PEER-REVIEWED ARTICLE

Evann L. Dufort,1 Mark R. Etzel1 and

Food Protection Trends, Vol 37, No. 6, p. 409–418
Barbara H. Ingham1* Copyright© 2017, International Association for Food Protection
6200 Aurora Ave., Suite 200W, Des Moines, IA 50322-2864
1
Dept. of Food Science, University of Wisconsin-Madison,
1605 Linden Drive, Madison, WI 53706, USA

Thermal Processing Parameters to Ensure a 5-log

Reduction of Escherichia coli O157:H7, Salmonella
enterica, and Listeria monocytogenes in Acidified
Tomato-based Foods

ABSTRACT ference in the rate of pathogen inactivation. Extrapolating

Under 21 Code of Federal Regulations Part 114, man- 5-log pathogen reduction times to relevant processing tem-
ufacturers of acidified canned foods must apply a process peratures, E. coli O157:H7 was the most heat resistant
that ensures destruction of pertinent vegetative bacterial at < 65°C (149°F), while L. monocytogenes was most heat
pathogens and spoilage organisms. We used nonlinear resistant at temperatures above 65°C. Using a calculated
(Weibull) modeling to calculate thermal processing param- z-value of 13.3°F (7.4°C) and F-value of 0.51 min at 160°F
eters sufficient to inactivate Escherichia coli O157:H7, (71.1°C), time/temperature combinations to achieve a
Listeria monocytogenes, and Salmonella enterica in tomato minimum 5-log pathogen reduction in tomato purée, pH ≤
purée at 54°C. Inoculum (1 ml of a 5-strain single patho- 4.5, ranged from 13.83 min at 141°F (60.6°C) to 0.02
gen cocktail) was heated in 99 ml tomato purée at pH min at 180°F (82.2°C). Results can be used to inform
4.50 (unacidified) or in purée acidified to pH 3.80 or 4.20 development of scheduled processes and to support FDA
by use of acetic or citric acid. D54°C ranged from 20.85 min process filings for tomato-based acidified foods.
for E. coli O157:H7 in purée at pH 4.5; as well as 0.63
min for S. enterica in purée acidified to pH 3.8 by addition INTRODUCTION
of acetic acid. Acetic acid was significantly more effective From January 2008 to October 2012, 1,693 new acidified
than citric acid in ensuring pathogen inactivation (P < tomato-based products were introduced into the United
0.05). E. coli O157:H7 was significantly more heat- and States market, including salsas, pasta sauces, and other table
acid-resistant than S. enterica and L. monocytogenes in sauces. Consumer demand of these products is projected to
tomato purée at pH 4.5, as well as in purée acidified increase and to prompt continued growth of this category
to pH 4.2 by addition of citric or acetic acid (P < 0.05). (13). Regulations require that acidified foods be thermally
L. monocytogenes was the most heat- and acid-resistant in processed to an extent sufficient to destroy the vegetative
purée acidified to pH 3.8 by addition of acetic acid, but in cells of microorganisms of public health significance and
purée acidified to pH 3.8 with citric acid, there was no dif- those of non-health significance capable of reproducing in

*Author for correspondence: Telephone: +1 608.263.7383; Fax: +1 608.262.6872; E-mail: bhingham@[Link]

November/December Food Protection Trends 409

the food under the conditions in which the food is stored, tion with citric and acetic acids on thermal inactivation of
distributed, retailed and held by the user (22). For acidified these vegetative pathogens in tomato purée at 54°C. Time/
foods, thermal processing is normally a mild heat treatment, temperature processing conditions were estimated that would
such as steam pasteurization or an inverted hot-fill hold that, ensure a minimum 5-log pathogen reduction of vegetative
for low pH products, is generally sufficient to guarantee pathogens and could be used to support safe manufacture of
safety (2). tomato-based acidified canned foods.
In the manufacture of acidified canned foods, acidification
is most often accomplished by addition of organic acids such MATERIALS AND METHODS
as acetic, lactic or citric acid, singly or in combination, and Strain selection and maintenance
employed to control microbial growth, improve sensory Strains of E. coli O157:H7, S. enterica, and L. monocytogenes
attributes, and reduce microbial spoilage of foods (1, 9, 15). were used in this study (Table 1). With one exception, these
In a review of 179 tomato-based acidified food products strains were the ones used in work by Breidt et al. (5–7);
found in local grocery stores, we noted that citric and acetic E. coli O157:H7 strain ATCC 43895 (American Type
acids were the most commonly used acidulants, usually in the Culture Collection, Manassas, VA) replaced strain ATCC
form of lime juice or distilled vinegar. Citric acid has been 43888 used in the prior work. Stock cultures of each strain
shown to inactivate Salmonella Typhimurium in tahini (1), were maintained in Tryptic Soy Broth at pH 7.1 (TSB;
while acetic acid has been shown to inactivate Escherichia coli Difco, Becton, Dickinson and Company, Sparks, MD)
O157:H7 in simulated pickle products (4) and apple-carrot containing 10% (v/v) glycerol (Fisher Scientific, Itasca,
juice blends (23). Acetic acid was also effective at inactivating IL) and stored frozen at -20°C. Bacterial strain identity
Salmonella, E. coli O157:H7, and L. monocytogenes to various and purity were assessed and confirmed by Gram reaction,
degrees in asparagus purée (20). Acetic and citric acids have cell and colony morphology, and responses to biochemical
been shown to inactivate E. coli O157:H7, Salmonella, and profiling kits (bioMerieux, Durham, NC) for E. coli
L. monocytogenes on fresh apples and lettuce (18) and in O157:H7 and S. enterica (API 20E), and L. monocytogenes
cucumber purée (3). (API Listeria). Working cultures were prepared monthly
Processors of acidified canned foods are required to regis- by streaking for isolation from partially thawed stock
ter their manufacturing facility and file scheduled processes cultures onto appropriate selective plating medium: L.
for their products with the Food and Drug Administration monocytogenes on Listeria Selective agar (LSA; Oxoid LTD,
(FDA) (22). FDA filings for acidified canned foods must be Basingstoke, Hampshire, England) with added Listeria
supported by research that validates destruction of import- Selective Supplement (Oxoid), and S. enterica and E. coli
ant foodborne pathogens, and E. coli O157:H7, S. enterica, O157:H7 on modified Levine’s Eosin Methylene Blue agar
and L. monocytogenes are considered targets for thermal pro- (mLEMB; Difco). mLEMB was prepared from Levine’s
cessing and acidification (5–7). The work of Breidt and col- Eosin Methylene Blue agar, with addition of D-sorbitol
leagues used a re-parameterized Weibull model to estimate (10 g/liter; Fisher) and NaCl (5 g/liter, Fisher). Working
thermal processing conditions necessary to achieve a 5-log culture plates were incubated at 35°C for 24 h for S. enterica
reduction of target pathogens in an acidified cucumber and E. coli O157:H7, or 48 h for L. monocytogenes, followed
juice medium (5–7). Recent studies in our laboratory have by storage at 4°C for ≤ 40 days. E. coli O157:H7 colonies
confirmed the appropriateness of the non-linear Weibull appear colorless to pale pink on mLEMB agar, S. enterica
approach for calculating accurate thermal processing D- and colonies appear dark red to iridescent green on mLEMB agar,
z-values (10). and L. monocytogenes colonies appear pale yellow or grey
The process authority, in using laboratory data to establish surrounded by a black halo on LSA.
a scheduled process for an acidified food, must rely on an
understanding of heat penetration in a particular food matrix, Inoculum preparation
the impact of formulation on microbial inactivation, and For each trial, a 5-strain single-pathogen cocktail was
heat tolerance of spoilage organisms and target pathogens. prepared. Inocula were prepared by the selection, for each
It is in the best interest of manufacturers to have appropri- strain, of a single colony from a working culture plate, the
ate evidence to validate thermal processing conditions that cells of which were suspended into 9 ml of TSB +1% added
produce a safe but not over-processed product, leading to a glucose (Fisher) and statically incubated at 35°C for 24 ± 2
balance between food safety and food quality. In the ab- h to obtain stationary-phase cells (108–109 CFU/ml) and
sence of current recommendations developed with use of a expose growing cells to acid, as previously described (5, 8).
tomato-based medium, this study was conducted to calculate Contents of all five tubes for a species were combined in one
thermal processing parameters that would inactivate vegeta- conical 50 ml centrifuge tube and harvested by centrifugation
tive cells of the target pathogens E. coli O157:H7, S. enterica (4,500 g, 7 min, 21°C, Marathon 21K, Fisher). The super-
and L. monocytogenes in tomato purée. This study expanded natant was discarded and the pellet was re-suspended in 4.5
on previous work (10) and examined the effect of acidifica- ml Butterfield’s Phosphate Diluent (BPD; Nelson Jameson,

410 Food Protection Trends November/December

 TABLE 1. Bacterial strains used in this study

Strain name IDa Origin

Escherichia coli O157:H7 SRCC 1675 Apple cider-linked outbreak
Escherichia coli O157:H7 SRCC 1486 Salami-linked outbreak
Escherichia coli O157:H7 SRCC 2061 Ground beef
Escherichia coli O157:H7 SRCC 1941 Pork
Escherichia coli O157:H7 b
ATCC 43895 Ground beef
Salmonella Braenderup c
SRCC 1093 Eggs
Salmonella Cerro SRCC 400 Cheese powder
Salmonella Enteritidis SRCC 1434 Ice cream
Salmonella Newport SRCC 551 Broccoli with cheese
Salmonella Typhimurium SRCC 1846 Liquid egg
Listeria monocytogenes SRCC 529 Pepperoni
Listeria monocytogenes SRCC 1791 Yogurt
Listeria monocytogenes SRCC 1506 Ice Cream
Listeria monocytogenes SRCC 1838 Cabbage
Listeria monocytogenes SRCC 2075 Diced Coleslaw

a
ID = identification; SRCC strains obtained from Silliker, Inc., Chicago, IL.
b
Previous publications included ATCC 43888, not 43895, as part of an E. coli O157:H7 cocktail in determination of 5-log
pathogen reduction times for heat processed, acidified vegetable brine (5).
c
Salmonella enterica serotype.

Marshfield, WI) at 21°C and vortexed to obtain an inoculum tomato-based acidified foods that had been developed by the
cocktail (~1010 CFU/ml) (n = 116). corresponding author over a 1-year period.
Tomato samples at the target pH were aseptically trans-
Preparation of tomato purée ferred (99 ml) into 710 ml Whirl-Pak filter bags (Nasco, Fort
Tomato purée was prepared from a single lot of locally Atkinson, WI), air was expelled, and each bag was preheated
grown late-harvest Roma tomatoes that had been blanched to the target temperature prior to inoculation. After thermal
(98°C for 3 min), peeled, and placed into 3.8-liter freezer processing, each purée sample was cooled on ice to 21°C and
bags (Ziploc, S.C. Johnson, Racine, WI) in 453 g portions. the final pH was measured.
Samples were held frozen at -20°C and used within 12
months. Prior to each experiment, tomatoes were thawed Thermal processing
overnight (18–24 h) at 13°C, then brought to 21°C and For each trial, preheated tomato purée (99 ml) was
homogenized for 5–7 s (Cuisinart Smart Stick Hand Blender, inoculated with 1 ml of a 5-strain pathogen cocktail (~1010
Stamford, CT). Tomato purée was adjusted to a target pH CFU/ml) to produce a starting inoculum concentration
value prior to inoculation and heating. Unacidified tomato of 108–109 CFU/ml. Inoculation and sampling occurred
samples were adjusted to pH 4.50 ± 0.05 by addition of through a narrow (2.5–4 cm) access point in the top of
granular NaOH (Fisher) (0.15–0.35 g/99 ml tomato purée). the bag to restrict movement of air into the bag. Upon
Acidified samples were adjusted to the target pH (4.20 ± inoculation, bag contents were briefly stirred, using a pipette
0.05 or 3.80 ± 0.05) with addition of citric acid monohydrate tip, and samples of inoculated purée (1 ml) were taken
(Fisher) (0.1–0.2 g, 0.4–0.5 g) or glacial acetic acid (17.5 for enumeration at pre-determined intervals, depending
N, Fisher) (0.3 ml, 2.7 ml). Target pH values of 4.5, 4.2, and on the pathogen, temperature, and pH/acidulant. Total
3.8 were based on a review of 117 scheduled processes for heating times in unacidified tomato purée (pH 4.5) ranged

November/December Food Protection Trends 411

from 150 min for E. coli O157:H7 heated at 52°C, with determined using the re-parameterized non-linear Weibull
30 min sampling intervals, to 6 min for S. enterica heated model of Breidt et al. (5) according to Eq. (1):
at 58°C, with 1 min sampling intervals. Heating times in
acidified tomato purée at 54°C ranged from 60 min for E. log(N) = log(N0) – 5(t/t*)β, (1)
coli O157:H7 at pH 4.2 (citric acid), with 10 min sampling
intervals, to 4 min for S. enterica at pH 3.8 (acetic acid), where N is the cell count (CFU/ml) at time t, N0 is the
with 30 s sampling intervals. At least four independent trials cell count at time zero, t* is the time when the log reduction
were conducted for each of the temperature/pH/acidulant/ value (LRV = log(N0/N)) equals 5, and β is a curve shape
pathogen combinations. parameter (β = 1 is linear, β < 1 is concave up, β > 1 is
Whirl-Pak bags containing inoculated tomato purée concave down). As described in (10), D* = t*/5, where D*
were kept fully submerged in a circulating water bath is the decimal reduction time.
(Thermo Scientific Phoenix II, Newington, NH) during For experiments with unacidified tomato purée (pH 4.5),
heating. The water level in the circulating water bath was analysis of the temperature dependence of D* was according
maintained at 1–2 cm above the level of the purée in the to the z-value of Eq. (2):
Whirl-Pak bag at all times. Internal purée temperature was
measured using a digital data logger (HH506RA Multi- log (D*T1/D*T2) = (T2-T1)/z, (2)
logger thermometer, OMEGA Engineering, Inc., Norwalk,
CT) connected to sterile type “K-TC” thermocouple where D*T is the D*-value at temperature T, and z is the
probes placed in the top 2–3 cm of the purée and in the temperature increase required for a decimal reduction in the
geometric center of the purée. The water bath tempera- D*-value.
ture was monitored using a calibrated mercury-in-glass The mean and standard error for each of the three pa-
thermometer. Preliminary experiments indicated no rameters of the re-parameterized nonlinear Weibull model
significant temperature differences (≤ 0.5°C) attributable (log(N0), t*, β) were determined by use of the regression
to location or time in tomato purée; the system was protocol of the R software package (R Foundation for Sta-
isothermal and at steady state. tistical Computing, Vienna, Austria) from all time points of
the thermal death curve. Significant differences of t* values
Enumeration of surviving cells for all trials in unacidified tomato purée at pH 4.5 were
At designated intervals, 1 ml samples of inoculated and determined with use of Tukey’s HSD test in the R statis-
heated tomato purée were drawn and diluted into pre-chilled tical software. Statistical differences between curve shape
(2–3°C) 9 ml BPD. Time-zero sampling occurred within parameter β and 1 for assessment of linearity for each of
20 s of inoculation and mixing. Comparison of the cocktail the 116 trials were calculated by use of a one-sample Wald
inoculum count to the time-zero count indicated an average z-test. D54°C values from acidification trials were compared
recovery of 102.6% (range 92.6–108.2%) for inoculum in in PROC MIXED (20) (SAS version 9.4, SAS Institute,
tomato purée (data not shown). Cary, NC) using a 3*2*2 factorial design with all possible
At each time point, samples were serially diluted in BPD interactions for three pathogen types (E. coli O157:H7, Sal-
at 21°C and 0.1 ml aliquots were spread plated onto tryptic monella, Listeria) * two acidulants (acetic and citric acids) *
soy agar (TSA; Difco) within 10 min of sampling. E. coli two pH levels (4.2 and 3.8). Additionally, D54°C –values for
O157:H7 and S. enterica culture plates were incubated at tomato purée at pH 4.5 (acid control) and at pH 4.2 and
37°C for 24 h, and L. monocytogenes culture plates at 37°C for 3.8, acidified with citrate or acetate, were compared using
48 h, and surviving pathogens enumerated. A sample of each a 3*5 full factorial design for all five treatments for each
453 g bag of blanched, frozen, and thawed tomato purée was pathogen, with no treatment factorial structure. Differences
plated on TSA to estimate native microbiota. Neither native in D54°C –values for acidification trials were determined by
microbiota nor comtaminants were ever counted, and they Fisher’s LSD test in the SAS statistical software. A signifi-
were rarely encountered in inoculated samples. cance cut-off value of 0.05 was used for all analyses.

Mathematical modeling and statistical analysis RESULTS

Log-linear plots of the bacterial count (CFU/ml) versus Pathogen D-values in tomato purée at 54°C ranged from
time were made for all trials; 52 trials were conducted with 20.98 min for E. coli O157:H7 (pH 4.5, unacidified) to 0.63
pathogen cocktails in unacidified tomato purée (pH 4.5) min for S. enterica (pH 3.8, acidified with acetic acid) (Table
at 52–58°C, and 64 trials were conducted with pathogen 2). Pathogen thermal inactivation curves were often not
cocktails in tomato purée acidified to pH 4.2 or 3.8 with linear (Weibull shape parameter (β) significantly different
citric or acetic acid. A successful trial included at least six from 1, P < 0.05) As Dufort et al. reported (10), 16 of 52
sampling points and resulted in a reduction of at least three- (31%) pathogen inactivation trials in unacidified tomato
log units in surviving cells. Five-log reduction times were purée resulted in nonlinear inactivation curves; likewise,

412 Food Protection Trends November/December

 TABLE 2. Decimal reduction times (D-value) for pathogen cocktails in tomato purée at 54°C

Pathogen E. coli O157:H7 L. monocytogenes S. enterica

pH Acidulanta nb D* (SD) minc n D* (SD) min n D* (SD) min

4.5d None 4 20.85 (1.67)A,a 6 10.63 (0.71)A,b 4 6.97 (0.81)B,c
4.2 Citric 4 16.39 (2.23)B,a 7 10.63 (1.20)A,b 4 6.12 (0.26)B,c
4.2 Acetic 4 9.40 (0.76)C,a 4 6.73 (1.10)B,b 14 3.79 (1.49)C,c
3.8 Citric 4 10.05 (0.87)C,a 4 8.68 (0.90)B,a 4 9.16 (2.24)A,a
3.8 Acetic 4 2.71 (0.32)D,b 4 3.49 (0.19)C,a 7 0.63 (0.21)D,c

a
Acidulant: None = unacidified, Citric = citric acid monohydrate, Acetic = glacial acetic acid.
b
n = Number of independent experiments.
c
Mean (standard deviation) of the D-value (min). Mean values within a column with different uppercase superscripts (A–D) are
significantly different (P < 0.05). Mean values within a row with different lowercase superscripts (a–c) are significantly different (P < 0.05).
d
Data at pH 4.5 (unacidified) from Dufort et al. (10).

in 38 of 64 (59%) thermal inactivation trials in acidified or acetic acid (P < 0.05). L. monocytogenes was the most
tomato purée (this study), the shape of the pathogen heat- and acid-resistant in purée acidified to pH 3.8 using
inactivation curve differed significantly from linearity acetic acid (P < 0.05), but in purée acidified to pH 3.8 using
(P < 0.05; data not shown). citric acid, there was no difference in pathogen inactivation
β values averaged 0.93 (± 0.20), 0.55 (± 0.15) and 0.91 (± (P > 0.05). Thermal tolerance in heated tomato purée
0.14) across all pH/acid combinations for E. coli O157:H7, at pH 4.5 or 4.2 followed the trend E. coli O157:H7 > L.
S. enterica, and L. monocytogenes, respectively. Inactivation monocytogenes > S. enterica (P < 0.05); at pH 3.8 with acetic
kinetics for E. coli O157:H7 in tomato purée acidified to pH acid as the acidulant, thermal tolerance followed the trend L.
3.8 with citric acid are depicted in Fig. 1. β values across the monocytogenes > E. coli O157:H7 > S. enterica (P < 0.05).
four trials ranged from 1.20 to 1.38; β = 1.30 for the average Data from previous work (10) were used to calculate
curve fit with the Weibull model (Fig. 1). Post-processing 5-log reduction times for E. coli O157:H7, S. enterica, and
pH measurements indicated that pH of heated tomato purée L. monocytogenes in tomato purée, pH 4.5 (Table 3). From
deviated only slightly from the initial pH value, fluctuating data on each pathogen, z-values were estimated based on the
-0.06 to 0.05 pH units across all heated purée samples and all D-values from nonlinear modeling of microbial inactivation
target pH values (n = 116; data not shown). as described previously (10, Table 4). The z-value for L.
Results from a 3*2*2 factorial statistical analysis indicated a monocytogenes, 7.95°C (14.3°F), was significantly higher than
high level of interaction between the acidulant, pH level, and the z-value for E. coli O157:H7 and S. enterica (P < 0.05),
pathogen (data not shown; P < 0.01). Pathogen inactivation which were not significantly different (P > 0.05). Using a
trends were not uniform across acidulant and pH level reference temperature of 71.1°C (160°F), the F-160 value
combinations. Subsequently, using a 3*5 factorial design, all was estimated for each pathogen using the respective z-values
D-values at 54°C were statistically compared as a one-way in °C (Table 4). Using pathogen z-values and D*ref , the
ANOVA with 15 treatments: three pathogens processed in data were extrapolated to the range of common processing
five different tomato purée conditions (unacidified at pH 4.5, temperatures (141–181°F/60.6–82.2°C); data for E. coli
citrate-acidified at pH 4.2 and 3.8, and acetate-acidified at O157:H7 and L. monocytogenes are shown in Fig. 2. E. coli
pH 4.2 and 3.8). D54°C–values were not significantly different O157:H7 was the most heat resistant of the three pathogens
at several combinations of pathogens and treatments (P > at temperatures below 65°C (149°F), but L. monocytogenes
0.05) (data not shown). Meaningful and relevant statistical was the most heat resistant at temperatures above 65°C
comparisons of pathogen inactivation across treatments are (Fig. 2). To establish a single set of parameters that ensure
shown in Table 2. Under experimental conditions, E. coli lethality of vegetative pathogens in the range of temperatures
O157:H7 was significantly more heat- and acid-resistant tested as well as those relevant to commercial processors,
than S. enterica and L. monocytogenes in tomato purée at pH the two endpoints for the plotted 5 log-reduction times for
4.5 (unacidified) and in purée acidified to pH 4.2 using citric the most heat tolerant organism at 52°C and 82.2°C, E. coli

November/December Food Protection Trends 413

 10.0

9.0

8.0
log CFU/ml

7.0

6.0

5.0

4.0
0 10 20 30 40

Time (min)

FIGURE 1. Thermal inactivation curve for E. coli O157:H7 pathogen cocktail at 54°C (129.2°F)
in tomato purée at pH 3.8, acidified with citric acid monohydrate. Four independent replications of
the data are shown (∆) and then fit using the re-parameterized Weibull model (–).

TABLE 3. Estimated 5-log reduction times in unacidified tomato purée pH 4.5a

Pathogen Temp °C Temp °F 5LR (SD)b

52 125.6 206.38 (22.44)
54 129.2 104.27 (8.36)
Escherichia coli O157:H7
56 132.8 42.59 (2.21)
58 136.4 27.63 (2.30)
52 125.6 100.75 (5.09)
54 129.2 53.14 (3.55)
Listeria monocytogenes
56 132.8 33.14 (2.74)
58 136.4 17.09 (1.79)
52 125.6 74.66 (6.40)
54 129.2 38.84 (4.06)
Salmonella enterica
56 132.8 19.24 (2.70)
58 136.4 9.00 (2.31)
a
Based on data from Dufort et al. (10).
b
Mean (standard deviation) calculated 5-log reduction (5LR) time (min).

414 Food Protection Trends November/December

 TABLE 4. z and F values in unacidified tomato purée at pH 4.5a

Pathogen z-val (°C)b z-val (°F)b F-160c (min)

Listeria monocytogenes 7.95A 14.3A 0.39

Escherichia coli O157:H7 6.64B 12.0B 0.27
Salmonella enterica 6.57 B
11.8B
0.09
Calculated 7.37 13.3 0.51
a
Summarized from Dufort et al. (10).
b
Estimated z-values within a column with different superscripts (A–B) are significantly different (P < 0.05).
c
Estimated 5-log reduction time at 160°F (71.1°C) (min).
d
Calculated from fitted line (Fig. 2).

3.0

2.0

1.0
Time for 5 logs (min)

0.0

-1.0

-2.0

-3.0
52 55 58 61 64 67 70 73 76 79 82

Temperature (°C)

FIGURE 2. The 5-log reduction times for cocktails of E. coli O157:H7 (circles) and L. monocytogenes (squares)
in the range of temperatures tested and at typical processing temperatures. The solid line is the time to attain at
least a 5-log pathogen reduction for E. coli O157:H7, L. monocytogenes, and S. enterica across all temperatures.

November/December Food Protection Trends 415

 TABLE 5. Calculated processing times to achieve at least a 5-log pathogen reduction
versus temperature in tomato-based acidified foods (pH 4.5 and below)

T (°F) T (°C) Time (mina) T (°F) T (°C) Time (min)

141 60.6 13.83 161 71.7 0.43
142 61.1 11.63 162 72.2 0.36
143 61.7 9.78 163 72.8 0.30
144 62.2 8.22 164 73.3 0.26
145 62.8 6.91 165 73.9 0.21
146 63.3 5.81 166 74.4 0.18
147 63.9 4.88 167 75.0 0.15
148 64.4 4.10 168 75.6 0.13
149 65.0 3.45 169 76.1 0.11
150 65.6 2.90 170 76.7 0.09
151 66.1 2.44 171 77.2 0.08
152 66.7 2.05 172 77.8 0.06
153 67.2 1.72 173 78.3 0.05
154 67.8 1.45 174 78.9 0.04
155 68.3 1.22 175 79.4 0.04
156 68.9 1.02 176 80.0 0.03
157 69.4 0.86 177 80.6 0.03
158 70.0 0.72 178 81.1 0.02
159 70.6 0.61 179 81.7 0.02
160 71.1 0.51 180 82.2 0.02

O157:H7 and L. monocytogenes, respectively, were used to fit the calculated z-values for pathogen reduction in cucumber
a straight line across the range of temperatures (solid line, Fig. juice medium (17.4, 15.6, and 16.7°F for E. coli O157:H7,
2). The fitted line was used to generate a table of processing Salmonella, and L. monocytogenes, respectively), were
conditions that ensure a reduction of at least 5 log units for higher than the z-values for the same pathogens in tomato
all pathogens tested across all temperatures (Table 5) and, purée, 11.9, 11.8 and 14.3°F (Table 4). The differences in
along with calculated values for z = 13.3°F (7.4°C) and F160 = D- and z-values may be attributed to the presence of natural
0.51 min (Table 4), can be used to support scheduled process inhibitory organic acids present in the tomato purée that
development and FDA process filing for tomato-based were not present in the cucumber juice medium as well as
acidified foods. to differences in experimental design and methodology that
could have impacted pathogen thermal tolerance, such as the
DISCUSSION level of dissolved oxygen present in the system (14).
Previous studies that established thermal processing The D-values of all three pathogens decreased significantly
conditions for acidified foods (pH 4.1–4.6) were conducted (P < 0.05) as the pH decreased (Table 2). Usaga et al. (24),
using a cucumber juice medium with acetic acid as the Gabriel (11), and Steenstrup et al. (21) each studied the
acidulant (pH 4.6) (6). The 5-log reduction times in thermal tolerance of a single strain of E. coli O157:H7 and
cucumber juice medium at 56°C were 126.10 min, 150.73 observed a similar pH-dependent trend when the pathogen
min, and 156.70 min, for E. coli O157:H7, S. enterica, and L. was heated in an apple-carrot juice blend (pH 3.3 to 4.5), in
monocytogenes, respectively. The calculated 5-log reduction a model fruit juice (pH 3.0 to 6.0), and in apple cider (pH
times for similar pathogen cocktails in tomato purée at 56°C 3.1 to 4.2), respectively. Across the pH and acid treatment
were significantly shorter, ranging from 42.59 min for E. coli combinations in this study at 54°C, E. coli O157:H7 was
157:H7 to 19.24 min for S. enterica (Table 3). Additionally, significantly more heat and acid resistant than S. enterica

416 Food Protection Trends November/December

and L. monocytogenes (P < 0.05) (Table 2). Breidt et al. (4) across a wide range of potential processing temperatures
found that thermal inactivation rates of E. coli O157:H7 and (60.6–82.2°C/141–180°F) and showed that L. monocytogenes
L. monocytogenes in pickle brine (pH 4.1) were identical, was more heat resistant than E. coli O157:H7 and S. enterica
while S. enterica was significantly less heat tolerant. Similarly, at temperatures below 74°C (166°F), but E. coli O157:H7
Mazzotta (16) found that, in heated fruit juices (pH 3.9), was the most heat resistant at temperatures above 74°C. The
S. enterica was heat sensitive while E. coli O157:H7 was present work found that when 5-log reduction values were
heat resistant. Acidification of a product reduces the heat plotted against a wide range of typical processing tempera-
resistance of vegetative cells and reduces the possibility of tures, data lines for E. coli O157:H7 and L. monocytogenes
recovery from sub-lethal injury (24). In acidified tomato intersected at 65°C; E. coli O157:H7 was the most heat-
purée samples, all three pathogens were significantly more resistant pathogen at temperatures below 65°C (149°F), but
heat tolerant when purée was acidified with citric acid rather L. monocytogenes was the most heat resistant above 65°C
than with acetic acid (P < 0.05; Table 2). When tomato (Fig. 2). While the physiological explanation for this differ-
purée was acidified to pH 3.8 with acetic acid rather than ence in z-values between the two pathogens is not completely
with citric acid, the average D54°C –values were 7.34 min, 8.53 understood, the difference in z-values is important to consid-
min, and 5.19 min lower for E. coli O157:H7, S. enterica, er when calculating processing time/temperature combina-
and L. monocytogenes, respectively. In heated cucumber tions. Therefore, to establish a single set of parameters that
purée, Bae et al. also found that E. coli O157:H7, S. enterica ensure lethality of vegetative pathogens across a wide range
and L. monocytogenes were more resistant to citric acid than of relevant temperatures, a straight line was created that
to acetic acid (3). The antimicrobial effects of citric and ensured at least a 5-log pathogen reduction across all temp-
acetic acid are dependent on their pKa values (23). Acetic eratures. The linear equation that defined the straight line was
acid (pKa 4.76) is a stronger acid than citric acid, which is a used to generate a table of calculated time-temperature pairs
triprotic acid (pKa1 3.13, pKa2 4.77, pKa3 6.39). In acidified (Table 5) that would ensure at least a 5-log reduction for all
tomato-based foods with pH 4.5 and below, acetic acid has pathogens tested, similar to the method of Breidt et al. (6).
a greater proportion of undissociated acid than citric acid Processors of tomato-based acidified foods should consult a
has, and acetic acid is more likely than citric acid to diffuse competent process authority to determine additional heating
into the cell, lower cytoplasmic pH, and inhibit metabolic requirements beyond the minimal recommended here, to
reactions (1, 17). While this phenomenon may at least account for formulation effects on thermal processing, to ad-
partially explain our findings, our experimental design dress spoilage, and/or to address destruction of acid-resistant
was aimed only at achieving a target pH-value and did not microbiota and other relevant factors in the establishment of
take into account either buffering capacity of the system the scheduled process.
or actual quantity of acid added. For industrial food safety
applications, it is advantageous to use organic acids with RECOMMENDATIONS
a pK a near the pH of the food environment. However, the Using the re-parameterized Weibull model, we have
amount and concentration of acid required to achieve the calculated time/temperature conditions that ensure at least
target pH value, the associated flavor characteristics, and a 5-log reduction of vegetative pathogens of concern in
the labeling provision of an acidulant are also important tomato-based acidified canned foods (pH ≤ 4.5).
factors to consider when selecting acidulants for product • Processing time and temperature combinations were
formulation (12). calculated that can be applied to acidified products
In previous work from our laboratory, pathogen inactiva- with tomato as the primary ingredient/formulation
tion in unacidified tomato purée (pH 4.5) was modeled over base. These recommendations support safety while
a range of temperatures (52–58°C) using the re-parameter- avoiding over-processing. Processors will wish to take
ized Weibull method and the shape of the inactivation curve into account product formulation and the potential
(β) was evaluated, with nonlinearity frequently observed presence of spoilage organisms in using these minimal to
(10). In the current study, when pathogen inactivation was establish operational processing parameters.
modeled in acidified tomato purée (pH 4.2 and 3.8) as • Acidification of tomato-based canned foods using
compared to unacidified purée, nonlinearity in heat inacti- citric or acetic acid will increase pathogen lethality
vation curves increased from 31% to 59% of trials (data not across all heating temperatures, with acetic acid having
shown), suggesting that a nonlinear curve-fit model, such greater antimicrobial activity than citric acid.
as the Weibull method, may be of even greater utility when • Processing times and temperatures achieve at least
establishing accurate thermal processing parameters in low a 5-log pathogen reduction across a wide range
pH food systems. of anticipated processing conditions. Modeling
Using data derived from a cucumber juice medium, pathogen survival across typical commercial processing
Breidt et al. (6) calculated thermal processing parame- temperatures indicated that the target organism depends
ters for pathogen destruction in acidified cucumber juice on temperature; E. coli O157:H7 had the greatest

November/December Food Protection Trends 417

 heat tolerance at temperatures below 65°C, while ACKNOWLEDGMENTS
L. monocytogenes had the greatest heat tolerance at The National Institute of Food and Agriculture, United
temperatures above 65°C. States Department of Agriculture provided funding for this
• Calculated minimum processing time/temperature work under agreement 2011-51110-31019. Gratitude is
combinations (Table 5) and use of a z-value of 13.3°F expressed to Jonathan Sogin, Wan Mei Leong, and Tyler
(7.4°C), along with consideration given to intrinsic Laux for laboratory assistance, and to Dr. Fred Breidt, USDA-
product characteristics and acid tolerant spoilage ARS for expert advice in the conduct of this research and
microbiota, can be used by Process Authorities to interpretation of data.
develop scheduled processes or support FDA process
filing for tomato-based acidified food products.

REFERENCES
1. Al-Nabulsi, A. A., A. N. Olaimat, T. M. Osaili, 10. Dufort, E. L., J. Sogin, M. R. Etzel, and B. H. 18. Park, S. H., M. R. Choi, J. W. Park, K. H.
R. R. Shaker, N. Z. Elabedeen, Z. W. Jaradat, Ingham. Inactivation kinetics of pathogens Park, M. S. Chung, S. Ryu, and D. H. Kang.
A. Abushelaibi, and R. A. Holley. 2014. Use during thermal processing in acidified broth 2011. Use of Organic acids to inactivate
of acetic and citric acids to control Salmonella and tomato purée (pH 4.5). J. Food Prot. Escherichia coli O157:H7, Salmonella
Typhimurium in tahini (sesame paste). Food (In Press.) Typhimurium, and Listeria monocytogenes on
Microbiol. 42:102–108. organic fresh apples and lettuce. J. Food Sci.
11. Gabriel, A. A. 2012. Influences of heating
76:M293–M298.
2. Albert, I., and P. Mafart. 2005. A modified temperature, pH, and soluble solids on the
Weibull model for bacterial inactivation. decimal reduction times of acid-adapted 19. Saxton, A.M. 1998. A macro for converting
Internat. J. Food Microbiol. 100:197–211. and non-adapted Escherichia coli O157:H7 mean separation output to letter groupings in
(HCIPH 96055) in a defined liquid Proc. Mixed. In Proc. 23rd SAS Users Group
3. Bae, Y. M., and S. Y. Lee. 2015. Combined
heating medium. Internat. J. Food Microbiol. Intl., SAS Institute, Cary, NC, p. 1243–1246.
effects of organic acids and salt depending
160:50–57.
on type of acids and pathogens in laboratory 20. Shin, J. H., S. Y. Lee, R. H. Dougherty, B.
media and acidified pickle. J. Appl. Microbiol. 12. Hartwig, P.D., and M.R. McDaniel. 1995. Rasco, and D. H. Kang. 2006. Combined
119:455–464. Flavor characteristics of lactic, malic, citric, effect of mild heat and acetic acid treatment
and acetic acids at various pH levels. J. Food for inactivating Escherichia coli O157: H7,
4. Breidt, F., J. S. Hayes, and R. F. McFeeters.
Sci. 60:384–388. Listeria monocytogenes and Salmonella
2004. Independent effects of acetic acid and
typhimurium in an asparagus purée. J. Appl.
pH on survival of Escherichia coli in simulated 13. International Marketing Bureau. March 2013.
Microbiol. 101:1140–1151.
acidified pickle products. J. Food Prot. Consumer Trends — Sauces, Dressings and
67:12–18. Condiments in the United States — 6362- 21. Steenstrup, L. L., and J. D. Floros. 2002.
[Link]. Available at: [Link] Statistical modeling of D- and z-value of
5. Breidt, F., Jr., J. S. Hayes, J. A. Osborne, and
resources/prod/Internet-Internet/MISB- E. coli O157:H7 and pH in apple cider
R. F. McFeeters. 2005. Determination of
DGSIM/ATS-SEA/PDF/[Link]. containing preservatives. J. Food Sci.
5-log pathogen reduction times for heat-
Accessed 28 February 2017. 67:793–796.
processed, acidified vegetable brines. J. Food
Prot. 68:305–310. 14. Kreske, A. C., K. Bjornsdottir, F. Breidt, Jr., 22. United States Food and Drug Administration.
and H. Hassan. 2008. Effects of pH, dissolved 2016. Code of Federal Regulations. Acidified
6. Breidt, F., K. Kay, J. Osborne, B. Ingham,
oxygen, and ionic strength on the survival Foods, 21 C.F.R. § 114. Available at http://
and F. Arrrtt. 2014. Thermal processing of
of Escherichia coli O157:H7 in organic acid [Link]/cgi-bin/text-idx?SID=307216
acidified foods with pH 4.1 to pH 4.6. Food
solutions. J. Food Prot. 71:2404–2409. 6014a8e3925533f307c464efc5&mc=true&n
Prot. Trends 34:132–138.
ode=pt21.2.114&rgn=div5#se21.2.114_13.
15. Mani-Lopez, E., H. S. Garcia, and A. Lopez-
7. Breidt, F., K. P. Sandeep, and F. M. Arritt. Accessed 10 January 2017.
Malo. 2012. Organic acids as antimicrobials
2010. Use of linear models for thermal
to control Salmonella in meat and poultry 23. Usaga, J., R. W. Worobo, and O. I. Padilla-
processing of acidified foods. Food Prot.
products. Food Res. Internat. 45:713–721. Zakour. 2014. Thermal resistance parameters
Trends 30:268–272.
of acid-adapted and unadapted Escherichia
16. Mazzotta, A. S. 2001. Thermal inactivation of
8. Buchanan, R. L., and S. G. Edelson. 1999. coli O157:H7 in apple-carrot juice blends:
stationary-phase and acid-adapted Escherichia
pH-dependent stationary-phase acid effect of organic acids and pH. J. Food Prot.
coli O157:H7, Salmonella, and Listeria
resistance response of enterohemorrhagic 77:567–573.
monocytogenes in fruit juices. J. Food Prot.
Escherichia coli in the presence of various
64:315–320. 24. Zareifard, M. R., T. Savard, M. Marcotte,
acidulants. J. Food Prot. 62:211–218.
J. Y. Lecompte, and S. Grabowski. 2015.
17. Mortimore, S., and C. Wallace. 2013.
9. Cortesi, M. L., A. Panebianco, A. Giuffrida, Combined effects of mild acidification and
HACCP, a Practical Approach, third ed.
and A. Anastasio. 2009. Innovations in thermal treatment on color and texture
Springer, New York (Chapter 3).
seafood preservation and storage. Veter. Res. of green beans. Intern. J. Food Propert.
Comm. 33:S15–S23. 18:1847–1862.

418 Food Protection Trends November/December