Journal of Food Engineering 97 (2010) 497–503

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Modelling of Escherichia coli O157:H7 growth at various storage temperatures

on beef treated with electrolyzed oxidizing water
Tian Ding, S.M.E. Rahman, U. Purev, Deog-Hwan Oh *
Department of Food Science and Biotechnology and Institute of Bioscience and Biotechnology, Kangwon National University, Chuncheon, Gangwon 200-701, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: The influence of storage temperature (4, 10, 15, 20, 25, and 30 °C) on the growth of Escherichia coli
Received 3 August 2009 O157:H7 in beef untreated (control) and treated by acidic electrolyzed oxidizing water (AcEOW) or
Received in revised form 29 October 2009 slightly acidic electrolyzed oxidizing water (SAcEOW) was examined. A Baranyi model was employed
Accepted 14 November 2009
to describe growth parameters such as specific growth rate (SGR) and lag time (LT) as a function of stor-
Available online 18 November 2009
age temperature. SGR increased and LT declined with rising temperatures in all samples. There were no
significant differences between the SGR and LT values obtained from beef treated with AcEOW or SAc-
Keywords:
EOW. Secondary models were established for SGR and LT to evaluate the effects of storage temperature
Escherichia coli O157:H7
Baranyi model
on the growth kinetics of E. coli O157:H7 in treated and untreated beef. Mathematical evaluation was car-
Beef ried out to validate the performance of the developed models.
Acidic electrolyzed oxidizing water Ó 2009 Elsevier Ltd. All rights reserved.
Slightly acidic electrolyzed oxidizing water
Validation

1. Introduction oxidation reduction potential ORP (>1000 mV) is generally pro-

Escherichia coli O157:H7 was first identified in 1982 during an
investigation of two outbreaks of bloody diarrhoea in Oregon and
Michigan (Doyle et al., 2001). It is currently known to be one of
the most important food-borne pathogens with respect to public
health worldwide (Hodges and Kimball, 2005). In infected individ-
uals, E. coli O157:H7 can result in diarrhoea, haemorrhagic colitis,
haemolytic uremic syndrome, and thrombotic thrombocytopenic
purpura. CDC has reported that 1.03, 0.9, and 1.06 illnesses caused
by E. coli O157:H7 were found in 100,000 populations for 2003,
2004, and 2005, respectively (Anon, 2005).
Washing is the only way for consumers to reduce the popula-
tions of pathogens in their foods. Several sanitizers such as organic
acids (George et al., 1996; Pipek et al., 2006; Sommers et al., 2003),
ozone (Olmez and Akbas, 2009), trisodium phosphate solution
(Somers et al., 1994) and compounds of chlorine (Kim et al.,
2006; Weissinger et al., 2000) have been investigated for use in
controlling bacteria on fresh produce. Electrolyzed water is re-
garded as one of the most promising, with a high efficacy for the
inhibition of food-borne pathogens, but it was also less expensive
and safer than other traditional sanitizers. There are several differ-
ent types of electrolyzed water, including acidic electrolyzed oxi-
dizing water (AcEOW) and slightly acidic electrolyzed oxidizing
water (SAcEOW). AcEOW with low pH values (<2.7) and a high
* Corresponding author. Tel./fax: +82 33 250 6457.
E-mail address: deoghwa@kangwon.ac.kr (D.-H. Oh).
duced by the electrolysis of an aqueous sodium chloride solution
in an anode cell and has been estimated that has strong antimicro-
bial effects on many food-borne pathogens, including E. coli
O157:H7 (Kim et al., 2000; Ozer and Demirci, 2006). SAcEOW with
a pH of 5.0–6.5 and an ORP of 600 mV is generated by the electrol-
ysis of a dilute hydrochloric acid in a chamber without a mem-
brane (Cao et al., 2009). Yoshifumi (2003) demonstrated that the
effective form of chlorine compounds in SAcEOW is mainly the
hypochlorous acid (HOCl) and that it is nearly 80 times more effec-
tive as a sanitizer than an equivalent concentration of the hypo-
chlorite ion ClO for inactivating E. coli at a set contact time
(Anon, 1997; Cui et al., 2009). As SAcEOW is regarded as an advan-
tageous antimicrobial activity agent with a near neutral pH value
and low available chlorine (Izumi, 1999; Park et al., 2004), it has
been examined in several studies. Koide et al. (2009) reported that
slightly acid electrolyzed water (SLAEW: pH 6.1, 20 mg/L available
chlorine) reduced about by 1.5 log CFU/g for total aerobic bacteria
and 1.3 log CFU/g for moulds and yeasts, compared to fresh cut
cabbage before dipping, and Cao et al. (2009) investigated the effi-
ciency of slightly acidic electrolyzed water (SAEW) at different
temperatures (4, 20 and 45 °C) for inactivation of Salmonella ente-
ritidis in broth and on the surface of shell eggs, they found that the
antimicrobial effect of SAEW was higher than acidic electrolyzed
water (AEW). However, few literature has been applied the SAc-
EOW in muscle foods (e.g., meat, chicken, and beef).
Predictive microbiology aims to predict the microbial behaviour
in foods over time as a function of different influencing factors

0260-8774/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2009.11.007
498 T. Ding et al. / Journal of Food Engineering 97 (2010) 497–503
Nomenclature
Af accuracy factor
Bf bias factor
n the number of observations
Nt the microbial counts in units of log CFU/g at time (t)
N0 the logarithm of initial microbial counts (log CFU/g)
Nmax the logarithm of maximum microbial counts (log CFU/g)
obs the observed values
pred the predicted values
R2 correlation coefficient
R2Adj adjusted determination coefficient
t time (h)
T storage temperature (°C)
(Bang et al., 2008; McMeekin et al., 1997). Temperature is generally
regarded as the most important environmental factor that affects
bacterial growth in food. This factor constantly changes during
the processing, storage, and distribution of food products (Fujika-
wa et al., 2004). Predictive models are excellent tools for assessing
and controlling food safety, particularly when the models are able
to cope with dynamic conditions such as changing temperatures.
Many literatures have been published on development of models
on growth of microorganisms at different temperatures. Yaghlene
et al. (2009) reported a new predictive dynamic model describing
the effect of the ambient temperature on growth of E. coli, and Ding
et al. (2009) developed a model of E. coli O157:H7 growth in lettuce
treated with alkaline electrolyzed water at different storage tem-
peratures. In addition, many models have been developed relating
to AcEOW (Koseki and Isobe, 2001, 2005a,b), but still no predictive
models have been established regarding SAcEOW.
The objective of this study was to investigate and compare the
disinfection efficacy of SAcEOW and AcEOW for inactivating E. coli
O157:H7 in beef, and to develop predictive models for the growth
kinetics of E. coli O157:H7 in untreated and treated beef with AcE-
OW or SAcEOW. Furthermore, the SGRs and LTs obtained at various
storage temperatures were compared with those calculated using a
pathogen modelling programme (PMP).
2. Materials and methods
2.1. Bacterial strains
Two strain suspensions of E. coli O157:H7 used in this study
were B0259 and B0299 which obtained from the Department of
Food Science, University of Georgia (Griffin, GA, USA). All strains
were maintained at 70 °C in tryptic soy broth (TSB, Difco, Sparks,
MD, USA) with a 0.6% yeast extract (YE, Difco, USA) containing 20%
glycerol. An aliquot (10 lL) of the stock culture of each strain was
transferred into 10 mL TSBYE and then incubated at 35 °C for 24 h.
When the cultures reached the late stationary phase, the cells were
harvested by centrifugation for 5 min at 5000g and washed twice
in sterile 0.1% (w/v) buffered peptone water (Difco, USA). The cells
were subsequently adjusted to a final inoculum level of 9.0 log
CFU/mL.
2.2. Inoculation of beef
Beef samples from rib cut were purchased from a local super-
market in Chunchon, Korea. Fats and fascia were removed and then
stored in a refrigerator at 4 °C for at most 1 day. Beef samples were
cut into pieces of similar size (3  3 cm) using a sterile knife. Be-
fore inoculation, each sample weighed 10 ± 0.1 g.
T0 notional minimum temperature (°C) for growth
v the rate of increase of the limiting substrate (log CFU/h)
Greek symbols
l specific growth rate (log CFU/h)
k lag time (h)
Abbreviation
AcEOW acidic electrolyzed oxidizing water
ORP oxidation–reduction potential (mV)
SAcEOW slightly acidic electrolyzed oxidizing water
An inoculum solution was prepared by transferring 1 mL of the
mixed strain cocktails (9.0 log CFU/mL) of each bacterial species
into 2 L of sterile distilled water. The pieces of beef were then im-
mersed in the inoculum solution for 3 min and then dried com-
pletely on a clean bench with absorbent paper. Through this
process, the beef samples were inoculated with approximately
5.0 log CFU/g E. coli O157:H7. Inoculated samples without AcEOW
and SAcEOW treatments were used as positive controls.
2.3. Preparation of treatment solutions
AcEOW was obtained from a continuous EO generator (A2-
1000, Korean E & S Fist Inc., Seoul, Korea) at 220 V and 30 A. A con-
tinuous supply of softened tap water and 12% sodium chloride
solution at room temperature (23 ± 2 °C) was pumped into this
equipment. Generally, the EO generator electrolyzed at 13.5 V,
12 A and flow rate at around 1500 ml/min. When stable amperage
was reached after 15 min, AcEOW (available chlorine of 50 ppm,
pH value of 2.3–2.7 and oxidation–reduction potential of 1110–
1200 mV) was collected. SAcEOW, with a pH of 6.2, ORP of 500–
520 mV, and an available chlorine concentration of 5 ppm, was
produced through the electrolysis of a dilute NaCl solution in a
chamber without a membrane using an electrolysis device (model
D-7, sl no. 001171, Dolki Co. Ltd., Wonju, Korea) at a setting of 3 V.
A pH/ORP meter (YK-2001PH, Taiwan China) was used to deter-
mine the pH and ORP of each treatment solution. Available chlo-
rine concentrations were determined by a colorimetric method
using a digital chlorine test kit (RC-3F, Kasahara Chemical Instru-
ments Corp., Saitama, Japan). All measurements were carried out
immediately before each bactericidal experiment.
2.4. Experimental procedure
Preliminary experiments were carried out to determine the dip-
ping time (0.5, 1, 3, 5, and 10 min) of AcEOW and SAcEOW treat-
ments which suggested that 3 min dipping was the most
satisfied (data not shown). Inoculated beef samples (10 g each)
were immersed in 100 mL of AcEOW and SAcEOW for 3 min at
room temperature (23 ± 2 °C), and then transferred into a sterile
beaker filling with 100 mL neutralizing agent (0.5% sodium thiosul-
fate solution) in order to stop the bactericidal effects of AcEOW and
SAcEOW. After 3 min, all treated beef samples were transferred
into new sterile stomacher bags (Nasco Whirl-Pak, Janesville,
WI). Untreated inoculated beef was used as a control.
Each 10-g sample of beef without inoculation, untreated or
treated beef inoculated with E. coli O157:H7 was mixed with
90 mL of 0.1% sterile peptone water and then pummeled for
2 min in a stomacher (Lab-blender 400, Seward, London, UK). This
 T. Ding et al. / Journal of Food Engineering 97 (2010) 497–503
mixture was serially diluted and plated onto tryptic soy agar (TSA,
Difco Co., USA) to enumerate total bacterial count on beef sample
without inoculation and MacConkey sorbitol agar (Difco Co.,
USA) to enumerate E. coli O157:H7 on treated or untreated beef
samples, respectively. All plates were incubated at 35 °C for 24 h.
After incubation, colonies of naturally total bacteria and E. coli
O157:H7 were enumerated and expressed as log colony forming
units per gram (log CFU/g).
For shelf-life study, the untreated and treated beef samples
were air-packaged using stomacher bag and marked carefully be-
fore storage at different storage temperatures (4, 10, 15, 20, 25,
and 30 °C). During storage, sampling was carried out at 2-day
intervals for 4 °C, while only 3-h intervals for 30 °C. Generally, low-
er temperatures resulted in longer sampling intervals, while short-
er intervals were chosen for higher temperatures. At each sampling
interval, 90 mL of 0.1% sterile peptone water were poured into the
stomacher bag containing the sample. After 2 min of homogeniza-
tion (Lab-blender 400, Seward, London, UK), 0.1 mL aliquots of the
appropriate dilution were spread on the surface of duplicate plates
of MacConkey sorbitol agar (Difco Co., USA), and the plates were
then incubated at 35 °C for 24 h. The colonies were then counted
using the standard plate count (SPC) method. Each experiment
was replicated three times, and means of bacterial populations
(log10 CFU/g) from different storage temperatures were calculated.
2.5. Model development
Baranyi model (Eq. (1)) described by Baranyi and Roberts (1994)
was used to describe the bacterial growth curves at different stor-
age temperatures and growth parameters (SGR: log units/hour; LT:
hour) were estimated analysed by DMFit Excel Add-In software
(courtesy of J. Baranyi, Institute of Food Research, Norwich, UK)
 
elf ðtÞ  1
Nt ¼ N0 þ lf ðtÞ  ln 1 þ ðNmax N Þ
e 0
1   1.72 ± 0.09 log reductions were observed, respectively. Therefore,
f ðtÞ ¼ t þ ln ev t þ elk  eðv tlkÞ
v
where Nt is the microbial counts in units of log CFU/g at time, t(h),
N0 is the logarithm of initial microbial counts (log CFU/g), Nmax is
the logarithm of maximum microbial counts (log CFU/g), l is the
specific growth rate (log CFU/h), v is the rate of increase of the lim-
iting substrate, assumed to be equal to l and k is the lag time (h).
Pathogen modelling programme (PMP) is one of the most well-
known predictive software packages in the world, which can be
downloaded freely (Koseki and Isobe, 2005a). SGRs and LTs calcu-
lated from the growth models of E. coli O157:H7 of PMP were em-
ployed to compare with ones obtained from the developed models
in this study.
The obtained SGR values were analysed in SPSS v13.0 (Statisti-
cal Package for the Social Sciences, Chicago, IL, USA) to develop a
square root model as a secondary model using the equation ex-
pressed as follows (Ratkowsky et al., 1982, 1983):
pffiffiffiffi
l ¼ b  ðT  T 0 Þ
A natural logarithm model was better fitted for LT than the
square root model as a secondary model in SPSS. The equation is
described below:
Lnk ¼ b  ðT  T 0 Þ
where l is the growth rate of E. coli O157:H7, k is the lag time, b is
the regression constant, T is the temperature (°C), and T0 is a no-
tional minimum temperature for growth.
499
2.6. Validation
Validation is an important step for evaluating the ability of a
new model to interpolate, and it is a critical step in their develop-
ment (Bang et al., 2008). In this study, the goodness-of-fit of the
models was evaluated by the correlation coefficient (R2) and ad-
justed using the determination coefficient (R2Adj ), moreover, the bias
factor (Bf, Eq. (4)) and the accuracy factor (Af, Eq. (5)) were used to
validate the performances of the developed models. In all cases, the
Bf and Af values were close to unity, which indicated a good fit be-
tween the observations and the predictions. Dalgaard and Jorgen-
sen (1998) also reported that these two factors were regarded as
valuable tools for the evaluation of predictive models. Bf is an esti-
mate of the extent of under- or over-prediction by the model as it
gives the structural deviations (te-Giffel and Zwietering, 1999), but
it cannot indicate the average accuracy of the estimates. Therefore,
Af can be calculated, as it averages the distance between each point
and the line of equivalence as a measure of how close, on average,
the predictions are to the observations (Ross, 1996)
 
P
n
log ðobs=predÞ=n
Bf ¼ 10 i¼1 ð4Þ
 
Pn
j logðpred=obsÞj=n
Af ¼ 10 i¼1 ð5Þ
where n is the number of observations, obs is the observed value,
and pred is the predicted value.
3. Results and discussion
3.1. Primary modelling of E. coli O157:H7 growth
The initial population of E. coli O157:H7 on inoculated beef
ð1Þ samples in the experiment was approximately 5.0 log CFU/g. After
treatment with AcEOW or SAcEOW, about 1.64 ± 0.13 or
the population of E. coli O157:H7 on beef samples treated with
AcEOW or SAcEOW was approximately 3.36 and 3.28 log CFU/g.
Fig. 1 illustrates the growth curves of E. coli O157:H7 on un-
treated beef and beef treated with AcEOW or SAcEOW at different
storage temperatures (4, 10, 15, 20, 25, and 30 °C). Baranyi model
was fitted well with these growth curves that provided a good sta-
tistical fit (R2 > 0.97). The SGR and LT of E. coli O157:H7 in each
sample are shown in Table 1, together with those calculated using
PMP. As expected, higher incubation temperatures resulted in
higher SGRs (Allende et al., 2007), and at the same time, shorter
LTs were observed on control beef and treated beef samples. In
addition, the SGR values obtained from different storage tempera-
tures were significantly different, and there was also a significant
difference among the LT values, with the exceptions of 20 and
25 °C. PMP is the most well-known predictive software package
in the world (Koseki and Isobe, 2005a). Thus, the PMP growth mod-
el was chosen for a comparative growth study with the control
beef and treated beef samples. There was a great difference be-
ð2Þ tween the specific growth rate and lag time obtained from the beef
experiment and that predicted using PMP; the discrepancy in-
creased with the storage temperature.
Koseki and Isobe (2005b) reported that total bacteria counts
could also influence growth parameters during storage processing.
In this study, the total bacteria counts on untreated and treated
ð3Þ beef samples with AcEOW and SAcEOW were 5.21, 3.89 and
3.92 log CFU/g, respectively. The ratio of total bacteria counts to
inoculated E. coli O157:H7 was an important impact factor on the
growth of E. coli O157:H7. There were significant differences be-
tween the untreated and treated beef samples for SGR and LT;
500 T. Ding et al. / Journal of Food Engineering 97 (2010) 497–503

Fig. 1. Growth curves of E. coli O157:H7 in control, beef treated with SAcEOW and AcEOW at different storage temperatures (4, 10, 15, 20, 25, and 30 °C).

Table 1
Specific growth rate (SGR) and lag time (LT) of E. coli O157:H7 on beef with untreated and treated by acidic electrolyzed oxidizing water (AcEOW) or slightly acidic electrolyzed
oxidizing water (SAcEOW), and calculated using a pathogen modelling programme (PMP).

Temperature (°C) Specific growth rate (log CFU/h)a,b Lag time (h)a,b
c d e f
Control SAcEOW AcEOW PMP Control SAcEOW AcEOW PMP
4 A0.02a A0.025b A0.026b NPg E121.67a E120.73a E119.52a NP
10 B0.044a B0.053b B0.056b 0.057 D62.46b D56.34ab D55.64a 53.97
15 C0.078a C0.113b C0.121c 0.137 C20.23b C17.8ab C16.8a 16.9
20 D0.183a D0.225b D0.231c 0.282 B11.3b B9.73a B9.57a 6.58
25 E0.309a E0.356b E0.37b 0.498 B9.74b B8.53a AB8.47a 3.19
30 F0.377a F0.44b F0.445b 0.751 A5.5b A4.88a A4.85a 1.92
a
Within the same column, values not preceded by the same capital letter are significantly different (p < 0.05).
b
Within the same row, values not followed by the same letter are significantly different (p < 0.05).
c
Not treated with a sanitizer. The natural total bacteria counts were 5.21 log CFU/g.
d
Treated with 5 ppm slightly acidic electrolyzed oxidizing water (SAcEOW) for 3 min. The natural total bacteria counts were 3.92 log CFU/g.
e
Treated with 50 ppm acidic electrolyzed oxidizing water (AcEOW) for 3 min. The natural total bacteria counts were 3.89 log CFU/g.
f
Data from the USDA-PMP.
g
No prediction.

higher SGRs and lower LTs were observed in both groups of treated lower than control; also the time required to arrive at the final sta-
beef compared to control. However, it is observed from Fig. 1 that tionary phase was prolonged after treatments, in other words, the
the number of bacteria on treated samples was always significantly beef treated with AcEOW and SAcEOW can be stored for a longer
 T. Ding et al. / Journal of Food Engineering 97 (2010) 497–503 501

time compared with control. Similar conclusions have been pub-
lished in lettuce treated with alkaline electrolyzed water by Ding
et al. (2009).
3.2. Secondary modelling of E. coli O157:H7 growth
The SGR and LT values obtained from the Baranyi model were
used to develop the secondary models in order to describe the rela-
tionship between the growth parameters and the storage temper-
ature. The square root model and the natural logarithm model
were established for SGR and LT, respectively, to determine the ef-
fect of temperature on the growth kinetics of E. coli O157:H7 on
beef with untreated and treated by AcEOW or SAcEOW. The sec-
ondary models developed for SGR and LT are summarised in Ta-
ble 2. The correlation coefficient (R2) is often used as an overall
measure of the prediction, and it measures the fraction of the var-
iation about the mean that is explained by the model. A higher R2
value signifies a better prediction by the model (Grau and Vanderl-
inde, 1993; Duffy et al., 1994; Sutherland et al., 1994; te-Giffel and
Zwietering, 1999). As listed in Table 2, the R2 values of all of the
predictive models were greater than 0.94, indicating that the mod-
els could give good predictions. In addition, the values of the R2Adj
(R2Adj >0.93) indicated a high degree of correlation between the ob-
served and predicted values, which suggested that less than 7% of
the total variation could not be explained by the current models.
Figs. 2 and 3 show the scatter plot diagrams of the observed values
versus the predicted values calculated from the developed second-
ary models presented in Table 2. The data points lay near the line of
unity, and the correlation coefficients (R2) were greater than 0.98,
indicating that the models satisfactorily predicted growth under
the studied conditions.
3.3. Validation

Table 2
Table 3 shows the Bf and Af values of the statistical indices for
Secondary models of E. coli O157:H7 inoculated on untreated beef (control), beef
treated with acidic electrolyzed oxidizing water (AcEOW) or slightly acidic electro- the developed secondary models of SGR and LT. All Bf values ob-
lyzed oxidizing water (SAcEOW), and calculated using a pathogen modelling tained from the square root models and the natural logarithm
programme (PMP). models were in the range of 0.99–1.01, which were close to unity,
Equations a 2
R b 2 indicating a perfect concordance. For models describing pathogen
RAdj
pffiffiffiffiffiffiffiffiffi
growth kinetics, Ross (1996) published that Bf in the range of
Control SGR = 0.020T + 0.031c 0.975 0.968 0.9–1.05 could be considered good for models describing a patho-
Ln (LT) = 0.121T + 5.147d 0.955 0.944
pffiffiffiffiffiffiffiffiffi gen growth rate, 0.7–0.9 or 1.06–1.15 was considered acceptable,
SAcEOW SGR = 0.021T + 0.048 0.986 0.982
and <0.7 or >1.15 was considered unacceptable. Therefore, the Bf
Ln (LT) = 0.124T + 5.107 0.950 0.938
pffiffiffiffiffiffiffiffiffi values obtained in our study were acceptable and indicated that
AcEOW SGR = 0.021T + 0.054 0.986 0.982
Ln (LT) = 0.124T + 5.081 0.944 0.930 there was only a minimal difference between the predicted and ob-
pffiffiffiffiffiffiffiffiffi
PMP SGR = 0.032T  0.094 0.998 0.997 served data. Compared with other studies published concerning
Ln (LT) = 0.167T + 5.438 0.977 0.969 the bias factor, the results obtained in our study are generally bet-
a
Correlation coefficient.
ter (Lebert et al., 2000; Neumeyer et al., 1997; Garcia-Gimeno
b
Adjusted determination coefficient. et al., 2003; Grau and Vanderlinde, 1993; Patterson et al., 1993).
c
Square root model equation for specific growth rate. For the accuracy factor, Ross et al. (2000) reported that predictive
d
Natural logarithm model for lag time. models should ideally have an Af = 1.00, indicating a perfect model

A 0.4 B 0.5
R2 = 0.9801 0.4 R2 = 0.9876
Predicted value

0.3
Predicted value

0.3
0.2
0.2
0.1
0.1

0 0
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4 0.5
Observed value Observed value

C 0.5 D 0.8

0.4 R2 = 0.9848 R2 = 0.9989

Predicted value

0.6
Predicted value

0.3
0.4
0.2

0.1 0.2

0 0
0 0.1 0.2 0.3 0.4 0.5 0 0.2 0.4 0.6 0.8
Observed value Observed value
Fig. 2. Observed value versus predicted specific growth rate (SGR) obtained from developed secondary models for the growth of E. coli O157:H7 in control beef (A), beef
treated with SAcEOW (B), beef treated with AcEOW (C), and calculated from PMP (D).
502 T. Ding et al. / Journal of Food Engineering 97 (2010) 497–503

A 140 B 140
R2 = 0.9856 R2 = 0.9868
105

Predicted value
105

Predicted value 70 70

35 35

0 0
0 35 70 105 140 0 35 70 105 140
Observed value Observed value

C 140 D 60

R2 = 0.9853 R2 = 0.9836
105

Predicted value
Predicted value

40

70
20
35

0 0
0 35 70 105 140 0 20 40 60
Observed value Observed value

Fig. 3. Observed value versus predicted lag time (LT) obtained from developed secondary models for the growth of E. coli O157:H7 in control beef (A), beef treated with
SAcEOW (B), beef treated with AcEOW (C), and calculated from PMP (D).

Table 3 4. Conclusion
Bf and Af values of the mathematical and statistical indices for the secondary models
in this study.
The bactericidal effects of AcEOW and SAcEOW against E. coli
SGR LT O157:H7 on beef were determined and approximately 1.64 ± 0.13
Bfa Afb Bf Af or 1.72 ± 0.09 log reductions were observed, respectively. The re-
sults showed that both had strong antimicrobial effects, and there
Control 0.996 1.217 1.009 1.243
SAcEOW 1.007 1.143 0.994 1.263 were no significant differences between 50 ppm AcEOW and
AcEOW 1.009 1.147 1.001 1.282 5 ppm SAcEOW. To investigate the influence of storage temperature
PMP 0.994 1.052 1.004 1.189 (4, 10, 15, 20, 25, and 30 °C) on the growth of E. coli O157:H7 in un-
a
Bias factor. treated beef and beef treated with AcEOW or SAcEOW, a Baranyi
b
Accuracy factor. model was employed to generate the SGR and LT during a shelf-life
test. A square root model and a natural logarithm model were
established to describe the relation between temperature and the
fit, where the predicted and actual response values are equal. How- growth kinetics of E. coli O157:H7. The bias and accuracy factors
ever, Ross et al. (2000) and Carrasco et al. (2006) proposed that an of the mathematical and statistical indices were used to validate
acceptable Af could be determined by the effect of the number of the performance of the current models. The results demonstrated
environmental parameters in a kinetic model with a increase of that the square root models developed for SGR could make a good
0.10–0.15 units for each predictive variable. In this study, the best prediction, while the predictions of the natural logarithm models
performance that might be expected from the kinetic model were not so accurate. In the future, a new model should be estab-
encompassing the effects of temperature on SGR and LT is approx- lished for lag time to provide reliable predictions.
imately 10–15%, or an accuracy factor of 1.10–1.15. As shown in
Table 3, only three Af values obtained from the square root models
References
for SGR were within the acceptable range, while the others were all
greater than 1.15. For the predictive models of LT, the results indi- Allende, A., Martinez, B., Selma, V., Gil, M.I., Suarez, J.E., Rodriguez, A., 2007. Growth
cated that the predictions differed from the observations by an and bacteriocin production by lactic acid bacteria in vegetable broth and their
average of 24.3%, 26.3%, 28.2%, and 18.9%, respectively. Schaffner effectiveness at reducing Listeria monocytogenes in vitro and in fresh-cut lettuce.
Food Microbiology 24, 759–766.
(1995) summarised that most of the models were not good at pre- Anon, 1997. Principle of Formation of Electrolytic Water. Hoshizaki Electric Co. Ltd.,
dicting lag times. Even previous results showed that square root Sakae, Toyoake, Aichi, Japan.
models predicted lag times with a very high average error (36– Anon, 2005. Preliminary FoodNet Data on the Incidence of Infection with Pathogens
Transmitted Commonly Through Food-10 Sites, vol. 54. United States, 2004.
40%). Moreover, Robinson et al. (1998) and Ross et al. (2000) have
Morbidity and Mortality Weekly Report, pp. 352–356.
reported that it was much more difficult to create a model to pre- Bang, W.S., Chung, H.J., Jin, S.S., Ding, T., Hwang, I.G., Woo, G.J., Ha, S.D., Bahk, G.J.,
dict lag time than other growth parameters, because it depended Oh, D.H., 2008. Prediction of Listeria monocytogenes growth kinetics in sausages
on many factors, for instance, the physiological state of the bacte- formulated with antimicrobials as a function of temperature and
concentrations. Food Science and Biotechnology 17, 1316–1321.
rial cells, the size of the inoculum, the previous growth conditions, Baranyi, J., Roberts, T.A., 1994. A dynamic approach to predicting bacterial growth in
and so on. food. International Journal of Food Microbiology 23, 277–294.
 T. Ding et al. / Journal of Food Engineering 97 (2010) 497–503
Cao, W., Zhu, Z.W., Shi, Z.X., Wang, C.Y., Li, B.M., 2009. Efficiency of slightly acidic
electrolyzed water for inactivation of Salmonella enteritidis and its contaminated
shell eggs. International Journal of Food Microbiology 130, 88–93.
Carrasco, E., Garcia-Gimeno, R., Seselovsky, R., Valero, A., Perez, F., Zurera, G., 2006.
Predictive model of Listeria monocytogenes’ growth rate under different
temperatures and acids. Food Science and Technology International 12 (1),
47–56.
Cui, X.D., Shang, Y.C., Shi, Z.X., Xin, H.W., Cao, W., 2009. Physicochemical properties
and bactericidal efficiency of neutral and acidic electrolyzed water under
different storage conditions. Journal of Food Engineering 91, 126–132.
Dalgaard, P., Jorgensen, L.V., 1998. Predicted and observed growth of Listeria
monocytogenes in seafood challenge tests and in naturally contaminated cold-
smoked salmon. International Journal of Food Microbiology 40, 105–115.
Ding, T., Jin, Y.G., Rahman, S.M.E., Kim, J.M., Choi, K.H., Choi, G.S., Oh, D.H., 2009.
Prediction of growth of Escherichia coli O157:H7 in lettuce treated with alkaline
electrolyzed water at different temperatures. Journal of Food Hygiene and
Safety 24, 232–237.
Doyle, M., Zhao, T., Meng, J., Zhao, S., 2001. E. coli O157:H7 in food. In: Doyle, M.P.,
Beuchat, L.R., Montville, T.J. (Eds.), Microbiology: Fundamentals and Frontiers.
ASM Press, Washington, DC, pp. 171–192.
Duffy, L.L., Vanderlinde, P.B., Grau, F.H., 1994. Growth of Listeria monocytogenes on
vacuum-packed cooked meats: effects of pH, Aw, nitrite and ascorbate.
International Journal of Food Microbiology 23, 377–390.
Fujikawa, H., Kai, A., Morozumi, S., 2004. A new logistic model for Escherichia coli
growth at constant and dynamic temperatures. Food Microbiology 21,
501–509.
Garcia-Gimeno, R.M., Hervas-Martinez, C., Barco-Alcala, E., Zurera-Cosano, G., Sanz-
Tapla, E., 2003. An artificial neural network approach to Escherichia coli
O157:H7 growth estimation. Journal of Food Science 68, 639–645.
George, S.M., Richardson, L.C.C., Michael, W., 1996. Peck predictive models
of the effect of temperature, pH and acetic and lactic acids on the growth of
Listeria monocytogenes. International Journal of Food Microbiology 32,
73–90.
Grau, F.H., Vanderlinde, P.B., 1993. Aerobic growth of Listeria monocytogenes on beef
lean and fatty tissue: equations describing the effects of temperature and pH.
Journal of Food Protection 56, 96–101.
Hodges, J.R., Kimball, A.M., 2005. The global diet trade and novel infections.
Globalization and Health 1, 1–7.
Izumi, H., 1999. Electrolyzed water as a disinfectant for fresh-cut vegetables.
Journal of Food Science 64, 536–539.
Kim, C., Hung, Y.C., Brackett, R.E., 2000. Efficacy of electrolyzed oxidizing (EO) and
chemically modified water on different types of foodborne pathogens.
International Journal of Food Microbiology 61, 199–207.
Kim, H., Ryu, J.H., Beuchat, L.R., 2006. Survival of Enterobacter sakazakii on
fresh produce as affected by temperature, and effectiveness of sanitizers
for its elimination. International Journal of Food Microbiology 111, 134–143.
Koide, S., Takeda, J., Shi, J., Shono, H., Atungulu, G.G., 2009. Disinfection efficacy of
slightly acidic electrolyzed water on fresh cut cabbage. Food Control 20 (3),
294–297.
Koseki, S., Isobe, S., 2001. Prediction of microbial growth in fresh-cut vegetables
treated with acidic electrolyzed water during storage under various
temperature conditions. Journal of Food Protection 64, 1935–1942.
Koseki, S., Isobe, S., 2005a. Growth of Listeria monocytogenes on iceberg lettuce and
solid media. International Journal of Food Microbiology 101, 217–225.
Koseki, S., Isobe, S., 2005b. Prediction of pathogen growth on iceberg lettuce under
real temperature history during distribution from farm to table. International
Journal of Food Microbiology 104, 239–248.
Lebert, I., Robles-Olvera, V., Lebert, A., 2000. Application of polynominal models to
predict growth of mixed cultures of Pseudomonas spp. and Listeria in meat.
International Journal of Food Microbiology 61, 27–39.
McMeekin, T.A., Brown, J., Krist, K., Miles, D., Neumeyer, K., Nichols, D.S., Olley, J.,
Presser, K., Ratkowsky, D.A., Ross, T., Salter, M., Soontranon, S., 1997.
503
Quantitative microbiology: a basis for food safety. Emerging Infectious
Diseases 3, 541–550.
Neumeyer, K., Ross, T., Thomson, G., McMeekin, T.A., 1997. Validation of a model
describing the effects of temperature and water activity on the growth of
psychrotrophic pseudomonads. International Journal of Food Microbiology 38,
55–63.
Olmez, H., Akbas, M.Y., 2009. Optimization of ozone treatment of fresh-cut green
leaf lettuce. Journal of Food Engineering 90, 487–494.
Ozer, N.P., Demirci, A., 2006. Electrolyzed oxidizing water treatment for
decontamination of raw salmon inoculated with Escherichia coli O157:H7 and
Listeria monocytogenes Scott A and response surface modeling. Journal of Food
Engineering 72, 234–241.
Park, H., Hung, Y.C., Chung, D., 2004. Effects of chlorine and pH on efficacy of
electrolyzed water for inactivating Escherichia coli O157:H7 and Listeria
monocytogenes. International Journal of Food Microbiology 91, 13–18.
Patterson, M.F., Damoglou, A.P., Buick, R.K., 1993. Effects of irradiation dose and
storage temperature on the growth of Listeria monocytogenes on poultry meat.
Food Microbiology 10, 197–203.
Pipek, P., Houska, M., Hoke, K., Jeleníkova, J., Kyhos, K., Sikulova, M., 2006.
Decontamination of pork carcasses by steam and lactic acid. Journal of Food
Engineering 74, 224–231.
Ratkowsky, D.A., Olley, J., McMeekin, T.A., Ball, A., 1982. Relationship between
temperature and growth rate of baterial cultures. Journal of Bacteriology 149,
1–5.
Ratkowsky, D.A., Lowry, R.K., McMeekin, T.A., Stokes, A.N., Chandler, R.E., 1983.
Model for bacterial culture growth rate throughout the entire biokinetic
temperature range. Journal of Bacteriology 154, 1222–1226.
Robinson, T.P., Ocio, M.J., Kaloti, A., Mackey, B.M., 1998. The effect of the growth
environment on the lag phase of Listeria monocytogenes. International Journal of
Food Microbiology 44, 83–92.
Ross, T., 1996. Indices for performance evaluation of predictive models in food
microbiology. Journal of Applied Bacteriology 81, 501–508.
Ross, T., Dalgaard, P., Tienungoon, S., 2000. Predictive modeling of the growth and
survival of Listeria in fishery products. International Journal of Food
Microbiology 62, 231–245.
Schaffner, D.W., 1995. The application of the Wlf equation to predict lag time as a
function of temperature for three psychrotrophic bacteria. International Journal
of Food Microbiology 27, 107–115.
Somers, E.B., Schoeni, J.L., Amy, C.L., 1994. Wong effect of trisodium phosphate on
biofilm and planktonic cells of Campylobacter jejuni, Escherichia coli O157: H7,
Listeria monocytogenes and Salmonella typhimurium. International Journal of
Food Microbiology 22, 269–276.
Sommers, C.H., Fan, X., Handel, A.P., Sokorai, K.B., 2003. Effect of citric acid on the
radiation resistance of Listeria monocytogenes and frankfurter quality factors.
Meat Science 63, 407–415.
Sutherland, J.P., Bayliss, A.J., Roberts, T.A., 1994. Predictive modeling of growth of
Staphylococcus aureus: the effects of temperature, pH and sodium chloride.
International Journal of Food Microbiology 21, 217–236.
te-Giffel, M.C., Zwietering, M.H., 1999. Validation of predictive models describing
the growth of Listeria monocytogenes. International Journal of Food
Microbiology 46, 135–149.
Weissinger, W.R., Chantarapanont, W., Beuchat, L.R., 2000. Survival and growth of
Salmonella baildon in shredded lettuce and diced tomatoes, and effectiveness of
chlorinated water as a sanitizer. International Journal of Food Microbiology 62,
123–131.
Yaghlene, H.B., Leguerinel, I., Hamdi, M., Mafart, P., 2009. A new predictive dynamic
model describing the effect of the ambient temperature and the convective heat
transfer coefficient on bacterial growth. International Journal of Food
Microbiology 133, 48–61.
Yoshifumi, H., 2003. Improvement of the electrolysis equipment and application of
slightly acidic electrolyzed water for dairy farming. Journal of the Japanese
Society of Agricultural Machinery 65, 27–29.