 ºC was typically 7.5 hours for 1,000 ml formula in the measuring beaker and 3.5 hours for the 120 ml volume in the two types of bottles. In an air-ventilated refrigerator, cooling times typically were 2 hours less.

Figure 2A. Example of temperature profiles of reconstituted powdered infant formula in (black solid line), 1-liter measuring beaker; (grey solid line), Type 1; and (black dashed line), Type 2 bottles, cooled down in a household type refrigerator. The refrigerator had an average air temperature of 6.3 ºC.

147

Chapter 7

Figure 2B. Example of temperature profiles of reconstituted powdered infant formula in (black solid line), 1-liter measuring beaker; (grey solid line), Type 1; and (black dashed line), Type 2 bottles, cooled down in an air-ventilated refrigerator. The refrigerator had an average air temperature of 6.3 ºC.

With the three containers, 25 cooling experiments were performed in nine different household refrigerators and 9 experiments were performed in an air-ventilated refrigerator. All cooling profiles (3×25 + 3×9) were fitted to Equation 1 to obtain the overall heat transfer coefficient (α [W m-2 °C-1]), a parameter independent of the reconstitution temperature, the air temperature, and within a limited range independent of the size of the object. No consistent differences were found between the α-values determined for the refrigerators that were empty and those that contained also other materials, therefore the data were pooled and used as a single data set. Figure 3A shows the α-values with their standard deviation as a function of air temperature in household refrigerators. An effect of air temperature on the heat transfer coefficient (α) was not observed indeed. The α-values ranged from 5.4 to 12.9 [W m-2 °C-1] in this series of experiments in which temperature settings and filling rates of the 9 household refrigerators were variable.

148

Figure 3B. □).) and 9 experiments in a single air-ventilated (vent. +).Growth during cooling Figure 3A. Δ). as determined in a selection of 11 experiments performed in a single. Markers indicate (×.h. 1-liter measuring beaker. as determined in 9 different household refrigerators at various temperatures.) refrigerator. (▲. Overall heat transfer coefficient (α) (Wm-2 °C-1) with their standard deviations. 149 . Type 2 bottle. Type 1 bottle. (■. Overall heat transfer coefficient (α) (Wm-2 °C-1) with their standard deviations. empty household refrigerator (h.

was filled to its maximum volume.9-19.8 – 13. each containing 120 ml.4 – 11. Estimated overall heat transfer coefficient (α) [Wm-2 °C-1] of bottles Type 1 and Type 2. When a circular flow does develop in a 150 .0  10-4 and 2.7 Number of experiments Standard deviation of average 11 9 11 9 11 9 1. both empty except for the experimental setup consisting of the 3 containers.4 The average values for the beaker and the Type 1 bottle corresponded to cooling rates of 1.4 – 20. and a measuring beaker containing 1000 ml of reconstituted infant formula as measured in one household-type refrigerator and one air-ventilated refrigerator Type of refrigerator Measuring beaker Bottle Type 1 Bottle Type 2 household air-ventilated household air-ventilated household air-ventilated Average overall α [W m-2 °C-1] 9. P=0.4 2.0 10.4 1. The Type 2 bottle. The coefficients in this well-controlled refrigerator varied over the range of 9.9 7. identical to the cooling rates assumed in the FAO/WHO microbial risk assessment model (25).001) than the corresponding values of the measuring beaker and bottle Type 1.0 Wm-2°C-1.7 7.5 2.9 14. It might be that the area available for heat transfer was slightly less than assumed in Equation 2.3 9.01) than in the household type refrigerators. internal heat transfer can not be overlooked.02 m.5 4. since the bottle is rounded at the bottom and not a perfect cylinder. The fitted heat transfer coefficients as shown in Figure 3B are significantly higher (P = 0.7 6.8 95%confidence interval [W m-2 °C-1] 6. In this well-defined series of experiments the range of α-values (6. Table 3. α was significantly lower in household refrigerators (ANOVA univariate analysis.5 1.0  10-4 s-1.Chapter 7 Figure 3B shows a selection of the α-values.5 2.4 – 19. Another explanation is that the lower α might be due to the fact that in this narrow bottle.7 – 13. Average values and confidence intervals are shown in Table 3. with a radius of 0. which is the smallest baby bottle available on the Dutch market.8 – 21. which were measured in one single household refrigerator and one air-ventilated refrigerator.1 13.1 to 12.0 15.0 8.9 [W m-2 °C-1]) in the single household refrigerator was almost as wide as in the 9 different household refrigerators shown in Figure 3A. For bottle Type 2.

were calculated using Equations 4 through 7.2 h. 0. With the temperature at each time point known. Growth of the organism was apparent in the 1-liter measuring beaker that was cooling down from 38 °C in the household-type refrigerator set at 7.1 °C. during cooling were continued using bottle Type 1. To rule out the potential effect of insufficient liquid mixing. both the specific growth rate and the lag time of Cronobacter spp. and thus is in the same order of magnitude as the resulting overall coefficient (see Figure 3). while the logarithmic model (Equation 7) predicted a lag time of 4.Growth during cooling container during cooling. The increase in numbers during the 151 . Figure 4A shows the temperature of a 1-liter beaker placed in a household refrigerator at 7.3 h (data not shown).52 W m-1°C-1) and radius of the container (0.5 °C between core and the wall. According to the inverse of the Ratkowsky model (Equation 5) the lag time was 3. It may therefore have contributed significantly to heat transfer resistance. Furthermore this bottle size and type is used by the majority of the caregivers in The Netherlands . In the same 5-h period. Figure 4B shows that the specific growth rate dropped sharply during the first 5 hours of the cooling process. the lag time increased to values of 20 h and more.02 m). the internal heat transfer coefficient depends solely on the ratio of the thermal conductivity of the reconstituted infant formula (kliquid. The resulting predicted and experimental numbers of Cronobacter spp. If in a narrow bottle such a circular flow does not develop.8 h. the temperature in a bottle or measuring beaker could be predicted during the entire cooling process.1 °C. Growth during cooling On the basis of the estimated overall heat transfer coefficient.05 [-] taken from previous research (17) led to a predicted lag time of 3. An initial estimate for the k-value of 4. the experiments testing the growth of Cronobacter spp. The construction of the lag time under dynamic temperature conditions is shown in Figure 4C. the internal heat transfer coefficient (α) can be estimated to be in the range of 150 to 250 W m-2 °C-1 using the theory of free convection from a vertical plate (5) assuming a temperature difference of 0. The resulting internal heat transfer coefficient is 26 W m-2 °C-1. under dynamic temperature conditions are shown in Figure 4D.

measured and (solid line). lag times at the prevailing temperature. predicted. the temperature of the formula (Figure 4A) will be lower at the moment that growth commences. lag = 0.7 log CFU/ml. With a longer lag time (Figure 4C). A: (×). whereas also the specific 152 . Figure5 shows graphs similar to Figure 4D for both the 1-liter measuring beaker and the Type 1 bottle at 5. Impact of lag time In the predictions a small difference in lag time resulted in considerable differences in the estimated growth. including the 30% of household refrigerators that exceed standard temperatures (10. specific growth rate and (lines).1 °C. The reconstituted formula was prepared from artificially contaminated powder and had a temperature of 38 °C at the start of the experiment. measured and (lines). This range of temperatures was chosen to mimic the range of temperatures that can be found in household refrigerators.Chapter 7 entire 24 hours was 0. Figure 4. 7. 10 and 16 ºC. (black dashed line ▪▪▪▪▪). lag time predicted with the logarithmic model. B: Predictions of (♦). lag time according to Ratkowsky. and (grey line). Various lines represent: (grey dashed line). sakazakii (×). 31). C: Fraction of lag time elapsed and D: number of C. predicted temperature with 95%-confidence interval (dashed lines). Prediction and experimental values of growth during cooling of infant formula in a 1liter measuring beaker in a household refrigerator set at 7.

Growth during cooling

growth rate (Figure 4B) will be lower, together resulting in a considerably smaller increase in cell count. When the lag time was assumed to be zero (grey dashed lines), the exponential growth function overestimated the experimental values considerably in all cases shown in Figure 5. Although Cronobacter spp. cells are known to commence their growth quickly in infant formula, they thus seem to require some time to adjust to their new environment and start multiplying. The logarithmic model overestimated the lag time and thus underestimated the growth in all cases (25). The results of the use of a k-value were often comparable to the Ratkowsky model. The Ratkowsky model (black dashed lines) predicted the growth during cooling at 7 and 10 °C quite well, but at other temperatures this approach underestimated growth. Optimal fits to the experimental data were obtained with k = 1.75 at 4 °C, k = 3.49 at 7 °C, k = 4.07 at 10 °C, k = 2.06 [-] at 16 °C (grey solid lines). The best fit over all 8 experiments was obtained with k = 2.88 [-] (black solid line). These values are all within the range of lag values previously reported for individual experiments at a constant temperature (17, 34). It should be noted that the experiments shown in Figures 4 and 5 were performed with Cronobacter spp. cells that were present in the dry powder, which can be expected to be injured and therefore show relatively long lag phases. It could be that, even longer lag times might have been observed when lower levels of inocula were used than in the current study and when naturally contaminated powered infant formula had been used. On the other hand, in an additional series of experiments using an inoculum of cells grown overnight in BHI, lag times were shorter and could be predicted best by assuming k-values ranging from 1.2 to 1.9 [-] (results not shown).

153

Chapter 7

Figure 5. Growth during cooling of infant formula in (left) 1-liter measuring beaker and (right) the Type 1 baby bottle at air temperatures of 5 °C, 7 °C, 10 °C, and 16 °C in an empty household-type refrigerator. The markers represent experimental data, while the lines are predictions, assuming (grey dashed line ▪▪▪▪), lag = 0; (light grey solid line) lag time according to Ratkowsky; (black solid line), lag time as predicted with k = 2.88; (grey solid line), lag as predicted with an optimal fit of the k-value to the two experiments at each specific temperature: k = 1.75 at 5 °C; k = 3.49 at 7 °C; k = 4.07 at 10 °C; k = 2.06 at 16 °C.

154

Growth during cooling

Table 4. Prediction of the time required to cool down from 40 to 10 °C and the increase in cell counts after 24 h in bottles Type 1 and Type 2, each containing 120 ml, and a measuring beaker containing 1,000 ml of infant formula. Infant formula was reconstituted at 40 °C and placed in refrigerators with an air temperature of 7 °C. Overall heat transfer coefficients with their confidence intervals were taken from Table 3 and for lag time the optimal value k =2.88 [-] was assumed. In parentheses are 95% confidence intervals

Type of refrigerator

Time to reach 10 °C [h]

Measuring beaker Bottle Type 1 Bottle Type 2

household air-ventilated household air-ventilated household air-ventilated

6.8 (5.3 –10.0) 4.8 (3.5 – 8.1) 3.4 (2.6 – 5.1) 2.2 (1.6 – 3.5) 3.6 (2.5 – 6.2) 2.0 (1.3 – 3.7)

Predicted increase in cell counts in 24 h [log CFU/ml] 1.3 (0.8 – 2.3 )

0.7 (0.2 – 1.7 ) 0.2 (0.0 – 0.7) 0.0 (0.0 – 0.2) 0.3 (0.0 – 1.1) 0.0 (0.0 – 0.3)

Prediction of growth during cooling
Using the parameters for heat transfer and lag time derived from the experiments, a series of simulations were performed, to evaluate the effect of reconstitution temperature and refrigerator temperature. Regarding the lag time, the value k =2.88 [-] as found in this study was assumed in all predictions.

Cooling time and growth during 24 hours
Table 4 presents an overview of the estimated time required to cool down infant formula from 40 °C to 10 °C in refrigerators set at 7 °C and the increase in Cronobacter spp. cells in bottles placed in such refrigerator for 24 h. In a household refrigerator, a 1-liter portion of formula reconstituted at 40 °C needed at least 5.3 and up to 10 h to reach a temperature of 10 °C and in an air-ventilated refrigerator at least 3.5 h. Bottles needed on average 3.5 h in the household type and 2.1 h in the air-ventilated type to cool down to 10 °C. In 1-liter portions, growth of Cronobacter spp. was predicted to vary in both types of refrigerators from 0.2 to 2.3 log CFU/ml. In bottles placed in household refrigerators, the

155

Chapter 7

increase in numbers was on average limited to 0.2 or 0.3 log CFU/ml. The maximum increase in bottles placed in household refrigerators was 1.1 log CFU/ml for refrigerators with the poorest heat transfer properties. In most air-ventilated refrigerators set at 7 °C no increase (0.0 log on average) was predicted in bottles during 24 h, while the upper limit of the confidence interval was 0.3 log CFU/ml. The variability of both the time to reach 10 °C and the increase in cells over 24 h mainly originated from the considerable variability of the overall heat transfer coefficient (Table 3).

Varying refrigerator temperature
Figure 6 shows the effect of air temperature in the refrigerator on the predicted growth of Cronobacter spp. during cooling down from 40 °C in the 1-liter-measuring beaker and in the Type 1 bottle filled with 120 ml. Assuming an average heat transfer coefficient of 10 W m-2 °C-1, no growth was predicted to occur in the bottle if the air temperature was 6 °C or lower. Over 6 °C, the numbers were predicted to increase slightly within 4 h, whereas up to 24 h more significant multiplication was predicted, depending on the temperature. In contrast to the bottle, the 1-liter container was predicted to support growth at all refrigerator temperatures. This can be explained by the fact that the resulting temperature of the formula (Tc) during cooling in this container is higher then in the bottles, due to the higher V/A ratio (Equation 1) in the 1-liter container. At the common range of refrigerator temperatures (below 7 °C), growth was predominantly in the first 4 h of cooling, while growth was limited in the period between 4 and 24 h. At higher temperatures, however, Cronobacter spp. was able to continue multiplying after the first 4 h. With the refrigerator set at 16 °C a 4.2-log increase in the bottle and an almost 5.5-log increase in the beaker were predicted.

156

or at the upper or lower limits of its 95%confidence interval. Increase in cell counts as a function of refrigerator temperature at (diamonds). Open symbols and solid lines represent the 1-liter measuring beaker and closed symbols. in formula placed for 24 h in a household refrigerator at 4 °C with a heat transfer coefficient that is either average. since our experiments and simulations were not suitable to measure and to predict cell numbers after thermal inactivation which likely plays a role at these temperatures. results obtained using the microbial risk assessment model (25) are shown in Figure 7B as an increase in relative risk compared to a base-line scenario. Below a reconstitution temperature of 25 °C no growth was predicted under any circumstances.Growth during cooling Figure 6. dotted lines the Type 1 baby bottle. Effect of reconstitution temperature Figure 7A shows the impact of reconstitution temperature on the predicted growth of Cronobacter spp. 24 hours after reconstitution at 40 °C. assuming an average (10 W m-2 °C-1) heat transfer coefficient. Reconstitution at more than 50 °C was not included. For comparison with our study. For the 120-ml bottle our model did not predict growth. 4 h and (triangles). In the 1-liter beaker growth was predicted to occur at reconstitution temperatures over 25. depending on the assumption for the heat transfer coefficient. 32 or 37 °C. unless the reconstitution was over 35 °C and the heat transfer coefficient was at its lower confidence limit. which is reconstitution at 157 .

45 log units in case of poor heat transfer (this study). and it may well affect growth of other micro-organisms. because the refrigerator may be packed with other food products. a shorter lag phase and quicker multiplication may be envisioned. When cooling 1-liter containers. Refrigerators in a home situations may have even poorer heat transfer properties. or bottles may be placed in the refrigerator door. Simulations were limited to 24 h cooling. Although the end-points were not identical. 158 .Chapter 7 25 °C. Our study predicted growth of Cronobacter spp. where heat transfer is at its minimum (15). Moreover. formula is contaminated during reconstitution with cells that are actively growing. however. cells used in this study were present in the powder and can be expected to have a longer lag phase than actively growing cells. already in 1 liter of formula after reconstituting at 26 °C when placed in a household refrigerator with the poorest heat transfer properties. or will be limited to 0. the microbial risk assessment model and our study concluded that in bottles growth will either not occur (both studies). and the door may be opened to frequently. possible present in the reconstituted infant formula. the Cronobacter spp. When. The heat transfer coefficient assumed in our study was measured in an empty refrigerator. the reconstitution temperature was found to affect the proliferation of Cronobacter spp. It should be noted that the latter situation is not a worst case scenario.

Open symbols represent a 1-liter beaker and closed symbols the Type 1 bottle filled with 120 ml of infant formula. (crosses) at lower (6.Growth during cooling Figure 7.7 W m-2 °C-1). and (squares) at upper (13 W m-2 °C-1) limit of confidence interval. The effect of reconstitution temperature on (A) increase in log count (this study) and (B) increase in relative risk (25) during cooling for 24 hours in a refrigerator with an air temperature of 4°C. 159 . average (10 W m-2 °C-1). Heat transfer coefficients are assumed to be (triangles).

in containers of reconstituted infant formula. may prevent multiplication quite drastically as compared to cooling of 1–liter portions. be preferred. however. provided that reconstitution is followed by immediate cooling of bottle-sized portions in a refrigerator. However. Reformulation to allow powders to dissolve at lower temperature would. our results suggests that this may have only a very limited effect on multiplication during cooling. in situations where bottles and/or syringes of infant formula have to be prepared in advance from powdered formula for use within 24 h. airventilated refrigerators can not totally prevent growth in 1–liter portions of prepared formula with reconstitution temperatures around 40 °C. This may suggest that limiting the reconstitution temperature to 25 °C might be an effective control measure.  Our study predicts no growth during cooling after formula reconstitution at or below 25 °C.  Replacing household-type refrigerators with air-ventilated refrigerators with better heat transfer properties is effective in reducing growth during cooling. For example:  Cooling reconstituted infant formula in portion size containers only. Only when the current set point is above 7 °C such measure might have a significant effect. This study provides measurements of both heat transfer characteristics and microbiological growth of Cronobacter spp.  Although lowering the set points of refrigerators in neonatal care units and/or households to 2 °C or less might seem a likely control measure.Chapter 7 Possible control measures Focusing on the cooling phase only. The formulae used in this study dissolved well at lower temperatures. along with a model that is able to estimate the variability of these processes taking place simultaneously 160 . quick cooling of bottle-sized portions (for instance using running water or ice) may be an alternative to prevent growth during cooling. such as bottles. there are several control measures that may help to prevent and minimize growth of Cronobacter spp. for those powders that may not there may be not dissolve at room temperature and need to be reconstituted at 40 °C or more.

Available at: http://en. Biering.engineeringtoolbox.Growth during cooling under dynamic temperature conditions. A comprehensive list of some common liquids and fluids as acetone. W. Weaver. This project was financially supported by Nestlé Research Center. Rigo.. Puntis. 161 . 39:320-322.org/wiki/Polycarbonate. Axelsson. Goulet. Available at: http://www. J. Shamir. exposure of consumers can be simulated for different scenarios of preparation and consumption of infant formulae. and L. References 1. Karlsson. Y. Preparation and handling of powdered infant formula: a commentary by the ESPGHAN Committee on Nutrition. Anonymous. I. water and many more and their specific heat capacities.wikipedia. C. Jonsdottir. The lag time. and potentially of other microorganisms that might be present in powdered infant formula. Anonymous.html. oil. T. Van den Plas. R. The predictive model proposed can be used to estimate the proliferation of Cronobacter spp. K. Szajewska. Accessed 18 February 2008. has a remarkable impact on the overall growth opportunities of Cronobacter spp. Nutr.. J.com/specific-heat-fluids-d_151. Ped. 2004. neonatal health care professionals and (inter)governmental organizations may take benefit from the specific findings to further improve guidelines and best handling practices (33). C. which is known to be highly variable. Agostoni. as it affects both the moment in time and the temperature at which the organism starts to grow. Using the predictive model. Clin. Accessed 18 February 2008. Liquids and Fluids . E. 3. S. F. Three cases of neonatal meningitis caused by Enterobacter sakazakii in powdered milk. 4. Lausanne. Polycarbonate. Switzerland. P. Ludvigsson. Turck. B. thereby aiding governments and industries to identify effective control measures to protect the vulnerable consumers of this product. Gastroenterol. Acknowledgements The authors wish to Kishor Khatri and Els Peters for their contribution to the experimental part of this work. Michaelsen. 1989. Powdered infant formula manufactures. paraffin. Koletzko.. K. Microbiol. and O. 2. Steingrimsson. J. G. N. Clark. 27:2054-2056. O. D. J. H.Specific Heat Capacities.

De Ware(n)chemicus. The Lancet. 18:1323-1329. J.. M. thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant formula milk.. Cronobacter sakazakii subsp sakazakii. M. Syst. D. 7:64. 11.. James. Iversen. Enterobacter sakazakii (emerging issues in food safety) ASM Press.Chapter 7 5. E. J. C. J. 30:569-584. 16. FAO/WHO. Correlating equations for laminar and turbulent free convection from a vertical plate. S. M. 38:378-382. 10. 2008.. Heat Mass Transfer. p. 13. Enterobacter sakazakii and Salmonella in powdered infant formula: meeting report. 1992. J. Int. O'Hara. 15:313-319. Hickman. M. and J. J. and E. Cronobacter sakazakii subsp malonaticus subsp nov. Mullane.. Evans. Farmer III.. and the Enterobacteriaceae study group. E. Occurrence of Enterobacter sakazakii in the food production environments and households. Biol. F. Appl. C. Cronobacter muytjensii sp nov. New Jersey. Chu. J. M. D. J. Lett. 1992. Zwartkruis-Nahuis. Microbiol. Dis. Heuvelink.C.. 15:307-312. Refrig. Ahmed. 2004. Diagn. N. BMC Evol. S. Forsythe.. Joosten. Refrig. In: J. McDowell. Flynn. O. Microbiological Risk Assessment Series No. M. S. and S. S. 2007. Kodde. Int. Refrig. L. D. Churchill. 2002. J. 15:299-306. I. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial outbreaks. I. and H. Gorris. A. and S. J. Kandhai. New Brunswick. Fanning. nov. 2004. O. S. 10 Rome. W.. J. nov. J. C. 1975. M. Cronobacter dublinensis sp nov and Cronobacter genomospecies I. comb. and R. The efficiency and consumer operation of domestic refrigerators. Cronobacter turicensis sp nov. Iversen. 1990. W. 1992. J. A. and S. 9. 6. Van Schothorst. F. J. T. S. Blair. Cooksey. Forsythe. M. 2006. 162 . and M. Italy. R. and H. Stephan. and J. Enterobacter sakazakii: A new species of "Enterobacteriaceae" isolated from clinical specimens. 32:1730. Guillaume-Gentil. Bacteriol. C.. 13:467-472. 14. H. Lane. 7. M. Consumer handling of chilled foods: Temperature performance. Enterobacter sakazakii in melkpoeder. Clark. C. 8. G. J. and descriptions of Cronobacter sakazakii comb. Reij. W. Int. N. Fanning.). The taxonomy of Enterobacter sakazakii : proposal of a new genus Cronobacter gen. Microbiol. J. Hill. C. The growth profile. 15. de Boer. Farber. and D.. Int. Steingrimsson.. 27-59... Isolation and identification of Enterobacter sakazakii. O. 12. Bidlas. C. J. James. M. Int. Forsythe (eds. Washington. The temperature performance of domestic refrigerators. Asbury. nov. A. B. Lehner. 363:39-40.. Marugg. 1980. Infect. Evans. Cleenwerck. Brenner..

C. 2007. Melchers. W. L. S. M. Microbiol. L. Appl. and J. Zwietering. H. J. 2003. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. A.. 29. Accessed 1 Augustus 2009. M. G. 1998. Muytjens. and J. 61:610-616. M. New Jersey. 18:115-120. M. 80:113-122. 2005. Jaspar. Bidawid. and E. 66:370-375. Microbiol. Kollee. 163 . 21. Pagotto. Experimental study of heat transfer by natural convection in a closed cavity: application in a domestic refrigerator. H. Available at: www..Growth during cooling 17. A. M. Bajard. Microbiol. M. Van Schothorst.Case reports and a review of the literature. 23. J. Time-temperature profiles of infant milk formula in hospitals and analysis of Enterobacter sakazakii growth. Effects of preculturing conditions on lag time and specific growth rate of Enterobacter sakazakii in reconstituted powdered infant formula. J. Farber. 20. Smeets. Paoli. 28. Baylac. H. J.mramodels. Nazarowec White.). Flandrois. L. L. 25. Microbiol. Meis. 27. Zanen. C. and G. Appl. 18:1412-1418. Zwietering. J. K. J. A.W. 3rd. C.. Nazarowec-White. children. Ben Amara. J. J. and M. Appl. H. Reij. and J. Noel. Laguerre.. M. W. 2001. Roelofs-Willemse. Comparison of Fecal Coliform Agar and Violet Red Bile Lactose Agar for Fecal Coliform Enumeration in Foods.. Enterobacter sakazakii: infectivity and enterotoxin production in vitro and in vivo. Rosso. 1983. and P. Environ. Lobry. New Brunswick. Clin.. C. infants. and J. Wanegue. Overview of a risk assessment model for Enterobacter sakazakii in powdered infant formula. L. and W. Genetische karakterisatie van Enterobacter sakazakii isolaten van 18. 72:2721-2729. R. 2006. Washington DC. Clin. Food Control. O.. Reij. Environ. J. Muytjens.. 68:1631-1638. 1997. Medicine. 1995. Lai. L. p. H. Voss. S. 2002. Morelli. H. V. M. Convenient model to describe the combined effects of temperature and pH on microbial growth. Farber. 22.. Incidence. 26:743-746. survival. Microbiol. Schaffner (Ed. Integrating concepts: a case study using Enterobacter sakazakii in infant formula. K. In: D. L. Muytjens. I. and adults . J. P. Food Prot. Flick. 19. Food Eng. F. Food Prot. M. and M. G. Microbial Risk Analysis ASM Press. G. Rosset. A. Prepared for WHO/FAO.org. J. H. Kandhai. 24. Analysis of eight cases of neonatal meningitis and sepsis due to Enterobacter sakazakii. 60:226-230. Gorris. F. 2006. 2008. Environ. and D. Sonderkamp.. 26. M. Hartnett. H. 177-204.. P. Farmer. and E. G. and growth of Enterobacter sakazakii in infant formula.. 1988. Leclercq. Enterobacter sakazakii infections among neonates. K. 70:523-537. Wachsmuth. J. H. C. Grognou.

1961. 5:139-153. D.int/foodsafety/publications/micro/pif_guidelines.. WHO. de Wit. E. H. The thermal conductivities of some biological fluids. Steenbekkers. and S. Med. 32. Berne. J.. M. C. Spells. de Koos. 33. 1991.Chapter 7 Nederlandse patiënten met neonatale meningitis.who. and K. De Maertelaere. P. 2005. M. 1960. A. 107:526-533. Food storage and disposal: Consumer practices and knowledge. Biol. 34. M. 30. guidelines. Available at http://www. B. Hasenack. J. N. Appl. Urmenyi. E. Environ. Nijhuis.pdf. storage and handling of powdered infant formula. J. Microbiol. Ned. Food J. 2007. Zwietering. C. Neonatal death from pigmented coliform infection. Accessed 13 February. M. 31. and M. Terpstra. Brit. T. Tijdschr. Modeling of bacterial growth as a function of temperature. 164 .. Safe preparation. A. K. Phys. L. 6:113-115. 1:313-315. 2009. 57:1094-1101. Med. Van 't Riet. C. Microbiol. The Lancet.

Inactivation of Cronobacter spp. 165 .

M. decline rates of 0. M. and enumeration of survivors with a minimum detection level of 100 CFU/g. Assuming a linear relationship between log transformed D-values and temperature. Journal of Food Protection.G. Cronobacter spp. This Chapter has been accepted for publication as “Inactivation rates of Cronobacter spp. around 104 CFU/g of powder. To describe the effect of temperature on survival. 2010. respectively.011 and 0. z-values estimated for C.M. both the Weibull distribution model and the log-linear model were tested. Van Schothorst. M.H.. it was found that the number of bacterial cells decreased faster with increasing temperature. while such a difference in survival was not apparent at other temperatures. Such differences found in resistance among Cronobacter spp.W. Zwietering. were 13.Chapter 8 Abstract The aim of the study was to determine the survival of two strains of Cronobacter (Enterobacter sakazakii) and six other bacterial strains inoculated into dry powdered infant formula (PIF) stored for 22 weeks at several temperatures between 7 and 42 ºC.C. For all strains tested. and M. At 22 °C. would be relevant to consider when establishing quantitative risk assessments on consumer risks related to PIF. respectively. 166 . L. The experimental setup involved a relatively high initial concentration of bacteria. and selected other bacterial strains in powdered infant formulae stored at different temperatures” Kandhai. sakazakii ATCC 29544 and Cronobacter MC10. Gorris.008 log units/day were found for Cronobacter sakazakii ATCC 29544 and Cronobacter strain MC 10.5 ºC. Reij. cells generally survived better at high temperatures (37 and 42 ºC) than the other bacteria.3 and 23.

being rarely involved in food-borne disease among neonates (14). 13. 167 . In several cases.. 17) as 6 species in a new genus. 26). dublinensis. is used consistently in the subsequent part of this paper when the micro-organisms are referred to in general terms or when historic information that is not specific to the various newly recognized species is cited. this Gram-negative micro-organism has been reclassified (16. malonaticus. Only recently. Cronobacter gen. sakazakii. C. C. As an example in one study it could not be isolated in milk. Cronobacter spp. Such levels. an overall mortality rate of 19% was estimated. based on data presented in the Food and Agricultural Organization and the World Helath Organization report (11). The sixth species is referred to as genomospecies I and currently includes two representative strains.Inactivation of Cronobacter spp. within the family of Enterobacteriaceae. 10). are opportunistic pathogens that have caused illness particularly in premature and neonatal infants.7% (29) and less recently 14% (22). Introduction The International Commission on Microbiological Specifications for Foods (ICMSF) has placed Enterobacter sakazakii in the category of “severe hazards for restricted population”. Prevalence figures for Cronobacter spp. nov. 6. in PIF are scarce. reflect operations that lack stringent hygiene measures as reductions ranging from about 10-4 to 10-6 CFU/g may be possible in adequately hygienic operations (5. The prevalence of these and other microorganisms in PIF may depend on many factors. C. has not been consistently isolated from PIF. when present at all. with the mean level in contaminated product possibly being in the order of 10-3 CFU/g (8. in PIF are thought to be low. the source of infection has been identified as powdered infant formula (PIF) (4. The collective of species originally known as E. can be referred to as Cronobacter spp. Cronobacter spp.or soy based infant formulae or the infant herb drinks (24). 11). including factors related to manufacturing process. turicensis. Five of the new species are Cronobacter sakazakii. The designation Cronobacter spp. but may differ widely as they have been reported to be about 1% and 3% (data quoted by Food and Agricultural Organization/World Health Organization in 2008) (11). Over the past fifty years. muytjensii and C. Levels of Cronobacter spp. however. choice of ingredients and postprocess handling.

In the case where micro-organisms are present in the factory environment. in dry residues and powder has been reported by several authors (3. recent research had shown that several isolates of Cronobacter spp. low water activity (aw) is an important determinant of the survival of Cronobacter spp. In this study. 23). Though PIF manufacturing is advocated to be undertaken respecting high hygiene conditions and keeping the environment essentially dry. Contamination after pasteurization can occur through at least two routes. Because the temperature in the factory environments may allow the micro-organisms to proliferate in residues to relatively high numbers in a short time. A second route is environmental contamination. recontamination after pasteurization may occur from such environmental sources (5). to multiply rapidly (18). The specific aim was to quantify key parameters that express the inactivation 168 . are vegetative micro-organisms that are not particularly heat resistant (23). in dry factory environments and the same may hold true for the final product during storage and retail. allowing Cronobacter spp. However. 6. Reconstitution with water of 70 ºC before final use of the dry PIF will decrease the number of micro-organisms by at least 100. 12).000 fold. were able to survive spray drying. was determined in PIF stored at a number of different temperatures and compared to those of six other microorganisms. Unfortunately. into the final product when they are present in the ingredients. the survival rates of two strains of Cronobacter spp. but the dynamics have not yet been described mathematically. which may have to be added after pasteurization and may introduce Cronobacter spp.Chapter 8 Cronobacter spp. The presence of the microorganisms in the factory environment may pose a special problem where factory environments become wet. It has been reported that the micro-organism will not survive pasteurization treatments normally given during manufacture of wet-processed infant formulae (15. the rate of survival in dried-up material is a key factor determining the extent of recontamination possible during PIF manufacturing. Quite long survival of Cronobacter spp. the impact of aw on the growth potential of these micro-organisms under realistic conditions has not been reported in great detail. for instance during wet cleaning. it is important to note that PIF is not manufactured to be a sterile product. rendering the final product safe (10). One such route is via heat-sensitive ingredients. Conceivably. a process that is not intended to be a pasteurization treatment (1).

The Netherlands). Cronobacter sakazakii ATCC 29544 (Enterobacter sakazakii). Bacillus cereus ATCC 14579 and Staphylococcus aureus ATCC 14458 SEB were used. Materials and Methods Preparation of the inoculum The following strains. St. Wageningen. Escherichia coli O2K.7 ml of the bacterial culture. The obtained bacterial suspension with a final concentration of around 107 CFU/ml was transferred into tubes of a perfume sprayer (Designed by Gérard Brinard. due to desiccation in PIF at low aw at several different temperatures. Switzerland) and 0. All the strains were pre-cultured by adding 100 l of the stock to 100 ml of Brain Heart Infusion Broth (BHI. kindly provided by Dr.ec0001. Cronobacter strain MC10 (Enterobacter sakazakii MC10). Grown cells were centrifuged for 10 min at 20 C at 2958 x g (MSE. These strains were obtained from the culture collection of the Laboratory of Food Microbiology (Wageningen University.Inactivation of Cronobacter spp.3 ml 87% glycerol (Fluka-Chemica GmbH. Becton Dickinson and Co. Formula manufacturers may be able to use the data in establishing adequate Good Hygienic Practices (GHP) and Hazard Analysis Critical Control Point (HACCP) plans. rate of Cronobacter spp.. Radboud. Nijmegen. Leusden. The Netherlands). Buchs. Klebsiella pneumoniae kl0001. Bacillus cereus ATCC 14579 is an air isolate. Germany). aureus ATCC 14458 SEB is a human feces isolate. Enterobacter cloacae eb0001. 169 . United Kingdom) and washed twice in 40 ml of 0. Le Pont de Claix. Frickenhausen. Such parameters are considered important for microbiological risk assessments and for product/process design studies. France) and incubating at 37 °C for 22 hours to obtain a mature culture. containing 0. The stock cultures were stored at -20 ºC in cryo vials (Greiner Bioone GmbH. Mistral 3000i.85% physiological salt solution. Cronobacter MC10 is a clinical isolate. Salmonella serovar Enteritidis (phage type 4). while the other strains were food isolates. The Netherlands) to artificially contaminate PIF (18). DA Drogisterij. Harry Muytjens (University Medical Center St. Leicester.

coli. the detection limit was 100 CFU/g. Heythuysen. The Netherlands). except on week-end days. VWR International Prolab. while contaminated powder stored at 37 and 42 ºC was sampled once every week.Chapter 8 Inoculation of the powdered infant formula Powdered infant formula for infants from 0 to 6 months was bought locally and aseptically divided in sterilized square boxes (Greiner bio-one GmbH. Alphen aan de Rijn. which has a constant aw of 0. Leerdam.. 30. The Netherlands). The survival of the micro-organisms during storage at 7. 170 . each containing 20 grams. The Netherlands). In this additional experiment. 8. using a perfume sprayer with an individual bacterial suspension and rigorously shaken. Samples of 1 g of PIF were reconstituted in 9 ml peptone saline solution (NaCl. 22. Normapur. 15. The aw of the powdered infant formula measured immediately after spraying with the bacterial suspension remained below 0. Basingstoke. England) by using a spiral plater device (Eddy Jet. Oxoid Ltd. Powdered infant formula was sprayed six times. IUL instruments. samples were taken every day. cloacae. The aw of the contaminated powder was determined using a water activity (aw) meter (Novasina Aw-box. and neutralized bacteriological peptone.3. Basingstoke. 1 g/L. As the decline in the numbers of bacteria was found to progress very rapidly at 37 and 42 ºC and very few data-points reflecting this decline were obtained. E. and 42 °C for 22 weeks. for 16 days. resulting in a final concentration of bacterial cells in the dry powder between 104 and 105 CFU/g. an additional experiment was conducted with five of the strains of bacteria (i.11. The desiccators with the inoculated PIF were incubated at the following different temperatures: 7. 37. 22 and 30 °C was determined by sampling the contaminated PIF once every two weeks. England). Fontenay sous Bois France).5 g/L. Inoculated plates were incubated for 20 to 24 h at 37 ºC and typical colonies were counted by manual enumeration. Oxoid Ltd. Cronobacter MC10. The contaminated powder was divided over small plastic containers and placed in desiccators containing saturated lithium chloride (LiCl. Pedak. and St.e. 15. E. IKS BV. aureus) with storage at 37 ºC. Cronobacter sakazakii ATCC 29544. and appropriate dilutions were surface plated in duplicate onto tryptone soya agar (TSA.

it is a parameter that describes the shape of the inactivation curve. The goodness of fit of both the Weibull model and the log- 171 . the experimental data had to meet the following requirements: 1) a minimum of three data points were required. and the other bacterial strains evaluated over time. 28).03. Survival data that did not meet these requirements were excluded from fitting using the Weibull model. D [days] is the time of the first decimal reduction (time required to reduce the population by 1 log unit from the initial level at t = 0) as obtained by the log-linear model (Dl) and the Weibull distribution model (Dw). p  t  log N t   logN 0    D    w log N t  log N 0    (1)       t D l     (2) where t [days] is the time. Germany). 2) at least two data points had to be above the detection limit of 100 CFU/g.Inactivation of Cronobacter spp. Determination of the D-value To describe the rate of survival of the two strains of Cronobacter spp. Parameter p [-] (p > 0) in equation 1 has no direct biological relevance. (Jandel Scientific. If p = 1 then equation 1 is identical to equation 2 and the decline of the Log count is linear (Equation 2). All parameters in the Weibull distribution model and the log-linear model were obtained via minimizing the residual sum of squares (RSS) using the solver option in Microsoft Office Excel 2003 and were verified using Table curve 2D. both the Weibull distribution model (Equation 1) and the log-linear function (Equation 2) were applied to the bacterial counts and used to determine the decimal reduction time Dw and Dl (20. N0 is the initial bacterial concentration at time t = 0. for Windows 2. the start of the incubation period. N(t) [CFU/g] is the bacterial concentration at time t. Erkrath. In order to obtain reliable estimates for Dw and Dl parameters.

e. Calculation of the z-value Traditionally. and this parameter was monitored during storage. This is done by dividing the RSS value by the degrees of freedom (DF). i. B. in powdered infant formula during storage for 22 weeks at 7. 15 and 22 ºC.  T  Tref logDref   logDrefmodel     z      (3) where log (DT) [log days].value the 95% confidence interval was calculated. Since the water activity varies with temperature. 37 and 42 ºC. coli. and Tref is the reference temperature [ºC]. Results This study assessed the survival of eight strains of bacteria. reaching values of 0. Cronobacter sakazakii ATCC 29544. and products were stored at different temperatures some level of inaccuracy may be associated to the results reported here. In addition. The estimated D-values were therefore log transformed in order to obtain a linear relationship with temperature (27). cloacae. 37 ºC and 42 ºC.152 [-] after 5 days storage. and 0. it has been assumed that the temperature dependence of the D-value versus the temperature is log linear and can be expressed with the z-value. The water activity decreased more slowly at 7. is the logarithm of the D-value [days]. the average water activity had dropped to 0. log (Dref model ) is the log Dref value at Tref. pneumoniae. The z-value is the increase in temperature (ºC) that is needed to reduce or extend the D-value by a factor of ten. 22. Equation 3 was used to estimate the z-values. then the log-linear model was also suitable to describe the data. E. when p = 1 falls within the 95% confidence interval. A key parameter determining the potential survival in PIF is water activity. 30. 172 . 15. Cronobacter MC10. cereus and S.Chapter 8 linear model was assessed by calculating the mean square error (MSE).286. aureus. E. After 5 days of storage at 30 ºC. Kl.129 (-). 0. for each p.180. Salmonella.

08 log units/days. while a somewhat faster decline could be seen at 30 oC and a somewhat slower decline was noticeable at 15 and 7 oC. E. namely C. Symbols on the detection limit. aureus could be recovered incidentally beyond the first week of storage. coli and E. the first strain that declined below the detection limit was B. inactivation in PIF at 37 oC shows a rather nonlinear curve. 100 CFU/g. indicate that the results were either on or below the detection limit. E. Cronobacter spp. with the latter at the higher level of the two. coli and S. strains. respectively. 173 . Solid and broken lines indicate fitting of the data-points with a nonlinear model (the Weibull model shown in Equation 1) and a linear model (the log-linear model shown in Equation 2). In this experiment. with S. All the other strains tested were below the detection limit of 100 CFU/g already after one week. The other strains were detectable significantly longer. cloacae.07 and 0. The results obtained are depicted in Figure 2. a second experiment was performed with five of the strains. the apparent decline at 37 and at 42 oC progressed very rapidly for all bacteria investigated. cereus at 7 weeks of incubation. coli. Because of the limited number of data points obtained in this experiment. cloacae up to 20 weeks. pneumoniae and Salmonella up to 18 weeks and E. In order to gather more data points at 37 oC. with more frequent monitoring of the levels of bacteria than in the first experiment. and S. Kl. C. at the end of storage.Inactivation of Cronobacter spp. E. decline rates at 42 ºC could only be estimated for C. cloacae. Figure 2 indicates that for several bacteria. incubation was for a period of 16 days. At the two highest temperatures evaluated (37 and 42 ºC). sakazakii and Cronobacter MC10 and these were calculated to be 0. E. Cronobacter MC10. Data fitting did include data points on or below the detection limit. at least a 2-log CFU/g decrease was found within the 16 days of the incubation period. After 22 weeks. although the data for individual strains showed a somewhat erratic profile. sakazakii and Cronobacter MC10 were still detectable. Figures 1 and 2 show the enumeration results for the various bacterial strains at the various temperatures over time. sakazakii. Compared to room temperature storage. Notably. At room temperature (Figure 1. For all the strains tested in the second experiment. aureus detectable up to 17 weeks. which would allow to more accurately quantify the decline in cell numbers during the initial incubation period. aureus. results at 22 ºC) the trend for all strains was gradual inactivation.

which for the log-linear model is equal to the rate of subsequent decimal reductions. D expresses the rate of the first decimal reduction. whereas for the Weibull model this Dw value can change depending on the value for parameter p. With 0 < p <1. two different modeling approaches were used to establish times for the first decimal reduction (D). In both cases. giving D values referred to as Dw. were calculated as a reflection of the inactivation rate (Table 1). or the loglinear model (Equation 2).Chapter 8 For the various data obtained in the first and second storage experiment the times for the first decimal reduction. From Table 1 it can be deduced for all bacteria investigated that. a fitted curve shows a decline in rate over time and thus a slow down of the speed of inactivation. Decimal reduction times were estimated by fitting the log-transformed data with either the Weibull distribution model (Equation 1). they were often in the same order of magnitude for individual bacteria at specific temperatures but more importantly consistently showed higher initial inactivation rates with an increase in storage temperature. where possible. 174 . giving D values designated Dl. while there were sometimes notable differences between the absolute values of Dw and Dl. To quantify the progressive decline in cell numbers during storage.

○ = S.  = Salmonella. coli ec0001. 15 ºC and 7 ºC for a period of 22 weeks. cloacae eb0001. 22 ºC. 37 ºC. □ = Kl.  = Cronobacter sakazakii ATCC 29544.  = Cronobacter MC10 . ● = B. ■ = E. Figure 1. ▲ = E.Inactivation of Cronobacter spp. Enumeration of various bacterial strains in powdered infant formula at 42 ºC. 175 . aureus 14458 SEB. cereus ATCC 14579. pneumoniae kl0001 . 30 ºC.

Survival of various bacterial strains in powdered infant formula at 37 ºC for a period of 16 days. aureus 14458 SEB (E). ■ = E. 176 .  =Cronobacter spp. the fit of the Weibull model (________) and the log linear regression (----------). cloacae eb0001 (D). coli ec0001 (C).  = Cronobacter sakazakii ATCC 29544 (A). ▲ = E. ○ = S.Chapter 8 Figure 2. (B) .

Inactivation of Cronobacter spp. 177 .

Chapter 8 178 .

Inactivation of Cronobacter spp. 179 .

Chapter 8 180 .

Inactivation of Cronobacter spp. Table 1 shows the values computed with the Weibull model for parameter p (Equation 1). were plotted against the incubation temperature (Figure 3). The estimated values for MSEw and MSEl. 181 .. the log-linear model. respectively. while these values were lower at 30 ºC or above. pneumoniae showed very fast decimal reduction (Dw-value of 22. sakazakii and Cronobacter spp. B. The Dw-values at 22 ºC calculated for the two Cronobacter strains (78 and 67 days) were comparable with the Dw-value estimated for the E. respectively for the Weibull model and the log-linear regression model. In an effort to characterize the effect of temperature on the decline of bacteria in PIF in quantitative terms. were similar in many instances. 14 and 13 days. For the two strains of Cronobacter spp. Comparing the Dw-values in Table 1 for the various bacteria. cereus and Kl. respectively). being 78 days and 95 days. cloacae. generally. Dw-values at 22 ºC were generally higher at 7 or 15 ºC. However. reflecting a rapid first decline followed by successively slower inactivation. whereas E. with better fits often at the lower storage temperatures. respectively 67 days and 119 days for C. the log transformed first decimal reduction times calculated with the Weibull model (Dw-values). overall. In the last column an indication is given whether the value for parameter p is equal to 1 within the confidence interval of the data obtained and thus Equation 1 equals Equation 2. From the total of 37 cases. The fit of the two models with the data points. expressed in mean square error (MSE) values.and Dl values at 22 ºC were quite different. coli strain (Dw-value of 64 days) and somewhat higher than that of Salmonella (Dw-value 47 days). In 31 (84%) of the 37 cases (as indicated in the former last column in Table 1). resulting in a Dw-value of 169 days. this model provided for the better fit of the data. the Weibull model showed equal or lower MSE values than the log-linear model meaning that. showed that the Weibull model generally provided the better fit as compared to the log-linear model (Table 1). MC10. St. differences in MSE values between models were relatively small and also on the basis of the p parameter in the Weibull model it appeared that either model might be applicable in certain cases. the log-linear model is valid 19 times (51%) on the basis of parameter p. aureus did not show a consistent decrease in numbers at 22 ºC. Dw. For individual micro-organisms.

z-values were calculated with Equation 3 using 22 oC as the reference temperature.  = Cronobacter sakazakii ATCC 29544 .Chapter 8 Dw values rather than Dl values were chosen as the Weibull model fitted the data generally better than the log-linear model. sakazakii ATCC 29544 (z=13. cloacae eb0001.2 C).3 C) seemed to be the most temperature depended while the inactivation rate of Kl. cereus ATCC 14579. aureus 14458 SEB. coli ec0001. B.  = Cronobacter MC10 . coli were all very close together (range: 23.  = Salmonella. Assuming a linear relation for the temperature impact on survival. Table 2 presents the calculated z-values and the log Dref model values for the various bacterial strains investigated. cereus. ■ = E. ● = B. pneumoniae was the least temperature dependent (z = 43. aureus was relatively low (19. □ = Kl.2 C) while the z-value for St.2 C) and that of Salmonella relatively high (28. The log transformed estimated first decimal reduction time (Dw-values) for several microorganisms in powdered infant formula at various temperatures.3 C) during incubation in powdered infant formula. Estimated z-values for Cronobacter MC10. ▲ = E. Figure 3.3 to 25. From these data the inactivation rate of strain C. cloacae and E. E. ○ = S. 182 . pneumoniae kl0001.

Cronobacter MC10. E.3 25.2 23. Though much has been done in manufacturing operations in terms of process optimization and control of hygiene.Inactivation of Cronobacter spp.59 1. Levels of Cronobacter spp. coli ec0001. S. pneumoniae kl0001 S.92 CFU/g (8). aureus 14458 SEB. during storage 183 . including Cronobacter spp. sakazakii ATCC 29544 Cronobacter MC10 Salmonella B. cereus ATCC 14579. PIF is not designed and processed to be a sterile product and as a consequence storage. cereus ATCC 14579 Kl. in contaminated PIF have been reported.2 23.3 19.5 43. from the processing environment in practice (10.81 1. aureus 14458 SEB E. in the range of 0. in three surveys. adherence to proper hygiene and manufacturing practices tailored to PIF production should ensure levels of. handling and preparation before consumption have to take account of the presence of potentially harmful bacteria.77 1. Salmonella at 22 ºC Strain z-value [ºC] Log (Dref model[days])a C. for instance. E. In effect. cloacae eb0001.25 1. pneumoniae kl0001. However. B.40 1.2 1. 21). coli ec0001 a 13.23 1.5 28. Whether and how the level of such bacteria changes from the manufacturing phase. Cronobacter spp.002 to 0. Estimates of the z-values for Cronobacter sakazakii ATCC 29544. the genus that has been identified as an opportunistic pathogen potentially present in PIF and a risk factor for specific vulnerable consumers.3 23.. Table 2. Discussion In this study we investigated the dynamics of several different bacteria possibly associated with powdered infant formula including two strains of Cronobacter spp. to be below 10-4 CFU/g (5).58 1.59 Dref with Tref 22 ºC. Kl. at present it is not possible to totally eliminate Enterobacteriaceae. bacteria may be present in PIF after manufacture and packaging. cloacae eb0001 E.

For the various bacteria tested. may depend on factors such as the length of storage and storage temperature. exhausting their energy reserves. At 37 and 42 oC. but then these temperatures are usually not temperatures at which PIF is stored in practice.Chapter 8 up to consumer use. The underlying mechanism for this may relate to metabolic exhaustion at elevated temperatures as reported before (19). respectively. aureus. as well as genetic and structural building blocks. strains were slightly lower (2 days at 37 ºC) than those of the others (3 to 5 days). Vegetative cells tend to initiate every possible repair mechanism to overcome the impact of temperature stress on metabolism and cell integrity. cloacae. reductions were generally very rapid. the estimated D-values at 37 ºC were 2 days (for Dw) and 7 and 12 days (for Dl). However. higher storage temperatures seem to cause or contribute to a progressively rapid die-off of micro-organisms. In terms of achieving a reduction in the level of potentially hazardous micro-organisms remaining in PIF when present from the manufacturing phase. In this study it was found for all bacteria. storage at room temperature or above may thus have an advantageous effect over cold storage. At temperatures of 37 and 42 oC. However. For C. the estimated values for D-values were found to decrease with increasing temperature (Table 1) putting the observations made regarding Figures 1 and 2 on a quantitative basis. strains). such temperatures (and higher) are not uncommon in environments surrounding spray-dryers. the rate of decline was found to be somewhat more rapid that at 22 oC. E. rather the low water activity of the product would render the micro-organisms metabolically inactive within a short space of time. Comparing the five strains of which the decline rate could be derived at 37 oC (St. Consistently. that levels in PIF declined over time and that the rate of decline was influenced by storage temperature (Figure 1). loss of viability is particularly rapid compared to storage at 22 oC. D-values between 30 ºC 184 . At 30 oC. Notably. while the former could be a valid storage temperature in countries with a tropical climate. Decline at 7 and 15 oC progressed more slowly than at 22 oC. The latter is a realistic storage temperature generally. coli and both Cronobacter spp. sakazakii and Cronobacter MC10. E. there may be a detrimental impact on product quality and/or nutritional properties of high temperature storage. it seems that the Dw-value of the two Cronobacter spp.

However. and together these findings underscore that that survival of Cronobacter spp. As indicated by the 95% confidence interval data for the Dw.014 log units per day. Thus. While the first rapid phase of inactivation was characterized by a decline rate of 0. temperatures that can be room temperature in practice.014 log units per day reported in literature (7. sakazakii and Cronobacter MC10. progressive inactivation of most bacteria may take place in the course of storage.013 log units per day for C. in dry powdered infant formulae only considered storage at room temperature and expressed the decline as a rate in log units per day. the data indicated that at 22 and 30 oC. these literature values were computed considering the inactivation curves to be biphasic. remained viable and well over the detection limit of the experimental set-up for the whole 22 weeks storage period. Previous studies on the inactivation of Cronobacter spp. Many other studies reported resistance to low aw (2. respectively. decline rates may gradually slow down in the course of storage. as seen frequently in the current study for this bacterium and several others investigated. Also the Dw. indicates that the micro-organisms posses a certain level of resistance or tolerance to dryness. At 7 and 15 oC.015 log units per day for Cronobacter MC10. and 7 ºC ranged between 49 and 713 days and between 49 and 332 days for C.011 and 0.001 log units per day (10). the second phase showed a much lower decline rate for Cronobacter spp. at these lower temperatures. a (so-called) tailing effect that was 185 . as calculated from the Dl. decline rates observed in this study at room temperature were found to be 0. Notably. these ranges occasionally varied considerably. 10). The fact that these and other bacteria tested here survived under the dry conditions prevailing in PIF. 15). 6. inactivation was limited to about 2 log cycles for the various bacteria. respectively. To compare the current findings with those of previous studies. of 0.Inactivation of Cronobacter spp.and Dl -values in Table 1. However. the decimal reduction times found at 22 oC were converted to decline rates (log units/day) by taking the reciprocal of the D-values.and Dw-values. in PIF is a matter to be carefully considered in assessing the safe preparation of infant formula for final consumption. 12. These values compare well with the decline rate of 0.008 and 0.and Dl -values estimated for the other six bacteria at the six different incubation temperatures evaluated showed significant individual variation. overall. sakazakii and 0. Nevertheless. 3. the two strains of Cronobacter spp.

a 1-log reduction may occur within 2 to 3 months while a 3-log reduction may take place in 11 to 25 months. though it is likely that in most cases PIF is consumed within a shorter time frame (10). reduction is faster than at room temperature. as this generally fitted the dynamic behavior of most bacteria evaluated here best. it would be more realistic to consider the actual dynamic change in decline rate over storage in predictions.g. for instance as part of a microbiological risk assessment study. At 22 oC. Interpretation of these calculations should take into account the significant variation in the estimated D-values determined in this study and that samples were taken during a 22-week period.Chapter 8 most pronounced at low storage temperatures (7 and 15 ºC). Extrapolations beyond that time frame have been indicated in parentheses in Table 3 and should be interpretated as orders of magnitude only. 186 . an 1-log inactivation in the level of Cronobacter spp. while under refrigeration. e. the most rapid as the best case or the slowest as the most conservative case. may occur within 3 to 24 months of storage. in PIF are projected as times required for a certain decimal reduction using both the Weibull and the log-linear model. The data obtained in the current investigation show that a dynamic change in decline rate can be reflected well by using the Weibull model. survival will exceed the maximum shelf life of the product. Comparing times predicted using the Weibull model. The projected times are in months and are best interpreted considering that the maximum shelf-life of PIF is about 30 months. Rather than choosing one decline rate. it can be seen that at 7 oC. In Table 3. At elevated temperatures. our findings considering the impact of storage temperature on the decline of Cronobacter spp.

Inactivation of Cronobacter spp. 187 .

MC10. It should be noted that z-values reported in this paper are based on Dw-values as estimated in dry PIF and do not apply to earlier reported thermal inactivation in liquid (25). it has been possible to estimate z-values which can be used to relate inactivation kinetics between temperatures. Cronobacter cells are rapidly inactivated in dry powder at temperatures of 37 °C and 42 °C.and zvalues and the quantitative substantiation of the very long survival time of Cronobacter spp. sakazakii and Cronobacter spp. D. As shown in Table 2. Our findings suggest that there may be wide differences in the inactivation rates of Cronobacter spp. The Weibull model in conjunction with the inactivation data generated in this study. Formula manufacturers may be able to use the model and other results in establishing adequate Good Hygienic Practices (GHP) and Hazard Analysis Critical Control Point (HACCP) plans.Chapter 8 Since inactivation kinetics data have been accumulated for several different temperatures in the present investigation. respectively. but that inactivation rates at ambient temperatures or lower tend to progressively slow-down. may be useful for reflecting the dynamic reduction of Cronobacter spp. z-value of 13. in 188 . Several observations may need to be considered in the design of Good Manufacturing Practices (GMP) and HACCP systems for PIF manufacture.001 log units per day assumed in the Food and Agricultural Organization/ World helath Organization risk assessment study (11). In most cases. for instance. including the impact of storage at temperatures over ambient on Cronobacter spp. quantification of survival rates of Cronobacter spp. in dry PIF may provide useful data for microbiological risk assessments and for product/process design studies. Further work may need to address interstrain differences of Cronobacter spp. in PIF. in microbiological risk assessments. Thus the given D/z-concept is specific for inactivation in dry PIF at moderate temperatures. inactivation. z-values for the various bacteria were established on the basis of the Dw-values at different storage temperatures.5 ºC were estimated for C. inactivation rates may be higher than the rather slow decline rate of 0. In conclusion.3 ºC and 23. in dessicated PIF. as it allows determining likely levels of the micro-organism in PIF subjected to different scenarios of time and temperature during storage and distribution. in PIF. however. Considering the better fit of the Weibull distribution model with experimental data.

. J. 2003. D. J. J. Washington.. L. and S. Switzerland. Edelson-Mammel. Clark. H. Mullane. 145-185. 27-59. p. p. Infect. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial outbreaks. Dis. New Jersey. Buchanan. 7.L. B. Enterobacter sakazakii (emerging issues in food safety).. S. 13:467-72. Food Prot. Enterobacter sakazakii (emerging issues in food safety). A. B. Food Prot. C. 2008. J. Forsythe (eds.. 3. 2004. 189 . J. O'Hara. Farber. Dairy Technol. Appl. 2008. P. Steingrimsson. D. Microbiol. 5. 70:2111-2117. E. 8. 61:102-108. Enterobacter sakazakii survives spray drying. L. New Brunswick. J. M. Edelson-Mammel. 2008.Inactivation of Cronobacter spp. is gratefully acknowledged. Farber. Our data. Food Prot. J.. Arku. Fox. N. Desiccation and heat tolerance of Enterobacter sakazakii. J. M. G. ASM Press. A. In: J. Acknowledgements The financial support by Nestlé Research Center. and H.). Fanning.. N. New Brunswick. and R. C. Forsythe (eds.C. Washington. Hill.C. Diagn.. 2. their ability to survive dry environments. and S. C. should not be assumed to represent a worst-case scenario of the most desiccation-resistant strains that may be present in the environment. Forsythe. Muytjens for providing the Cronobacter strain MC10. In: J. M. M. Jordan. Breeuwer. S. 68:1900-1902. and R. G. 2007. Forsythe. 6. 1990. Lausanne. Fanning. S. Int. 2005. O. Peterz. Dry stress and survival time of Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered infant formula. 4. and K. References 1. ASM Press. Lardeau. and S. 67:60-63. and S. Caubilla-Barron. J. L. Survival of Enterobacter sakazakii in a dehydrated powdered infant formula. Microbiol. Isolation and identification of Enterobacter sakazakii.). The authors thank Samia Amrouche for carrying out part of the experiments and Dr. Thermal inactivation of Enterobacter sakazakii in rehydrated infant formula. New Jersey. Buchanan... 95:967-973. Joosten. collected from two clinical isolates in this study. Cordier. Cooksey. J. J. S. and R. Production of powdered infant formulae and microbiological control measures. C.

and descriptions of Cronobacter sakazakii comb. C. 15. Cronobacter gen. 2006. Rome. Harris. C.. and M. Cronobacter malonaticus sp. nov. Cronobacter dublinensis subsp. I. water activity. The growth profile. Microbiological Risk Assessment Series No. C. Italy. Stephan.. Lehner. nov. Cronobacter dublinensis sp. Appl. Survival of Enterobacter sakazakii in powdered infant formula as affected by composition. 58:1442-1447. nov.. Springer Publishers. R. I. B. B. nov. 12. Microbiol. comb. Fannin. C. nov. McCardel. Enterobacter sakazakii (Cronobacter spp. 11. 2002. FAO/WHO. Tal. S. lactaridi subsp. J. FAO/WHO.. Mullane. J. 2004. Cronobacter muytjensii sp nov.. 17.. Bidlas.. 2002. Microbiol. Cronobacter sakazakii subsp sakazakii. Forsythe. nov. M. The taxonomy of Enterobacter sakazakii: proposal of a new genus Cronobacter gen. L. Joosten. 13. 2006. nov. 10. A. W. nov. and H.. B. 10 Rome. BMC Evol Biol. New York. 7:64. H. Appl. Lehner. Italy. Cronobacter turicensis sp. nov. and L. M. Lorch. comb.. nov. Marugg. Syst. Beuchat. 16. J. J. R. N. Iversen.) in powdered followup formulae: meeting report. Gorris. Cronobacter dublinensis sp nov and Cronobacter genomospecies I. R.Chapter 8 9. M. Mullane. and of three subspecies. V. S. Cronobacter genomospecies 1.. Cronobacter turicensis sp nov. 190 . 2008. Int.. Kandhai. Grognou. Iversen. a new genus to accommodate the biogroups of Enterobacter sakazakii... 18. Microbiological Risk Assessment Series No. ICMSF. nov. 287:2204-2205. 14. A. 70:1579-1586. 15 Rome.. N. and S. Zwietering.. 2007. Microbiological Risk Assessment Series No.. Micro-organisms in Foods 7: Microbiological Testing in Food Safety Management. Iversen. JAMA. 2008. E. Microbiol. E. M. Reij. Stephan. Environ. Dublinensis subsp. Enterobacter sakazakii infections associated with the use of powdered infant formulaTennessee. and H. 2001. Joosten. Evol. Cleenwerck. Van Schothorst. lausannensis subsp. nov. Cronobacter sakazakii subsp malonaticus subsp nov. Food Prot. 6. thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant formula milk. nov. 72:2721-2729. Enterobacter sakazakii and Salmonella in powdered infant formula: meeting report. Effects of preculturing conditions on lag time and specific growth rate of Enterobacter sakazakii in reconstituted powdered infant formula. . J. and temperature. and M. C. Anderson. G... 38:378-382. Cronobacter muytjensii sp. Cronobacter dublinensis subsp. and proposal of Cronobacter sakazakii gen. M. 2004. Fanning. FAO/WHO. 2007. and Cronobacter dublinensis subsp. Lane. Italy. Himelright. Lett. Gurtler. D. Enterobacter sakazakii and other micro-organisms in powdered infant formula: meeting report.

Negredo. de Smet. A. J. P. 20. G. L. Oh. heat. J. R. 28. Bougatef. and S.. M.. Appl. Fanning. G. Healy. 24:9-13. Van Asselt. 116:73-81. J. Int. Kim. Effect of desiccation. 21. Microbiol. J. G. S. Zwietering. Whyte. 107:73-82. M. Muyldermans. J. B. 2001. Food Sci. Farber. Fanning. 2007. 2008. 14:875-877. 23. S. Int. and G. M. 1995. Jaspar.. L. Trends Food Sci. Wall. Roelofs-Willemse. 1988. R. Forsythe. Clin. Naessens. 72:107-113. J. Leguerinel.. Food Microbiol. Gorris. A. Mafart. and I. 74:139159. A systematic approach to determine global thermal inactivation parameters for various food pathogens. 6:41-46. H. Food Microbiol. Shaker. S. Food Sci. S. J. H. Int. J. Food Microbiol.. J. M.. Van Boekel. starvation. Nazarowec-White. M. Clin. S. T. Isolation and genotyping of Enterobacter sakazakii from powdered infant formula manufactured in Korea. M. A. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. Quinn. T. A. and L. 24. Ayyash. 26:743-746. J. Abu Al-Hasan. Lett.. Osaili. M. 2005. Couvert.. C. E. 191 . Appl. H. and S. Van Acker. and M. J. and C. Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula. 25. 39:293-297.Inactivation of Cronobacter spp. Microbiol. On calculating sterility in thermal preservation methods: application of the Weibull frequency distribution model. Food preservation by hurdle technology. 29. W. Leistner. Biotechnol. 48:536-541. Food Microbiol. Mullane. 1997. and J. Microbiol. 22. R. Anderson. 2002. Technol. Application of pulsed-field gel electrophoresis to characterise and trace the prevalence of Enterobacter sakazakii in an infant formula processing facility. O. Muytjens. K. Iversen. 2009. Yoo. 27. Prevalence of Cronobacter species (Enterobacter sakazakii) in follow-on infant formulae and infant drinks. 2002. Thermal resistance of Enterobacter sakazakii in reconstituted dried-infant formula.. Microbiol. Int. and S. and S. Lett. 26. N. Lauwers. F. S. P. 2006. Gaillard. P. On the use of the Weibull model to describe thermal inactivation of microbial vegetative cells. S. M.. 19. J. O' Brien. D. 73:M354-M359. H. and cold stresses on the thermal resistance of Enterobacter sakazakii in rehydrated infant milk formula.

Chapter 8 192 .

Discussion and concluding remarks 193 .

previously referred to as Enterobacter sakazakii. though notification of the illness is mandatory in a few countries. For the last decade there has been a strong and focused effort amongst all relevant stakeholders (i. academia. industry. Since an outbreak of Cronobacter infections occurred in the United States (Figure 1).Chapter 9 Introduction Cronobacter spp. there have been recalls of Cronobacter spp. 194 . as is the number of confirmed death (n=29). consumers) to establish harmonized safety criteria. All known species of Cronobacter have been linked retrospectively in either infants or adults to clinical cases of infection (24). governments. is not clearly understood. health case professionals. In this regard. the number of cases of confirmed Cronobacter infections in children worldwide is limited (n=156).. Like many other governmental agencies. 82). contaminated infant formulae in several countries around the world (28). Cronobacter spp.. Sub-populations at risk for Cronobacter infection can be found among infants and elderly consumers (53). is more and more widely recognized as an opportunistic foodborne pathogen that can potentially cause illness in vulnerable sub-populations of consumers. good manufacturing practices and sound guidelines for PIF preparation and use in order to protect the vulnerable consumers from the threat of infection with Cronobacter spp. Epidemiological studies have implicated powdered infant formulae (PIF) as a prime vehicle for possible exposure of infants to Cronobacter species (1.e. Cronobacter infection is a rare illness and no active surveillance systems exist. Although the natural source or habitat of Cronobacter spp. it is increasingly apparent that this group of micro-organisms is present in many different environments and can be found in a wide range of foods. may not appear to be as important a pathogen as many other foodborne pathogens that challenge the safety of our food to date. but because it particularly affects very young infants it is causing a high level of public concern and outrage. the United States Food and Drug Administration (FDA) issued an alert to health care professionals regarding the risk associated with Cronobacter infections among neonates fed with milk-based powdered infant formulae (26). As detailed in the Chapter 1. 77.

are described. associated with powdered infant formulae. in 2002 in the United States (14). is intimately associated to the ability to detect and identify the micro-organism in foods and food environments. The differentiation of Cronobacter spp. potential exposure is simulated for different scenarios of preparation and consumption of infant formulae. the progress in methodologies is first discussed. cloacae. Reported outbreak related to Cronobacter spp. new insights from the research reported here as well as from other sources as related to our ability to establish sound assessments of consumer exposure to the micro-organism. agglomerans). gergoviae.Discussion and concluding remarks Figure 1. E. belonged to the genus Enterobacter that contains species such as E. using the results of this research and relevant external data to compile particular simulations and to draw conclusions relating to prevention and control of the risk to infants of Cronobacter spp. (formerly E. Methodology Before the new systematic categorization and nomenclature as described in Chapter 1. Finally. 195 . among these species is based on biochemical reactions. Because the ability to prevent and control Cronobacter spp. Cronobacter spp. Pantoea spp. Following that. aerogenes and E.

in powdered infant formulae. sensitivity and robustness. Cronobacter spp. In the (selective) enrichment step the specific growth conditions of the target organism(s) are met while growth of other micro-organisms is inhibited to the greatest extent possible. in particular methods aiming to detect and trace Cronobacter spp. can be detected at 1 CFU/100g of PIF. methods for the isolation. detection and enumeration of Enterobacteriaceae. typically. Selective media developed using this feature include “Enterobacter sakazakii Isolation Agar” – ESIA (AES. sakazakii chromogenic plating medium) and Druggan-Forsythe-Iversen-medium (DFI) (37). In an enrichment procedure. is the presence of α-glucosidase (60).have been developed to various degrees of specificity. when 100 g is pre-enriched. Many new methods have been introduced and further refined over the last decade by academics and media manufacturers (Table 1). Sometimes a resuscitation step is used for the recovery of stressed cells or cells injured by exposure to chemical or physical treatments used in food production or processing (79).Chapter 9 and serological and molecular techniques (25). but only after induction of the enzyme by a suitable substrate (Chapter 2). An enrichment procedure may be an essential step in enabling detection of low contamination levels as it can promote the amplification of a single cell to levels of > 103 CFU/ml of enrichment broth. Over the years. All methods have a number of steps such as pre-enrichment and (selective) enrichment. This method is elaborated on in Chapter 2 and it has been used with a range of isolates obtained from environmental samples taken in several different (food) factories and food products (Chapter 3 and 4). thus reducing interferences from the food in the detection assay (9). France). Based on this feature and the fact that Cronobacter spp. other Enterobacteriaceae species may possess α-glucosidase after 24 hours. as compared to other Enterobacter species. and Salmonella spp. The biochemical characteristic of α-glucosidase production has remained the basis for many subsequently developed detection media. R&F® (E. early in this research a new practical method for screening Cronobacter spp. 196 . typically produce yellow pigmented colonies when grown on tryptone soy agar. a 1:10 dilution of the food matrix is made. A key feature for Cronobacter spp.. in foods and food environments was developed. Reducing the detection level has been an important aim for method development and currently Cronobacter spp. Notably.

Discussion and concluding remarks The initial detection methods established by the US FDA (2002) (78) and ISO (TS 22964) (36) are currently being re-evaluated. Moreover. which is followed by a selective enrichment broth prior to chromogenic agar plating. a fluorescence in situ hybridisation (FISH) technique has been described for the detection of Cronobacter spp. therefore. efforts have to be made to improve the cultural methods such that all strains are captured. that target 16S rRNA of the micro-organism are highly sensitive and useful for rapid and direct detection of Cronobacter spp. in foods. would be missed using only the general Enterobacteriaceae test procedures. Conventional methods are generally more time consuming than the molecular methods described more recently (2. at a population 197 . (40) showed that current cultural media do not support the growth of all Cronobacter strains and. 62. as knowledge regarding Cronobacter further improves. 85). 55. the conventional methods showed to be efficient even when low levels of Cronobacter cells were found in a sample (0. therefore. the international standard method for detection of Cronobacter from milk products issued by the International Organization for Standardization (ISO) and designed by the International Dairy Federation (IDF) involves a two-step enrichment (36). Most published methods to date include a pre-enrichment step. It was also shown that some Cronobacter strains (2%) (43) are unable to grow in standard Enterobacteriaceae enrichment (EE) broth and. However. they are rather qualitative and still rely on a time consuming (pre-)enrichment step. Improvements in detection methods for Cronobacter have been focussing much on optimising the isolation of the micro-organism from powdered infant formulae. the reliance on yellow-pigmentation (25) in ISO TS 22964 is no longer accepted as valid because a high proportion of strains (up to 25%) are nonpigmented on tryptone soy agar (6).. This method is based on binding of specific probes to target nucleic acid regions and is able to detect Cronobacter spp. The selective enrichment method used in this international standard method has been adapted from the one which has been developed in the research described in this thesis as described in Chapter 3.36 to 66 CFU/100 g) (59). For example. Recently. Iversen et al. for this method a population of 102 CFU/ml is required to obtain a 100% detection probability (21). 76. Although real-time PCR methods for specific detection of Cronobacter spp. Currently.

A key physiological parameter of importance to consider in further methodology development is the lag time (i.e. Further research on methodology should best focus on improving the enrichment step. including recovery from an injured state) of the target micro-organism. they need to be used in conjunction with culture based confirmation methods. Therefore. Note that an enrichment step is necessary to detect such low levels of the micro-organism. detection may offer advantages in identification of the micro-organism. the time required for adjustment of the microbial cells to a new environment after transfer from another. Further optimising the sensitivity and speed of detection remains a challenging task. A major drawback of applying molecular based detection technologies as compared to traditional culture based methods is that they can not clearly differentiate living from dead bacteria. while PCR methods for Cronobacter spp.Chapter 9 level of 10-2 CFU/ml after an enrichment of 8 h (2). as this so far remains an important bottle-neck. considering the relatively low level of the micro-organism amidst many other micro-organisms and given the heterogeneous distribution within a bulk production of PIF and the complexity and diversity of the food matrix itself. ideally providing a more specific and sensitive enrichment procedure. as it may add several hours to the overall detection time. Molecular based detection technologies offer opportunities for increased sensitivity but they may be sensitive to product matrix interference. 198 . Another disadvantage is that the micro-organism itself is not isolated and can not be studied further.

Discussion and concluding remarks 199 .

016.5% prevalence) and patients in the age group 61-70 year (51). Cronobacter spp. and cereals and also in various foods present on the Dutch market. could be isolated from a diarrhoeal child's stool (69) and from stool samples of hospitalised infants (0. as well as in foods of animal origin like milk. have been isolated from various environmental and clinical sources and a wide range of food products. Chapter 5 showed that the prevalence of Cronobacter spp. fruits and vegetables. have also been isolated from household cleaning cloths (52) and kitchen brushes (61). meat and fish and products made from these foods (27). Cronobacter spp. herbs and spices. 27. were isolated from scraping or sweeping surfaces in the production-line environment of factories producing milk powder. contamination in PIF and the PIF batches that are marketed generally comply with stringent safety standards established for the presence on Cronobacter spp. More recently. legume products. Cronobacter spp. it should be recognized that PIF can be contaminated with Cronobacter spp. 45. Several other studies have since investigated the prevalence of Cronobacter spp. with the presence of the micro-organism being confirmed in food ingredients from plant origin. 39. the bacterium 200 . Although Cronobacter spp. these studies have established the ubiquitous or widespread occurrence of Cronobacter spp. cereals. like cereals. Chapter 4 and 5 showed that Cronobacter spp. However. so far only powdered infant formulae have been epidemiologically linked with Cronobacter infections (1. through human vectors both during manufacture and preparation. and pasta.065. Overall. in a wide variety of products (4. at ~1%. 13.077). Notably. as it was earlier isolated from various clinical sources. including the production facilities for milk powder. in food and beverages other than infant formulae. chocolate. potato flour.Chapter 9 Occurrence of Cronobacter spp. 54). The isolation of Cronobacter spp. as well as in vacuum cleaner bags from domestic environments (Chapter 4). chocolate. on human skin and in human fecal samples may be rather low.07.061. PIF manufacturing industries around the world have tightened control of Cronobacter spp. Consequently. were also isolated from vacuum cleaner bags collected in households. from human samples is not new. These findings may suggest that humans may be carrier of the micro-organism without any symptoms. Cronobacter spp. including a child’s stool (25).. can be found in the production environment of various (dry) food products.

. with most isolates being detected in air filtering samples (42. 32. Results obtained by sub-typing of Cronobacter isolates revealed a genetic diversity of isolates. Staphylococcus spp. 48.Discussion and concluding remarks has been isolated from powdered infant formulae by several investigators (13. and Enterobacteriaceae.. depending on many factors relating to choice of ingredients. Some isolates recovered from the PIF final product reported were identical to those from the environmental samples on the basis of similarity of PFGE 201 . It is generally accepted that preventive measures implemented to control Salmonella spp. Micro-organisms that can occasionally be present.. this is the case for consumption of PIF by infants less than two months of age. especially those that are of low birth weight.. 59). in certain environmental samples from processing lines. and Salmonella spp. (23). Clostridium spp. especially Cronobacter spp. 83) have issued warnings to professional and domestic carers that the safety of powdered infant formulae needs to include careful handling and preparation of PIF for consumption. 57. including pathogens. in infant formulae production facilities (17). comprise the basis of possible control measures and are a prerequisite to address and control Cronobacter spp. throughout the manufacturing operation. such high prevalence figures may be due to ongoing improvements in enrichment and biochemical methods. Notably. 58) possibly due to its ubiquitous nature. 56). processing. Surveys in powdered milk facilities have identified the presence of Cronobacter spp. 22-24. Regulatory agencies worldwide (22. Cronobacter spp. While generally healthy infants should be at no elevated risk consuming product prepared according to instructions provided by manufacturers. Most notably. can not be totally eliminated from milk powder production facilities (30. and other manufacturing activities. premature and or otherwise immunocompromised. certain more sensitive infant groups may require that additional care is taken in the preparation and handling of infant formulae. 39. contrary to what can be achieved for Salmonella spp. occurring in PIF. 23. include Bacillus spp. Until now. Prevalence figures of 39% have been recorded for Cronobacter spp. It needs to be stressed that PIFs are not designed to be sterile products and that they can occasionally be contaminated with micro-organism.

prevention and control measures need to ensure that exposure of consumers to Cronobacter spp. regarding contamination of product ingredients and recontamination from environmental sources). Preventive and control measures need to be taken 1) during production of powdered infant formula (i. 2) during reconstitution of PIF (e. the environment and people handling the product in the course of the manufacture. Because of the importance to control infections caused by Cronobacter spp.Chapter 9 profiles.e. Contamination routes of Cronobacter spp. in the product ready for consumption is prevented or that its presence in the prepared product is controlled to a level that is considered acceptable for consumer safety. Future research on occurrence might focus on determining the spatial and statistical distribution of contamination. regarding recontamination from environmental sources and product preparation conditions) and 3) before the reconstituted formula is consumed (e. 2) through adding dry ingredients after the pasteurization step. Together. can persist in the production environment (30). and could be found over a year period in product samples.g. it is important that the occurrence of Cronobacter spp. 202 . from the environment or due to poorly cleaned feeding bottles or poorly maintained equipment in hospitals and homes. Regarding the possible occurrence of micro-organisms in PIF. regarding product handling and storage and consumption conditions). which may suggest that Cronobacter spp. 3) through contamination during reconstitution prior to feeding. potentially present in PIF. especially in PIF. it is worthwhile to discuss a number of characteristics/issues such as their survival in dry powdered infant formulae. is minimized.g. There are three main routes by which micro-organisms may enter PIF during its production and at reconstitution: 1) through exposing the product to contaminated equipment.

Cronobacter spp. can be present in various dry environments (46) and that Cronobacter spp. there has been an increase of Rapid Alert System for Food and Feed (RASFF) notifications concerning potential microbiological 203 .002 to 0. by applying appropriate control measures such as the application of adequate dry cleaning procedures within the processing environment and on the processing equipment. Acidification could reduce the concentration of Cronobacter cells in different types of infant formulae even further (44). therefore. strains could survive the spray drying process (3). such as vitamins. Pasteurization is indeed effective in destroying Cronobacter spp. or via the factory environment. 67).Discussion and concluding remarks their inactivation by pasteurisation and their growth in reconstituted infant formulae Such considerations can lead risk managers to specific advice to caregivers of at risk consumers. the incidence of Cronobacter spp.2 [-]. could increase after a wet cleaning process or due to uncontrolled presence of water. can lead to elimination of specific micro-organisms. in infant formula production factories should be reduced by effectively eliminating water throughout the manufacturing operation. 64. 18). cells during dry powdered infant formulae manufacture even if the thermal tolerance of individual Cronobacter spp. The micro-organism is clearly able to survive in very dry environments such as those with a water activity < 0. though in general the incidence of Cronobacter in PIF remains low with levels. Conceivably. Previous studies conclusively showed that Cronobacter spp. In 2007. isolates may differ (41. cells are able to survive for at least twelve months in powdered infant formulae (12. occasional contamination of PIF may occur after the pasteurization step. 63). Pasteurization. either via addition of contaminated ingredients. Recent studies showed that Cronobacter spp. Notably. such as health care professionals in neonatal care units. the incidence of Cronobacter spp. for instance following an emergency shower (17). reportedly between 0. have the ability to survive under dry conditions (8) as has been documented in the research described in Chapter 8 and it can do so for a long period. In factories.92 CFU/g (21. and to further improve the existing guidelines and handling practices (83). a heat process applied typically in the range of 60-80 ºC for a few minutes.

A factor possibly contributing to exposure of infants to Cronobacter spp. These parameters could be used to calculate the proliferation of Cronobacter spp. such as process equipment. It is likely that storage at ambient temperatures may lead to substantial growth. Topt= 39. in reconstituted infant formula at any temperature below the maximum growth temperature of 204 . and the possible outgrowth of the Cronobacter spp. such as resulting from detergents and disinfectants. The bacteria were furthermore found to be able to grow at 7 ºC. (50). as described further on. and kitchens sinks (33) and such biofilms may be a source of contamination. Reconstituted infant formula can either be used immediately or.. (20). resulting in Tmin= 3.4 ºC.Chapter 9 contaminations mainly due to some batches of vitamins contaminated with Cronobacter spp. PIF reconstitution and growth of Cronobacter spp. Biofilm may also be formed in enteral feeding.6 ºC.31 h-1. Previous results (Chapter 4 an 5) showed the ubiquitous occurrence of Cronobacter spp. refrigerated and used within 24 h (83). Cells of Cronobacter spp. Tmax= 47. can be covered. if indicated on the package. formation of biofilms may protect bacteria from aggressive conditions. 47. Biofilms may be present in a variety of environments. In fact. Disinfectants are used routenily in hospitals. day care facilities and food service kitchens. and the growth data obtained were used to develop predictive growth models for Cronobacter spp. with an optimum being between 37 and 43 ºC (39. may recover quickly after dry stress (23). Some Cronobacter strains can produce exopolysaccharides and thus it is possible that biofilm formation may occur on surfaces commonly associated with PIF feeding equipment (41. that has received rather little scientific attention is biofilm formation. The key growth parameters were quantified applying the secondary growth model of Rosso. in Chapter 6 the quick recovery of Cronobacter cells at 37 ºC is highlighted. opt= 2. however they have been shown to be quite inadequate in eliminating biofilms comprised of Cronobacter spp. In Chapter 6 the effect of preculturing conditions on the growth was investigated at several temperatures between 8 and 47 ºC. in reconstituted infant formula (Chapter 6) (23). 49).6 ºC. 63).

infections have occurred in both hospitals and outpatient settings.6 ºC. 67). exponential phase and stationary phase. the relative risk associated with the various preparation practices can be assesed. Based on the results in this thesis. Cronobacter spp. using cells with a different pre-culturing background (being pre-cultured to lag phase. 205 . Chapter 7 focused on the cooling time required for reconstituted infant formula to reach the set refrigerator temperature. 70).Discussion and concluding remarks 47. In practice. various practices are followed in hospitals and in the home. Studying the microbial growth of Cronobacter spp. preparation and use practices should be performed (31). 31. immediate feeding or storage for short or extended times with or without refrigeration. in reconstituted infant formulae.. respectively). 23. The growth parameters obtained from the research in Chapter 6. although this will not be easy due to initially very low levels. it was found that the various physiological phases of the cells did not significantly influence its growth in reconstituted powdered infant formulae. during the cooling process of prepared formulae irrespective of container volume. storage. Future research might focus on determining the lag phase of the organism in naturally contaminated powder. thereby likely leading to a reduction in risk as was also reported elsewhere (23. were combined with this model into one that allows predicting growth of Cronobacter spp. The model enabled to show quantitatively to what extent practical control measures such as proper cooling or limiting the time between preparation and consumption could help reduce growth of Cronobacter spp. or several feedings from a single container or a single feeding taking an extended amount of time (11. rewarming or feeding without re-warming. Many professionals and home cares for infants lack information on proper handling. including mixing PIF with cold or hot water. A model was established to predict the cooling profiles of reconstituted infant formulae given different containers or refrigerators used.

and providing clear information to caregivers for them to adequately control microbiological hazards in reconstituted formula (10). Reconstituting PIF with water of at least 70 ºC. and/or suitable indicator micro-organisms in order to reduce the risk of Cronobacter infection (23). 23). especially regarding tube feeding. possibly present (23. into in manufacturing environment and avoid their persistence/growth (10.  Improving product labeling. e. 23.  Educating health care professionals to provide effective hygiene training to parents and professional caregivers to ensure that PIF is prepared. − For hospitals:      Prevention of prolonged storage time of prepared formulae. These include the following: − For manufacturers  Utilizing good manufacturing and good hygiene practices to further minimize the entry of Cronobacter spp.Chapter 9 Risk mitigation options Several risk mitigation options have been proposed for various stakeholders (10. 83) in order to prevent Cronobacter infections due to consumption of PIF. in order to take benefit of the heat to inactivate viable Cronobacter spp. 206 . 67.g. Quick cooling of reconstituted formula using ice or running water and storing only small volumes of prepared formulae in refrigerators (83). Restricting feeding times. handled and stored properly (10). 67). Exclusively using sterilized liquid infant formula for high risk infants. − For governments and intergovernmental bodies:  Setting a microbiological criteria for Cronobacter spp. including harmonized guidance to create awareness among caregivers that PIF is not a sterile product and could cause illness in infants.

10% of the parents stored the bottle at room temperature and 3% in a bottle warmer. 16% of the parents used the stored formula within 4 h after preparation (11). 4) add the PIF. These observations 207 . 3) use boiled water of a temperature no less than 70 ºC. 6) check the temperature before feeding in order to avoid scalding the infant’s mouth. 5) quickly cool prepared formula to feeding temperature by holding the bottle under running tap water. Risk mitigation option 1: Preparation of PIF with water of at least 70 °C The Food and Agricultural Organization anf the World Health Organization (FAO/WHO) guidelines for safe preparation. 7) discard any feed that is not consumed within 2 hours after feeding. while in other countries. employing the estimated parameters and predictive models developed in this thesis (Chapter 6 and 7) to determine the effectiveness of the mitigation options in reducing absolute or relative risks.g. Furthermore. These guidelines have been implemented or advised in some countries. where relevant. Following the advice provided in FAO/WHO guidelines to reduce the risk of Cronobacter infection from PIF. such as in the Netherlands. 2) wash hands with warm water and soap before preparation. storage and handling of powdered infant formula (83) advise several steps to reconstitute PIF. A study in Italy on home preparation of PIF showed that only 22% (n=29) of caregivers use water of at least 70 ºC when preparing infant formula. caregivers are still advised to prepare one or two bottles in advance following the instructions on the label. e. or by placing in a container of cold or iced water. If prepared formula is not to be consumed within two hours after preparation it should be quickly cooled immediately after preparation and stored in a refrigerator (at a maximum temperature of 5 ºC) and used within 24 h (83). reconstituting PIF with water > 70 ºC should result in at least a 4-log reduction of Cronobacter spp. namely 1) heat-treat the bottle at each feeding. and place the bottles immediately in a refrigerator with a temperature of no more than 4 ºC with a maximum storage time of 8 h (81). Italy (11) and Ireland (75).(67).Discussion and concluding remarks In the next paragraphs four mitigation options are discussed in some more detail. after preparing a bottle.

which may cause foodborne illnesses in infants such as diarrhea (73. it is advised to clean bottles using a dishwasher with a washing temperature of at least 55 ºC or using warm soapy water in combination with a dish brush (80).Chapter 9 indicate that the guidelines are implemented only by a certain percentage of the caregivers. Notably. it may take quite some time and effort for training caregivers to properly implement best practices in reliable way. such as B. the caregiver or the infant may be wounded by scalding. left at room temperature (70) and not properly cleaned and heat inactivated before re-use. Thus. Infant formulae are frequently contaminated with low numbers of Bacillus spores (5). more research is encouraged to investigate the effect of using hot water for the preparation of PIF on sporulation of spores. In the Netherlands. It has been reported that reconstitution at temperatures ranging from 65 to 95 ºC may lead to germination of endospores of B. Furthermore. This may allow Cronobacter spp. It would be relevant to dedicate further research on suitable control measures to minimizing an impact on vitamins and to minimizing the risk of scalding. A key beneficial effect of using water of at least 70 ºC is the on-site pasteurization of the bottle. cereus. to outgrow to hazardous numbers (Chapter 6). the high temperatures as recommended may encourage the proliferation of other bacterial cells. Note that Cronobacter spp. A potential disadvantage of reconstitution with hot water is its effect on heat sensitive vitamins. cereus. germination of spores and possible toxin production. 74). the effect of the high temperature on dissolving the powder should also be considered. Evedently. Baby bottles may be heavily contaminated if small amounts of infant formulae are not immediately discarded after feeding. For this reason. Another important factor that needs to be taken into account is the potential activation of Bacillus spores. 208 . may be introduced into a bottle due to its presence in for example a contaminated dish cloth or dish brush used for cleaning.

Discussion and concluding remarks Risk mitigation option 2: The relative risk reduction achieved when reconstitution is not with water of 70 °C but with water of 65 or 60 °C The expert consultation of the FAO/WHO (2006) pointed out that reconstitution with liquid of at least 70 ºC is an effective risk mitigation strategy (23). Where rapid cooling was part of the option it was part of the refrigeration step. b Refrigeration conditions used in the various scenarios included immediate refrigeration after preparation in still air or convection refrigerator. 209 . the following scenarios were calculated: use of water of 40. Preparation and storing conditions for scenario’s (Room) temperature [ºC] Time [h] Preparation Refrigeration Re-warming Feeding a 25a b 0. The initial Cronobacter spp. Table 2.5 6 37 30 Please note that the temperature of the water used for reconstitution is between 40 and 70 ºC whereas preparation was at room temperature is 25 ºC. the temperature of the formula was calculated by the web-tool at successive time points between preparation and consumption.25 8 0.5 0. Furthermore. 60. as shown in Table 2. The web-based tool of the risk assessment model of the FAO/WHO (68) was used to assess the relative risk associated with the various preparation and handling scenarios. the effect of rapid cooling on the relative risk was estimated. concentration was assumed to be 1 CFU/g. 50. 65 and 70 ºC to reconstitute the powder in either a can or a bottle and to store it in both a refrigerator with still air or convection. For the given preparation and storage scenarios. A relevant risk management question then is: how much less risk reduction is achieved at lower reconstitution temperature? In order to investigate the effect of different water temperatures. The baseline scenario for relative risk calculations was preparation of the formula with water of 40 ºC in a bottle and storing the bottle in a refrigerator with still air at 6 ºC for 8 hours.

Note that these comparisons are made in relation to the baseline scenario. Figure 2 shows the estimated relative risk for the various scenarios. The value of the dose-response parameter r has not been established with any certainty. Values between 10-10 and 10 -5 have been suggested (23). it is assumed that the risk (P) and the dose (D) are linearly correlated. with a reduction for all storage scenarios involving bottles but with increased risk for formula preparation in a can followed by storage in a still air refrigerator. growth or inactivation rates were estimated and the resulting cell concentration calculated.Chapter 9 From these. The probability of infection (P) resulting from consumption of an estimated average dose of Cronobacter spp. the duration of the lag time. The relative risk increases a lot (247%) when a can volume is used rather than a bottle. When water of 50 ºC is used for reconstitution in a bottle. even without knowing the exact value of the r value a relative risk can be determined.  1  exp  rD P   (1) where r [1/CFU] is the exponential dose-response parameter and D [CFU] the estimated dose consumed. there is a slight reduction in estimated risk for reconstituting PIF in the bottles for rapid cooling and convection refrigeration compared with the baseline scenario. 210 . can be described by the exponential dose-response model (Equation 1) (71). could lead to a decrease in relative risk of 68% compared to cooling in a refrigerator with still air. Therefore. For the low values of the dose (D). the relative risk increases (32%) for the scenario with still air and the bottle while it does not change for convection refrigeration but it is greatly reduced (38%) in the case of rapid cooling (for example using ice or running water). At a reconstitution temperature of 40 ºC. Reconstitution of PIF in a can shows an increase (100%) in the relative risk. Rapid cooling of PIF reconstituted at 50 ºC in a bottle. Reconstituting PIF with water of 60 ºC gives a variable effect in risk.

2. where a 1. Note that the chosen setting (at 40 ºC). as this action would yield at least a 4 to 6-log reduction of Cronobacter spp. reconstituting PIF with water of more than 70 ºC may have a safety margin. storage and handling practices for prepared formula as simulated using the web based tool (68). Reconstituting at a temperature somewhat colder than 70 ºC. However. when the temperature of the water is 60 ºC. 65 and 70 ºC.26-log reduction of Cronobacter populations was obtained (67). as shown in Table 2. At the higher temperatures. instead of 70 ºC. there was no apparent effect of convection and rapid cooling for both the bottle and can on the relative risk. subsequently in relative risk reduction. still can be expected to achieve a reduction in Cronobacter.Discussion and concluding remarks Figure 2. In all cases a similar risk reduction of at least 4 log units or more was observed. was used as baseline scenario for comparing relative risk levels. Results from these simulations can be compared with reported reductions achieved when water of 60 ºC and 70 ºC was used to reconstitute PIF. is achieved and the relative risk may even be increased compared with the baseline scenario in case PIF is prepared and stored in a can. depending on the type of formula used and the type of Cronobacter strain present. Therefore. a slight inactivation of Cronobacter spp. 211 ..and 5. Increase in relative risk of different preparation.

212 . Considering the advice from the FAO/WHO expert meetings (22. The growth of Cronobacter spp. which can be in place for various time periods. Research has shown that Cronobacter spp. will multiply in the tube. can form biofilms on plastic surfaces (41) and attach to enteral feeding tubes at ambient temperatures (49). or may accommodate microorganisms as these may possibly attach and form biofilms. which is at about body temperature and also in the feeding container (syringe) which is at ambient temperature.Chapter 9 Risk mitigation option 3: Impact of hang-time in case of continuous tubefeeding on consumer risk The WHO guidelines for safe preparation. which are the typical weak and the group at greatest risk (84).88 as described in Chapter 7. tube-fed formula is refreshed every 2 to 3 hours (34). The increase in cell count was predicted with the exponential growth function. using the specific growth rate calculated with the secondary growth model of Rosso with parameters as determined in Chapter 6. storage and handling of PIF do not consider the use of enteral feeding tubes. It is therefore reasonable to assume that Cronobacter spp. 23) and the EFSA panel (19) to limit the hang time of tube-fed prepared infant formulae to 4h. the impact of the following scenarios on the level of Cronobacter spp. The feeding time was simulated up to 6 hours in a warm room (35 ºC) and in a cold room (20 ºC). was estimated: the feeding container was assumed to be a baby bottle containing 120 ml of reconstituted infant formula was assumed to contain an initial concentration of Cronobacter spp. the estimated parameters for the baby bottle from Chapter 7 were presumed to be similar for the feeding container. The tube feeding started immediately after preparation and the formula was prepared with lukewarm water (40 ºC) in a baby bottle. was predicted. Therefore. at a relatively high temperature (warm rooms may be 35 oC) and therefore may be a cause of large outgrowth of bacteria. ranging from < 6 h to > 48 h (35). Enteral feeding tubes may be used to especially feed low birth weight infants in neonatal care units. Depending on the regime in the hospital. both neglecting the lag time and including the lag time according to the optimal fit of the k-value: k =  x λ = 2. of 1 CFU/ml.

88. Including the lag time.Discussion and concluding remarks Figure 3 gives the temperature profile. As can be seen. This prediction must be assumed to be an overestimation. When considering a lag time. while the temperature profile and predicted increase in cell count during tube feeding in a warm room are depicted in Figure 4. together with the predicted increase in cell count during tube feeding in a cold room (20 ºC).. Tube feeding in a warm room (35 ºC). which neglects lag time. growth will start after 2 hours. an increase of at least 4-log units is predicted with the exponential growth function. while cell numbers increase in time. overestimates growth as it predicts the increase to start immediately. These predictions show that the room temperature (Figures 3 A and 4A) and thus. as lag time is neglected (Chapter 7). an increase of 2. the decrease in temperature from 40 to 20 ºC is slow and takes more than 6 hours. The exponential growth function. The predicted growth shows a 1-log unit increase in 4 h.5 log units is predicted in 4 h. may result in a considerably larger increase in cell numbers as can be seen in Figure 4. the temperature of the formula being fed may significantly affect growth of Cronobacter spp. as estimated with a k-value of 2. and needs to be considered for tube feeding infant formula to sensitive infants. After 4 hours. 213 .

Chapter 9 Figure 3. Temperature profile (A) and the number of Cronobacter spp. black line = prediction with optimal fit k=2. (B) in a bottle used for tube feeding in a room which is at 20 ºC. 214 . Grey line = prediction with exponential growth function.88.

215 . (B) in a bottle used for tube feeding in a room which is at 35 ºC. black line = prediction with optimal fit k=2. Grey line = prediction with exponential growth function.88.Discussion and concluding remarks Figure 4. Temperature profile (A) and the number of Cronobacter spp.

and neonates. the group at greatest risk. in both homes and household settings. a mortality reduction of 90% (26/29) would be achieved. low birthweight (LBW) infants (< 2500 g). and the motility by 52% (15/29). low birthweight infants. 24). It is assumed that all infants have been formula fed only and that all cases were due to the presence of Cronobacter spp. would possibly lead to a 100% reduction of Cronobacter infection cases for this group and a 72% (111/155). If all neonates. 23).Chapter 9 Risk mitigation option 4: Use sterile formula for the population at greatest risk. Using liquid sterile formula for all infants of the group at greatest risk. and immuno-compromised infants. both assumptions are conservative and may cause overestimation of risks. Table 3 gives an overview of data on reported cases adapted from the FAO/WHO expert meetings. because they may specifically require infant formula and be more susceptible to infection (10. Infants of HIV-positive mothers are also at risk. and those infants that are less than 2 months of age. The two FAO/WHO expert meetings have established that all infants at or below 12 months are at particular risk for Cronobacter infections (22. Cases are divided in subgroups. Subsequently. 216 . those at greatest risk are neonates (< 28 days). particularly pre-term infants. are fed with sterile formula only a 48% (75/155) reduction of Cronobacter infections and a 24% (7/29) reduction in deaths would be reached. reduction in the total group of infants. It was estimated what the risk reduction could be of providing sterile formula to the whole or to a subset of the group at greatest risk. being all infants. Feeding only the LBW group with sterile formula is expected to have reduced the morbidity of Cronobacter infection with 37% (57/155). in PIF. Among these infants. A 100% risk reduction was assumed when using sterile formula for each individual infant.

in reconstituted infant formula is possible. Most of reported Cronobacter infections are related to PIF. also other strategies. Data adapted from WHO/FAO 2008 report (24) No.Discussion and concluding remarks Table 3. is an ubiquitous pathogen. the results described in this thesis can be used to estimate the exposure and thus the 217 . feeding the group particularly at risk and the at greatest risk group with sterile formula. For this reason. Estimation of the risk reduction. Overall. Although prematurity and low birth-weight are the most important risk factors for diseases. including PIF. as results in this thesis showed that potential rapid growth Cronobacter spp. should be considered to address effect in preventing illness. especially for the LBW group a larger reduction in the mortality rate was estimated compared to the neonate group. care should be taken while preparing. cases Recovered Outcome unknown Number of deaths All infants (≤ 12 months of age and including age not known)a Group indicated to be at greatest risk (≤ 2 months)a LBW group (< 2500g) Neonates (< 28 days) a 155 111 57 75 10 85 71 38 49 2 41 14 4 9 6 29 26 15 7 2 Age at moment of illness not known a in all groups is age indicated to be age at onset of the Cronobacter infection. It was isolated from a wide variety of of food production environments and food products. because of its widespread occurrence. Therefore. showing that young infants are at greatest risk. although infections were also reported in adults. results in this thesis showed that Cronobacter spp. it should be considered that PIF is not a sterile product. These estimates show that it would be worthwhile to consider using sterile formula for the group at greatest risk. like necrotising enterocolitis in newborn infants. depending on storage and feeding practices. storing and handling PIF. next to using sterile formula. Concluding remarks In conclusion.

(Enterobacter sakazakii) in different categories of ready-to-eat foods other than infant formula.. 39:293-297. Jordan. Food Prot. W. Forsythe. J. Shapiro. Int. Acker van J. Grand. Becker. Microbiol. Home preparation of powdered infant formula: is it safe? Acta Paediatr. W. Jaykus. Eur. Clin. J. Minster. Tyburski. 95:967-973. N. Liniger. J. Bougatef. S. 11. A. 2. 3.. Caubilla-Barron. 72:1774-1789. Young. Keevil. J. Carletti. 2003. Fanning. Maladen. 2008.. 2006.. 89:1309-1316. Terplan. Cluster of neonatal infections in Jerusalem due to unusual biochemical variant of Enterobacter sakazakii. Almeida. N. Azevedo. Development and application of a novel peptide nucleic acid probe for the specific detection of Cronobacter genomospecies (Enterobacter sakazakii) in powdered infant formula. 6. 61:102-108. C. Fanning. Peleg. Int.. References 1. Baumgartner. 75:2925-2930. B. and G.. Food Prot. A. and S. M. J. Detection and frequency of Cronobacter spp. C. Leclercq. Dis.. J. Iversen. Joosten. Lauwers. and K. 21:613-616. 7. I. Bacillus cereus in infant foods and dried milk products. Microbiol. Lardeau. A. Evaluation of the International Organization for Standardization-International Dairy Federation (ISO-IDF) draft standard method for detection of Enterobacter sakazakii in powdered infant food formulas. Peterz. 2008. E. and M.. 218 . 1994. Lombard. 9.Chapter 9 probability of illness due to Cronobacter spp. Desiccation and heat tolerance of Enterobacter sakazakii. 136:189-192. Appl. . Dairy Technol. Mullane. Int. A. CAC. and B.. Tortorello. C. Arad. and M. C. Microbiol. 5. S. C. O. Code of hygienic practice for powdered formulae for infants and young children. Appl. 2002. and M.. G. J. Schaller. 4. Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula. A. CAC/RCP 66-2008. G. 2009. J. J. Besse. N. A. G. 8. L. J. and C. 2008. Clin. Block. V. C. B. H. A. Vieira. AOAC Int. Von Wiese. 2009. Infect. 97:1131-1132. Dry stress and survival time of Enterobacter sakazakii and other Enterobacteriaceae in dehydrated powdered infant formula. L. J. 2001. 10. 2009. Arku. Iversen. Muyldermans. and H. Fox. Enterobacter sakazakii survives spray drying. Simhon. 23:1-15. Brehm-Stecher. Cattaneo. Environ. C. after the consumption of reconstituted infant formulae. Breeuwer. N. M. Sample preparation: The forgotten beginning. 2007. Food Microbiol.. and A. Microbiol. 70:2111-2117. P. and S. F. Food Microbiol. 12. M. J. Naessens. de Smet F.

Discussion and concluding remarks 13. M. T. O. J. Infect. Production of powdered infant formulae and microbiological control measures. Int. Fanning. 16. EFSA. 2008.C. 21. 2005. C. Cawthorn. Cooksey. 18. 2004. Forsythe. M. G. AOAC Int. and R. S. 145-185. C. 15 Rome. and K. Microbiological Risk Assessment Series No. Buchanan. FAO/WHO. FAO/WHO. Y. Food Prot.. Farber and S.. 65. Rome. 27-59. 68:1900-1902. S. MMWR. The EFSA Journal. Enterobacter sakazakii (Cronobacter spp. 2001. L. L. Enterobacter sakazakii (emerging issues in food safety) ASM Press. C. D. A. Enterobacter sakazakii and other micro-organisms in powdered infant formula: meeting report.C. K. J. CDC. N. Washington. 113:1-35. C. 6. S. C. 15. 2009. Diagn. J. p. Clark.) in powdered followup formulae: meeting report. and R. 13:467-472. Washington. 92:862-872. Italy. Evaluation of a revised US Food and Drug Administration method for the detection and isolation of Enterobacter sakazakii in powdered infant formula: pre-collaborative study. S. New Jersey. Evaluation of different methods for the detection and identification of Enterobacter sakazakii isolated from South African infant formula milks and the processing environment. 2006. New Brunswick.. Epidemiologic typing of Enterobacter sakazakii in two neonatal nosocomial outbreaks. 10 Rome. J. The Rapid Alert System for Food and Feed (RASFF) Annual Report 2007. Dis. Opinion of the scientific panel on biological hazards on the request from the commission related to the microbiological risks in infant formulae and follow-on formulae. Enterobacter sakazakii infections associated with the use of powdered infant formula . In: J. Microbiol. Lampel. 22. p. J. Italy.Tennessee. Hammack. Enterobacter sakazakii (emerging issues in food safety) ASM Press.). Microbiological Risk Assessment Series No. New Brunswick. 2008. Chen. Witthuhn. D. Isolation and identification of Enterobacter sakazakii. 2008. 2002. 17. 14. European Commission. Microbiological Risk Assessment Series No. Italy. p.. and R. 2008.J. and S. B. In: Official Journal of the European Union.). 19. J. 219 .M. 51:297-300. Hill. 127:129-138. 2004. In: J. Survival of Enterobacter sakazakii in a dehydrated powdered infant formula. FAO/WHO.. Y. Food Microbiol. 24. Cordier.. Song. New Jersey. Farber and S. 2007. Edelson-Mammel. Steingrimsson. 1990. 23. M. Luxembourg. Botha. O'Hara. Enterobacter sakazakii and Salmonella in powdered infant formula: meeting report. Forsythe (eds. Forsythe (eds.. 20. D.

fda. Health Res. K. J. J. Int. F. and S. ISO. Friedemann. and S. Kandhai. A simple and rapid cultural method for detection of Enterobacter sakazakii in environmental samples. 30. Forsythe. Hood. 2009. Schoder. Int.. Friedemann. 35. J. Neonatal enteral feeding tubes as loci for colonisation by members of the Enterobacteriaceae. 27. 1980. J. E. 32:1730. E. 68:64-69. Food Microbiol. D. 29. 2009. W. Path. Caubilla-Barron. A pilot study describing infant formula preparation and feeding practices. Grant. 9:146. S. 6:9-18.. 28. Food Prot. Gadzov. J. S. Sonnard.. Brenner. I. Kodde. Zwartkruis-Nahuis. Environ. B. ISO/TS 22964 and IDF/RM 210-first edition 2006-02-01: Milk and milk product. Kucerova. D. J. B. Microbiol. 28:12971304. M. Int. Zottola. H. J. Enterobacter sakazakii in food and beverages (other than infant formula and milk powder). 2006. Hurrell. Malorny. and E.Chapter 9 25. A. C. E. Biofilm in food processing. Foissy. Int. and M. 33. 31. F. Herbold. M. Food Control. Hurrell. and E. Scott. J. O. A. J. and H. Clin. A. 116:1-10. Dis. J.htm. M. M. de Boer. Salmonella serovars and other Enterobacteriaceae. Kucerova. 2009. 6:225-233. M. Joosten.. 2005. J. 2002. J. 220 . Hilton. Epidemiology of invasive neonatal Cronobacter (Enterobacter sakazakii) infections. V. J. Wagner. D. Loughlin.gov/Food/FoodSafety/ProductSpecificInformation/InfantFormula/ AlertsSafetyInformation/ucm111299. 11 April 2002. Date. FDA. Syst. Enterobacter sakazakii: A new species of "Enterobacteriaceae" isolated from clinical specimens. D. T. and the Enterobacteriaceae study group. N. M. 2007. 18:541-459. Eur. Heuvelink. M. Loughlin. BMC Infectious Diseases. 1995. Armstrong. A. E. 32. Hickman. Ahmed. and E. Health professionals letter on Enterobacter sakazakii infections associated with use of powdered (dry) infant formulas in neonatal intensive care units. Caubilla-Barron. Bacteriol. Accessed 5 August. Hein.. M. Food Microbiol. Temporal and spatial distribution of Cronobacter isolates in a milk powder processing plant petermined by Pulsed-Field Gel Electrophoresis. J. De Ware(n)chemicus. E. Smith. Guillaume-Gentil. 34. Enterobacter sakazakii in melkpoeder. Marugg. Biofilm formation on enteral feeding tubes by Cronobacter sakazakii. J. Foodb. Asbury. R. 2009. Dis. Forsythe. 136:227-231. 26.. 36. Available at: http://www. J.Detection of Enterobacter sakazakii. Shoo. Farmer III. 2009.. C. Infect. H.. 2008. 30:569-584.

C. Forsythe.. Appl. Lardeau. 17:87-92. M. 49. Kandhai. Lane. Kandhai.. J. 60:66-72. Van Schothorst. Occurrence of Enterobacter sakazakii in the food production environments and households. 363:39-40. P. H. G. K. M. W. Forsythe. Schumacher.. C. J. Food Sci. Feer. Archiv. 2004. J. and R. M. Environ. 15:152-157. Jung. R. J. 39. Iversen. Appl. Reij. 21:771-777. The growth profile. and A. 2004. and L. Microbiol. J. Biotechnol. T. thermotolerance and biofilm formation of Enterobacter sakazakii grown in infant formula milk. Iversen. Zwietering. C. 2007. C. Microbiol. 2006. 125:344-346. Appl. Comparison of media for the isolation of Enterobacter sakazakii. Beumer.Discussion and concluding remarks 37. Iversen. Int. C. Development of a novel screening method for the isolation of "Cronobacter" spp. M. Reij. Stephan. and S. Kandhai. Food Microbiol. 72:5846-5856. Enhanced microbiological safety of acidified infant formulas tested in vitro. Iversen. Microbiol. Joosten. H. Joosten. Forsythe. M. Genotyping of Cronobacter (Enterobacter sakazakii) strains isolated from an infant formula processing plant. a preliminary study. A selective differential medium for Enterobacter sakazakii. A. Int. Food Prot. Puyvelde. P. C. A new protocol for the detection of Enterobacter sakazakii applied to environmental samples. Microbiol. A rapid and reliable alternative to ISO 21528-1 : 2004 for detection of Enterobacteriaceae. Lett. Fricker-Feer. 2004. O. Food Microbiol. Fur Lebensmittelhygiene. 38:378-382. 46. Iversen. M.. R. K. R. Food Microbiol. Guillaume-Gentil. H. Ryu. Gorris. C. Jackson. J. 96:133139. Druggan. Effects of preculturing conditions on lag time and specific growth rate of Enterobacter sakazakii in reconstituted powdered infant formula. 43. 67:1267-1270.. and M. (Enterobacter sakazakii). C. Gorris. 74:2550-2553. and J. Iversen. Appl. Grognou. L. K. Forsythe. v. and M. Prevalence and thermal stability of Enterobacter sakazakii from unprocessed ready-to-eat agricultural products and powdered infant formulas. 2008. M. Environ. Stephan. H. 2004. 40.. SAJCN. Appl. Microbiol.. Environ.. 72:2721-2729. 38. Druggan. O. J. Gschwend.. and S. Reij. 2009. H. and R. M. Environ. J. A. 221 . Gschwend. Van Schothorst. H. C. K. Isolation of Enterobacter sakazakii and other Enterobacteriaceae from powdered infant formula milk and related products. Guillaume-Gentil. 2004. H. 2006. W. Attachment of and biofilm formation by Enterobacter sakazakii on stainless steel and enteral feeding tubes. A. C. 47.. 41. M. The Lancet. 45. M. L. Stephan. M. Marugg. and S. van Schothorst. 42. and S. 44. 48. 2004. Joosten. Lehner. G.. C. S. 2006. Beuchat. R.. M. and M. 2008. Lehner. Kim. 73:48-52. Klijn. C. Park. J. W. Iversen.. and C.

J. C. 55. and in a biofilm. 116:73-81. D. J. and adults .relevance to exposure assessment. S. P. Environ. B. Fanning. and J. and M. 73:1256-1265. H. Effectiveness of disinfectants in killing Enterobacter sakazakii in suspension. Prentice. H. Mullane. 58. N. Mullane. and S. 53. Medicine. C. and S. Analysis of eight cases of neonatal meningitis and sepsis due to Enterobacter sakazakii. dried on the surface of stainless steel. Whyte. Appl. Sonderkamp. Wall. K. and S. N. Clin. Microbiol. Lee. Clin. Y. Kusumaningrum.. Clin. J. B. J. Food Microbiol. 2007. Mullane.. and P. R. J. Surveillance of stool samples for the presence of Enterobacter sakazakii among Korean people. Whyte. S. H. and G. Appl. Park. 51. H. D. J... Kim. Leclercq. Lee. Fanning. 1988. N. P. J. Environ. H. Microbiol. 60. J. Lai. A. 74:5913-5917. Muytjens. 56. A. L. P. Wanegue. Van Druten. 222 . A. Wall. Wall. H. T.Case reports and a review of the literature. J. Microbiol. H. 26:743-746. Microbiol. Ryu. 18:115-120. Food Prot. Quinn. 72:63256330. Roelofs-Willemse. Wageningen. Microbiol. H. Kollee. Behaviour and cross-contamination of pathogenic bacteria in household kitchens . P. L. Farmer. 2002. 2008. 2005. Laboratory of Food Microbiology Wageningen University. Cho. Muytjens. 1984. Appl. and H. Appl.. Muytjens. Dissemination of Cronobacter spp. Environ. and L. R. 2006. infants. G. Detection of Enterobacter sakazakii strains by real-time PCR. Microbiol. 61. B. J. Microbiol. L. 2001. 80:113-122. 2007. Kim. H. Beuchat. N. Fanning.. H. L. In: PhD Thesis. Parton. The Netherlands. Murray. Meade. B. 2003. Quality of powdered substitutes for breast milk with regard to members of the family Enterobacteriaceae. I. A. B. 2008.. 1983. 3rd. 20:684-686.Chapter 9 50. J. 68:1623-1627. G. Malorny.. Int. Environ. J. K. Enterobacter sakazakii infections among neonates. Baylac. Detection of Enterobacter sakazakii in dried infant milk formula by cationic-magnetic-bead capture. Yonsei Medical Journal. K. 54. C. Healy. Enzymatic profiles of Enterobacter sakazakii and related species with special reference to the alpha-glucosidase reaction and reproducibility of the test system. Wachsmuth. Application of pulsed-field gel electrophoresis to characterise and trace the prevalence of Enterobacter sakazakii in an infant formula processing facility. Comparison of fecal coliform agar and violet red bile lactose agar for fecal coliform enumeration in foods. Kim. children.. and H. B. 59. G. Van der Ros-van de Repe. Zanen. J. J. Lee. H. P. 52. 57. Chae. Jaspar. Whyte. (Enterobacter sakazakii) in a powdered milk protein manufacturing facility. K. P. J.. 49:1017-1022. 68:1631-1638. R. Drudy. H. K. Wagner.

S. Food Microbiol. Appl.Discussion and concluding remarks 62. Heat resistance of Cronobacter species (Enterobacter sakazakii) in milk and special feeding formula. 64. A. Farber. and J. M. W. Aggarwal. L.. J. occurring in powdered infant formula. and M. C. C.aspx.A. Milk Sci. and R. Environ. and P. Frampton. B. Holley. Accessed 23 October 2009. 2009. Overview of a risk assessment model for Enterobacter sakazakii in powdered infant formula. Negredo. Food Prot. Farber. Reij. (Enterobacter sakazakii) in powdered infant formula. K. R. Lionberg. 63. J. Fanning. M. G.. survival. and environmental sources. J. 62:266-269. 67. The importance of hygiene in the domestic kitchen: Implications for preparation and storage of food and infant formula. N. 107:928-935. 2006. 1997. Perspectives in Public Health. Gorris. G. I.. E. Food Prot. Nair. L.. M. J.. O'Brien. 1997. A.. Microbiol. Lett. J. and C. Healy. Al-Haddaq. Venkitanarayanan. Jongenburger. Griffith.. 72. and J. S. Appl. Evaluation of a new one step enrichment in conjunction with a chromogenic medium for the detection of Cronobacter spp. Incidence. G. 66. Al-Nabulsi. Schwalbe. J. 223 . Rowan.. E. M. Isolation and characterization of Enterobacter sakazakii from milk foods and environment. 1998. 70. 2009. and R.. Pediatr. and J. M. K. and growth of Enterobacter sakazakii in infant formula. Paoli. 72:2539-2546. S. 1990. Restaino. W. 73. Osaili. 2006. 2009. 2009. J. Redmond. 60:226-230. E.S. 61:196-200. Int. Microbiol.org/esakmodel/ESAKRAModelWizard. 69:315-322. Appl. Martin. R. 65.. J. Prepared for WHO/FAO. M. Zwietering.. Kotloff. M. M.. W. Effectiveness of cleaning and disinfection procedures on the removal of enterotoxigenic Bacillus cereus from infant feeding bottles. A. Noriega.136:232-237. and C. and K. M. H. Gkogka. Nazarowec-White.. Food Prot. and E. J. 71. Raghav.. Microbiol. Available at: http://www. F. S. Shaker. Food Prot. M. Perspective on the risk to infants in the Netherlands associated with Cronobacter spp. Thermal resistance of Enterobacter sakazakii in reconstituted dried-infant formula.L. Nosocomial bacteremia caused by Enterobacter sakazakii and Leuconostoc mesenteroides resulting from extrinsic contamination of infant formula. 69. C. 2006. Infect. Dis. 2007. R. 129:69-76. Int. 68. A chromogenic plating medium for the isolation and identification of Enterobacter sakazakii from foods. K.mramodels. Hartnett. Anderson. Cloning and Sequencing of the ompA gene of Enterobacter sakazakii and development of an ompA-targeted PCR for rapid detection of Enterobacter sakazakii in infant formula. M. Becker. T. M. J. Iversen. 72:1472 -1475. food ingredients. 9:447-449. Milchw. Nazarowec-White. 24:9-13. R.

Food Microbiol. 317-328. Dev.int/foodsafety/publications/micro/pif_guidelines. Available at: http://www. Rowan. X. H. and A. Accessed 10 February. 83. Voedingscentrum. M. 2009. Londen. 224 . and J. 85. and W.voedingscentrum. Q. Safefood and Health Service Executive.aspx.Chapter 9 74. Hugo (eds. Resuscitation of injured bacteria in foods. Enterobacter sakazakii infections in neonates associated with intrinsic contamination of a powdered infant formula. Yang. Flesvoeding.. 79. 2009. Y. Academic Press. N. and R. and J. B. J.. H. 166:1570.safefood. P. (Londen).voedingscentrum. Accessed 13 February. Ferguson. Bacteriological quality of infant milk formulae examined under a variety of preparation and storage conditions. WHO. 2007.). 2002.pdf?epslanguage=en. W. B. storage and handling of powdered infant formula. Rapid. 2008. Simmons. M. Metts. Xu. 2009.. In: Skinner F. Can. Med. Early Hum. Gelfand. J. Wu. 2005. Anderson. p. 75. 10:398-401. specific detection of Enterobacter sakazakii in infant formula using a real-time PCR assay. Assoc. 60:1089-1094. Accessed 30 September 2009. K. Klaarmaken van flesvoeding. 2008.. Avalaible at: http://www. M.nl/nl/eten-gezondheid/zwangerschap-enkind/flesvoeding/klaarmaken-van-flesvoeding. 1976. Y. Haas. 1989.eu/Global/Publications/Consumer%20information/PIF%20le aflet. Ltd. Ye. J. 2009. Epidemiol. Van Schothorst. A. Brackett. 1999. Accessed 21 September 2009. How to prepare your baby's bottle available at: http://www. L. Y. P. 56:89115. 1997. 25:648-652. J. Inhibition and inactivation of vegatative microbes. V. Infect. 80. Accessed 28 September 2009.aspx. 68:59-63.. Anderton. Control Hosp. 84. Zhou. Available at: http://www. 78. Enteral feeding in the pre-term infant. U. K. Seo. Development of an immobilization and detection method of Enterobacter sakazakii from powdered infant formula. Zhang. Inc. Powdered infant formula and fatal infecton with Enterobacter sakazakii. X. M.. Voedingscentrum. J. 77. H. Isolation and enumeration of Enterobacter sakazakii from dehydrated powdered infant formula. E. J. Food Drug and Administration (FDA). G. Guidelines on Safe preparation. Yu. 82.S. 76. 81. S.who.pdf. Weir. Food Prot. E. 2002. Food Prot.nl/nl/eten-gezondheid/zwangerschap-enkind/flesvoeding.

225 .

226 .

maar kunnen resulteren in vormen van neonatale meningitis (hersenvliesontsteking bij pasgeborenen). geïsoleerd uit vuil uit stofzuigerzakken. Het zijn Gram-negatieve. die te vroeg zijn geboren. waaronder huishoudens. Het blijkt dat Cronobacter spp. maar het merendeel van de infecties komt voor bij zuigelingen jonger dan één jaar. voorkomt in onder andere fabrieken waar graanproducten. 227 . lopen een verhoogd risico. Cronobacter heeft ziektes veroorzaakt in alle leeftijdsgroepen. houdt een groot gedeelte blijvende neurologische complicaties. aardappelzetmeel of melkpoeder verwerkt worden. De meeste Cronobacter stammen maken een kenmerkend geel pigment aan als ze op Trypton Soya Agar gekweekt worden. In het bijzonder baby’s jonger dan twee maanden. Cronobacter is uit verschillende voedingsmiddelen en omgevingen geïsoleerd.) en is onderdeel van de familie van de Enterobacteriaceae. In nationaal en internationaal verband wordt sinds enige jaren aandacht besteed aan de mogelijkheid dat de bacterie via poedervormige zuigelingenvoeding ziekte kan veroorzaken. verzameld in verschillende huishoudens. sepsis (bloedvergiftiging) en necrotiserende enterocolitis (ernstige ontsteking en afsterving van het maagdarmkanaal). en tot 1980 werd het genus aangegeven met de naam “geelgekleurde Enterobacter cloacae”. ondergewicht of een verzwakt afweersysteem hebben. Zuigelingenvoeding is echter het enige levensmiddel dat als mogelijke infectiebron geïdentificeerd is. In dit proefschrift worden in hoofdstuk 2 en 3 methoden beschreven om snel en nauwkeurig vast te stellen of er Cronobacter voorkomt in een product of in een omgeving (detectie) en of de gevonden bacteriesoort inderdaad een Cronobacter is (identificatie). Cronobacter infecties zijn zeldzaam. Eveneens werd Cronobacter spp. staafvormige bacteriën die bij specifieke groepen gevoelige personen ziekte kunnen veroorzaken. De ontwikkelde selectieve methoden zijn toegepast in diverse producten en omgevingen. Vóór 2007 werd Cronobacter aangeduid als Enterobacter sakazakii.Samenvatting Het genus Cronobacter bestaat uit 6 verschillende species (= soorten = spp. Van de gerapporteerde patiënten is 19% overleden tussen 1961 en 2008 en van degenen die een infectie overleven.

000 cellen per voeding ofwel > 105) nodig zouden zijn om ziekte of dood te veroorzaken.Hoofdstuk 4 en 5 van dit proefschrift geven meer inzicht in het voorkomen van Cronobacter spp. met name in de afvulruimten. in de zuigelingenvoeding aanwezig is. voorkomen in de poedervormige zuigelingenvoeding kunnen deze redelijk goed overleven (hoofdstuk 8) en eventueel verder uitgroeien nadat de poedervormige zuigelingenvoeding aangemaakt is met water. door toepassing van kwaliteitssystemen zoals Good Manufacturing Practices (GMP) en Good Hygiene Practices (GHP). Het is belangrijk te realiseren dat zuigelingenvoeding niet een steriel product is en dat onjuiste bereiding en bewaring van de voeding de kans op besmetting kunnen vergroten. De in dit proefschrift beschreven resultaten kunnen onder ander door zuigelingenvoeding fabrikanten worden geïmplementeerd in de bestaande Hazard Analysis Critical Control Point (HACCP) plan om zo de veiligheid beter te beheersen. In hoofdstuk 6 worden 228 . goed kan overleven in droge omgevingen. (meer dan 100. in zuigelingenvoeding terechtkomen. in vergelijking tot andere bacteriën uit de Enterobactericeae familie. Wanneer lage aantallen Cronobacter spp. eventuele besmettingsroutes en bieden tevens mogelijkheden voor maatregelen om Cronobacter spp. Ondanks alle voorzorgsmaatregelen. Aangezien Cronobacter spp. kan het toch voorkomen dat Cronobacter spp. De snelheid en mate van uitgroei is sterk afhankelijk van de temperatuur van de aangemaakte voeding. is het zaak te voorkómen dat de bacterie in zuigelingenvoeding aanwezig is. dankzij een goede beheersing van de productie en de hygiëne bij het vullen van de verpakkingen. waarna tijdens het verdere productie proces geen hittebehandeling meer plaatsvindt. De bacterie kan dan ook via herbesmetting vanuit productieruimten. in aangemaakte zuigelingenvoeding was tot voor kort niet veel bekend. Eerder onderzoek heeft aangetoond dat Cronobacter spp. Deze hoge aantallen zijn nooit in droge poedervormige zuigelingenvoeding aangetroffen. in de fabrieksomgeving aanwezig kan zijn. droge ingrediënten aan de zuigelingenvoeding. Besmetting kan echter ook optreden tijdens het bereiden van de zuigelingenvoeding zowel in het ziekenhuis als thuis. infecties te voorkomen. Een andere besmettingsroute is via het toevoegen van besmette. Resultaten uit een experimentele dierstudie doen suggereren dat hoge aantallen levende cellen van Cronobacter spp. Over de groei van Cronobacter spp..

229 . en door de poedervormige zuigelingenvoeding aan te maken met water dat een temperatuur van 25 ºC of minder heeft. kan worden voorkomen door de aangemaakte zuigelingenvoeding in kleine porties ter grootte van één voeding te bewaren in de koelkast. in aangemaakte zuigelingenvoeding te voorspellen gedurende het afkoelproces. in de zuigelingenvoeding onder verschillende aanmaak. De temperatuur van huishoudkoelkasten is echter niet altijd beneden de 7 °C. Voorspellingen laten zien dat de groei van Cronobacter spp. wat in Nederland de voorgeschreven maximum temperatuur is. Aangemaakte zuigelingenvoeding kan worden bewaard in de koelkast. in zuigelingenvoeding is onderzocht. waaronder de groeisnelheid van Cronobacter spp. Na het plaatsen in de koelkast duurt het geruime tijd voordat de koelkasttemperatuur bereikt is.en bewaaromstandigheden om zo de effecten van opslag en bereiding van zuigelingenvoeding door te rekenen. De in dit proefschrift beschreven resultaten zijn gebruikt in de expert meetings van de FAO (Food and Agricultural Organization of the United Nations) en de WHO (World Health Organization) waarbij het risico van Cronobacter spp. bijvoorbeeld in het kader van microbiologische bepalingen (Microbiologisch Risk Assessments) die ontwikkeld worden om de kans op ziekte na consumptie te kunnen schatten.voorspellende modellen gebruikt om de groeiparameters. hebben optimale groei tussen 37 en 39 ºC. te kunnen bepalen bij diverse temperaturen. in aangemaakte zuigelingenvoeding. Cronobacter spp. De ontwikkelde modellen kunnen worden toegepast voor het kwantificeren van de groei van Cronobacter spp. Een nog belangrijker probleem is echter dat de voeding niet onmiddellijk dezelfde temperatuur heeft als de lucht in de koelkast. De verkregen groeiparameters uit hoofdstuk 6 zijn in hoofdstuk 7 gebruikt om een model te ontwikkelen om de proliferatie van Cronobacter spp.

230 .

Leon thank you for the useful discussions. I would also like to thank all the other staff of the Nestlé Food Safety Microbiology group at the Nestlé Research Center in Lausanne. Mats Peters. the members of Scientific Advisory Board and co-authors contributed positively to my research project. I reached the last pages of this thesis. Wilma Hazeleger. Furthermore. Debbie. and Tvzetomira your work contributed to get more insight in the behaviour of Cronobacter spp. Rijkelt Beumer.. I have learned a lot from your practical experiences. Also. Han Joosten. I would like to thank my supervisors and co-supervisor for their input during the past years. Gerben. thank you for supervising my work. I would therefore like to take the opportunity to thank the various people who helped to finalize this dissertation. I supervised several BSc and MSc students. Ellen Brinkman and Roelina Dijk thank you for the valuable discussions. I would like to thank all my colleagues of the Laboratory of Food Microbiology. Cecil. in reconstituted infant formulae resulted in Chapter 6. especially from the infant formula manufacturing environment. Mick thank you for giving me the opportunity to conduct this research. Olivier Guilluame-Gentil. Leon and Marcel. Switzerland for the nice collaboration and sharing their experience on Cronobacter spp. Enne de Boer. Thank you all for your interest in conducting your thesis within my research project. for the nice conversations during the coffee and lunch breaks. reports and manuscripts. almost two years from the start you joined this project. your insights in the microbiological food safety contributed to this broad research approach. Over the past 231 . First and foremost. resulting in jointed publications. Dennis. Frans Rombouts. Dominique. Tjakko Abee. and Etienne your work on the growth of Cronobacter spp.Acknowledgements Finally. Although you were at a distance you kept on following the progress of my work. being my daily supervisor you always could improve my draft presentations. Martine. Ingrid. Together with Martine. your experience on predictive modeling came along at the right moment during my research.. Pieter Breeuwer. Marcel. Annette Heuvelink. Kishor and Samia part of your work is described in Chapter 7 and 8.

and Ingrid thank you for listening to my complaints every now and then.years many of you have left the department. we shared one office at RIKILT. Late aunt Rinia and Ine. Thank you for being one of my paranymphs. Now we will have more time to spend together. Annette and Esther thank you for reading through the various Chapters. Els. Maarten. The cover is beautiful. you were always there for me during the past years. Esther. Ida. Ingrid. for designing it and for helping out with the layout of this thesis. Dorothée. sharing the same office with you for some years was a very pleasant time. Sabita. Sarada and Ishaan. love and encouragement. Thank you. thank you for the practical assistance. thank you for your motivation to finish this thesis. Gerda van Laar. I enjoyed the time and the nice conversations we had during the Food Micro 2004 Congress in Slovenia. I would also like to thank you for arranging all administration activities in the project. 232 . I learned a lot from you. Fons thank you for your patience. now I can answer your question that I finally have finished my “study”. Thank you! Geeta thank you for your support and the discussions we had especially on the last part of this thesis. Dear family thank you for all your support. Wilma. Gerda van Donkersgoed. Gerda van Laar. I hope our friendship will last forever. and together with Aditya. especially with all the enrichment experiments. And last but not least.

R. M. M. C. M. Journal of Food Protection. Kandhai. 2:2. M. Booij.01. Van Schothorst.. H. Van Schothorst. H. M. C. M.2010.1016/j. J. F. 2006. 2009. H. Accepted for publication in Risk Analysis. G . Van Schothorst. Borjesson. and C. S. T. De Heer. Dekkers. Van der Fels-Klerx. in The Netherlands between 2001and 2005. M. A.List of Publications Peer-reviewed papers Kandhai. Brynestad. and H. W. NJAS . J. Breeuwer. Kandhai. 2008.M. H. doi:10. 2010. Van der Fels-Klerx. H. Accepted for publication in Journal of Food Protection Kandhai.foodcont. Reij. H. M. C. C. H. G. Food Control. Gorris and M. Booij. Development of an European system for identification of emerging mycotoxins in wheat based supply chains. C.. and C. Kandhai. H. Kandhai. doi:10. Reij. 72:2489-2498. M. M. R. Van der Fels-Klerx. W. 233 ..003. M. S. Booij. and M. J.007. 2009. 2010. World Mycotoxin Journal. Morrison. C. 13-22. M. J. W. 72:4. Dijk. L. 1:1.njas. Martins. and other bacteria in dry powdered infant formula at various temperatures. Van der Fels-Klerx.. M. World Mycotoxin Journal. Inactivation rates of Cronobacter spp. Applied and Environmental Microbiology.Wageningen Journal of Life Sciences. Reij.. J. G. Booij. Kandhai. J. applied to mycotoxins in wheat based supply chains. 2010. J. C. and C. Zwietering. M. Jeurissen. and L. M. Beumer. The effect of pre-culturing conditions on the lag phase and the growth rate of Enterobacter sakazakii in reconstituted infant formula. J. A study into the occurrence of Cronobacter spp. W. H. P. M. A Delphi based expert study to identify indicators for emerging Fusarium related mycotoxins in wheat based supply chains. R.1016/j. E. M. A conceptual model for identification of emerging risks. S. E. C. 2010. C. M. Expert study to select indicators for identification of re-emerging mycotoxins.02. L.. H. M.. Zwietering. Kandhai. Gorris. Growth of Cronobacter sakazakii under dynamic temperature conditions during bottle cooling.. Gorris.2010. and M. Zwietering. M. C. Reij. Heuvelink. 2721-2729. 119-127. Dreyer.

J. under dynamic temperature conditions occurring during cooling of reconstituted powdered infant formula. USA. H. V. AA3. C. R. 2009. D. Reij. and M. J. G. and M. Journal of Food Protection. p. Reij. Food Micro 2008. M. Aberdeen. Van Schothorst. 46. H. H. p.. C. G. M. Gorris. L. Juneja & D. Sonnard. Z4.. W. Breeuwer. M.). Journal of Food Protection. Salmonella serotypes in the broiler supply chain. The Lancet. Early Identification of Emerging Mycotoxins in Wheat. W..Guillaume-Gentil. 1-4 September 2008. p. Non peer-reviewed scientific publications Kandhai. Van Puyvelde. Evolving microbial food quality and safety. A new protocol for the detection of Enterobacter sakazakii applied to environmental samples. G. C. In: The 21st international ICFMH Symposium. Van der Fels-Klerx. C.. M. Gorris. M.K. M. O. Evolving microbial food quality and safety. P. 6th International Conference Predictive Modeling in Foods Meeting Abstracts and Information Booklet. Growth of Cronobacter spp. Kandhai. E. Scotland. and M. D. J. Van Asselt. K. Washington DC. H. Kandhai. 68: 64-69. M. 67: 1267-1270. and C. Guillaume-Gentil. Guillaume-Gentil. Food Micro 2008. C. Van Schothorst. Booij. O.. 234 . Kandhai. Kandhai. 363: 39-40. 364: 414-415. C. Van Schothorst. Scotland. In: V. O. 2008. M. L. 2005. 2004. Van der Fels-Klerx. Reij. C. Reij. M. and M. Biesta-Peters. Aberdeen. H. O. G.W. Schaffner (Eds. Gorris. Enterobacter sakazakii in factories and households. W. Marugg. Kandhai. Beumer. A simple and rapid cultural method for detection of Enterobacter sakazakii in environmental samples. 2008. Van der Voet. M. M. 2004. Zwietering. L. Joosten. Kandhai. Guillaume-Gentil. Occurrence of Enterobacter sakazakii in the food production environments and households. W. and M. 2004. The Lancet. R. and H. J. E. M. In: The 21st international ICFMH Symposium. M...

Report R2008. 235 . M. RIVM report 320111002/RIKILT report 2008. Enterobacter sakazakii: a risk for infants? Prospects for health. In: New tools for improving microbial food safety and quality. C. Wageningen.004/PRI report 186. Reij. and C. The Netherlands. In: Van der Fels-Klerx. The Netherlands. European network of information sources for an identification system for emerging mycotoxins in wheat based supply chains (MYCONET). Wageningen : Stichting EFFI . Kandhai M. 78. C. p. Kandhai M. H. M. Zwietering. Enterobacter sakazakii en zuigelingen voeding Symposium 'Voedselvergiftigingen en -infecties' : EFFI symposium . Booij.. C. C. J. Development of a model to assess the occurrence of mycotoxins in wheat.nl/bibliotheek/rapporten/320111002. Available at: http://www. J. p.wur. H. C. Gorris. W. p.. Van der Fels-Klerx. J. G.008.html Van der Fels-Klerx. C. Indicators for emerging mycotoxins. S.rikilt. Food Micro 2004 Congress Slovenia. Bos. J. 2005. M. C. Reij.Dekkers. Kandhai. 1–4. J. W. J. 2003. The effect of pre-culturing conditions on the lag phase of Enterobacter sakazakii. de Heer. F. Fourth Nizo Dairy conference. and P. Available at: http://www.).nl/NR/rdonlyres/BDEEDD31-F58C-47EB-A0AA23CB9956CE18/83031/R2008008. 335. and M. 2003. J. H. Zwietering. M. H. 2004.. 2003-05-15 / Stichting EFFI. Kandhai M. H.pdf. Van Schothorst. Booij. H. S.rivm. 2008.. M. Jeurissen. RIKILT – Institute of Food Safety. H.. 2008. C. H Booij (Eds. maize and nuts using a holistic approach-draft report. Kandhai. well-being and safety. and M. and M. Wageningen UR. L. M. M.

236 .

Currently. As part of her MSc study thesis research projects were carried out within the Food chemistry and Food Microbiology specialization. The Netherlands. on microbiological safety of various food supply chains. (Enterobacter sakazakii)” at the European chair in Food Safety Microbiology and the Laboratory of Food Microbiology of the Wageningen University.Curiculum Vitae Morea Chantal Kandhai was born on April 3 rd 1974 in Paramaribo. Wageningen. Suriname. 237 . During her internship at Numico Research.Institute of Food Safety. The results obtained during this research are described in this thesis. From 2007 till 2009 she worked as a researcher at RIKILT . she worked on the effect of antioxidant supplementation on the antioxidant status of healthy volunteers. In April 2001 she started with her PhD project entitled: “Detection. Chantal is working at the Board for the Authorisation of Plant Protection Products and Biocides (Ctgb). occurrence. She completed her secondary education (Lyceum II in Paramaribo) in 1993. as the first daughter of Johan Kit Rameshchandra Narain and late Jien-Tjauw Melie Narain-Tjon Sien Foek. growth and inactivation of Cronobacter spp. From 1994 until 2000 she studied Food Science at the Wageningen University (then the Agricultural University). In 2006 she worked part time as education coordinator at the Graduate school VLAG.

Slovenia (2004) NIZO Dairy Conference: Prospects for health. 2002. 2004. June 2004 General meetings Laboratory of Food Microbiology (2001. 2005) 238 . well-being and safety. Brussel. 2003. France (2003) NVvM symposium: Microbiologische voedselveiligheid: van gen tot wet (2003) Nationale Hygiënedag (2003) EFFI symposium: Microbiologisch onderzoek van Levensmiddelen (2003) EFFI symposium: Microbiële gevaren: voorkomen en prevention (2004) EUROFORUM: 5e Nationaal Congres Voedselveiligheid (2004) NVVM symposium: Microbiologisch onderzoek van levensmiddelen: eigen werk (2004) Food Micro: New tools for improving microbial food safety and quality. Switzerland (2001) Foreign students trip Slovakia. Quimper. Papendal (2005) EFFI symposium: Micro-organismen op weg (2006) General courses English for PhD students (2001) Endnote course (2001) Presenting in English (2003) PE&RC Time and project management (2004) Career perspectives (2004) OWU Gespreksvaardigheden (2005) OWU Basis cursus didactiek (2004/2005) OWU Afstudeervak organiseren en begeleiden (2003) Other activities Preparation research proposal (2001) Training at Nestlé Research Centre in Lausanne. Belgium (2003) EFFI symposium: Voedsel vergiftigingen en – infecties (2003) Fourth International Conference on Predictive Modeling in foods.Completed Training Activities Courses VLAG Management on food safety and microbiological risks (2002) VLAG Physiology of food associated micro-organisms (2004) VLAG Food Fermentation (2004) Basic and advanced statistic course (2003/2004) Bionumerics software course (2002) Meetings EFFI symposium: Overleving en na-besmetting (2001) NVVM symposium (2001) HACCP en de garantie van voedselveiligheid (2002) Safe: new emerging pathogens.

239 .

240 . Switzerland. Cover picture and design by Dorothée Becu. Lausanne. | Ponsen & Looijen.The research described in this thesis was financially supported by Nestlé Research Center.nu Printed by GVO printers & designers B. Ede. The Netherlands.V.becu. www.

Sign up to vote on this title
UsefulNot useful