Full text of DOI:10.1111/j.1365-2621.2003.tb09668.x

JFS:FoodMicrobiologyandSafety
Effect of Modified Atmosphere Packaging
and Superchilled Storage on the
Microbial and Sensory Quality of
Atlantic Salmon (Salmo salar) Fillets
M. SIVERTSVIK, J.T. ROSNES, AND G.H. KLEIBERG
ABSTRACT: Fresh Atlantic salmon fillets packaged under modified atmosphere (MA) (CO2:N2 60:40) and air was
stored at superchilled (–2 °C) and chilled (+4 °C) temperatures. Changes in sensory scores, microbial growth, headspace
gas composition, water loss, and pH were monitored during 24 d of storage. The superchilled MA packaged salmon
maintained a good quality, with negligible microbial growth (<1000 colony-forming units [CFU]/g) for more than
24 d based on both sensory and microbial analyses (aerobic plate count, H2S-producing, and psychrotrophic
bacteria). Superchilled salmon in air had a 21-d sensory shelf life, whereas MA and air-stored fillets at chilled
conditions was spoiled after 10 d and 7 d, respectively.
Keywords: modified atmosphere packaging, superchilling, partial freezing, Atlantic salmon, shelf life
Introduction
ODIFIED ATMOSPHERE PACKAGING
(Dalgaard 2000). At high CO2 concentra- introduced onboard trawlers in the 1960s
tions, P. phosphoreum has been found to be (Pearson 1980), the temperature of the fish
the main spoilage bacteria in both MA-pack- is reduced to 1 °C to 2 °C below the initial
aged chilled cod (Dalgaard and others 1993; freezing point, and some ice is formed inside
Dalgaard 1995) and salmon (Emborg and the product (Haard 1992; Sikorski and Sun
others 2002). Application of knowledge 1994). How much of the water is frozen is
about specific spoilage organisms for deter- highly temperature-dependent (Huss
mination, prediction, and extension of shelf 1995). Freezing points for fishery products
life is of vital importance to extend the shelf vary from about –1 °C to –2.5 °C, such as in
life and to obtain a high raw material quality salmon, shrimp, and mackerel at about
M
(MAP) of fishery products has been
thoroughly investigated since the initial
works in the 1930s (Killeffer 1930; Coyne
1932, 1933; Stansby and Griffiths 1935) and
in particular for the last 2 decades (Siverts-
vik and others 2002). Shelf life extensions of
fish packaged under MAP has been report-
ed from 50% to 400% (Stammen and others
1990), but often it is only the period of mod-
erate to low quality that has been extended
and not the initial period of prime quality
(Fletcher and Statham 1988). Usually the
effect of MAP is conditioned by the concen-
trations and amount of CO2 available in the
gas atmosphere, availability of O2, the qual-
ity of the raw materials, and most important-
ly, the storage temperature (Sivertsvik and
others 2002). CO2 inhibits growth of the nor-
mal spoilage flora in air (such as Pseudomo-
nas and Shewanella putrefaciens) and during
storage; CO2-tolerant microorganisms (such
as Lactobacillus spp., Photobacterium phos-
phoreum, Brochothrix thermosphacta) will
dominate the microflora. Usually a single
species is responsible for the main sensory
spoilage: the specific spoilage organisms
(SSO). The numbers of SSO and the concen-
tration of their metabolites can be used as
objective indices of spoilage in shelf-life
determinations. With knowledge on the
microorganisms responsible for spoilage, a
close relation between log-numbers of SSO
and remaining shelf life can be expected
over a longer period.
–2.2 °C to carp at about –1.0 °C (Rahman
Atlantic salmon (Salmo salar) is an impor- and Driscoll 1994) and are usually depen-
tant product both from an economical and dent on the water content in the products
nutritious perspective. The maximum shelf (Chang and Tao 1981). Superchilling can
life for iced whole salmon is about at 20 d either be used prior to traditional chilled dis-
(Sveinsdottir and others 2002), and for tribution to store refrigerating capacity into
salmon steaks/fillets in MAP at chilled tem- the product (Magnussen and others 1998)
peratures (2 °C to 4 °C), shelf lives of 14 to 21 or the superchilling temperature is main-
d have been observed (Pastoriza and others tained throughout the storage and distribu-
1996; Randell and others 1999; de la Hoz tion. Storage at superchilled conditions may
and others 2000; Emborg and others 2002). enhance phospholipid hydrolysis and pro-
This shelf life is sufficient to distribute both tein denaturation (Ashie and others 1996),
whole salmon and MAP salmon within mar- but superchilling will inhibit most autolytic
kets closed to the salmon-farming industry, and microbial reactions, and thereby in-
but for markets farther away from the main crease shelf life (Huss 1995; Chang and oth-
salmon producers in Norway, Scotland, and ers 1998).
Chile, such as North America, Asia, and Oce-
Few articles on the combined effect of
ania, the relative short shelf life either ne- superchilling under modified atmospheres
cessitates expensive air freight, frozen dis- have been published and, to our knowl-
tribution, or inclusion of more preservative edge, none have been published on salmon
factors.
fillets. Bulk-packaged, superchilled, whole,
One of the few methods with the poten- gutted salmon combined with a high-CO2
tial to extend the period of prime quality in atmosphere maintained a high microbiolog-
fish is “superchilling” or “partial freezing” ical and sensory quality for more than 3 wk
(Haard 1992). In this technique, originally (Sivertsvik and others 1999). A shelf life of
© 2003 Institute of Food Technologists
Further reproduction prohibited without permission
Vol. 68, Nr. 4, 2003—JOURNAL OF FOOD SCIENCE 1467

Storage of superchilled MAP salmon . . .
Table 1—Experimental design and response variables
21 d was observed for mackerel at –2 °C in
100% CO2 (Hong and others 1996), and for
smoked blue cod, a doubling of shelf life
was observed when lowering the storage
temperature from 3 °C to –1.5 °C (Penney
and others 1994).
Design variables
Levels
Temperature (A):
Packaging method (B):
Storage time (C):
–2 or 4 °C
Air exposed (overwrap) or anoxic CO2-exposed (MAP)
7, 10, 14, 17, 21, or 24 d
Nr of experiments:
24 (full factorial design)
The objective of this article was to com-
pare the combined effect of MAP and su-
perchilled conditions on the storage quality
of salmon fillets.
Response variables
Analyses
Microbiological:
Aerobic plate count (APC), H2S-producing bacteria,
psychrotrophic bacteria
Raw and cooked odor, cooked flavor, firmness, and juiciness
Temperature, pH, water/drip loss, head-space gas composition
Sensory evaluation:
Chemical/other:
Materials and Methods
Raw material, packaging, and
storage
Three-d-old iced-stored, whole, gutted
Atlantic salmon (Salmo salar) of size 3 to 4 kg
was obtained from a commercial fish farm
(Marine Harvest AS, Hjelmeland, Norway).
The salmon were filleted and cut into small-
er fillets/steaks (131 3 g each) without
skin or bone. The filleting, deskinning, and
cutting were carried out at the production
area at the NORCONSERV Inst. of Fish Pro-
cessing and Preservation Technology under
high hygienic conditions to obtain raw ma-
terial with low initial bacterial counts. The
equal-sized fillets where individually pack-
aged under modified atmosphere (MA) or in
traditional overwrap packaging with expo-
sure to air. The MA packages were high-den-
sity polyethylene semirigid trays (Dyno-
pack, Polimoon, Kristiansand, Norway). The
air was evacuated, and the gas mixture (60%
CO2 and 40% N2) introduced into the pack-
age before heat sealing (lidding film: 15 my
PE/75 my PA, Dynoseal ST 1575, Polimoon)
tone (w/v) for 120 s in a Stomacher 400 Lab-
oratory Blender (Seward Medical, London,
U.K.). Total viable counts, measured as aer-
obic plate counts (APC) and H2S-producing
bacteria were measured after a suitable di-
lution had been added to melted and tem-
perate (44 °C) iron agar (Agar Lyngby, IA,
Oxoid CM 867, Basingstoke, Hampshire,
U.K.) supplemented with L-cysteine and
stored at 20 1 °C for 3 d. Black colonies
were counted as H2S-producing bacteria;
APC were counted as the total of black and
white colonies. The content of psy-
chrotrophic bacteria was determined by a
spread plate count method on solidified
chilled plate count agar (PCA, Merck, Darn-
stadt, Germany) added 1% NaCl, and incu-
bated at 8 °C for 5 to 7 d. Averaged results of
triplicate measurements are presented as
log colony-forming units (CFU) per gram
salmon.
Temperature measurements
Sample’s core temperatures were mea-
sured every 5 min during storage using elec-
tronic temperature loggers (Ebro Electronic,
Ingolstadt, Germany) at both temperatures
and in both packaging methods.
Gas analyses
The gas composition in the packages was
determined in triplicate by injecting an ali-
quot (30 mL) of the headspace gas of the
trays using an oxygen and carbon dioxide
analyzer (M.A.P. Test 4000, Hitech Instru-
ments, Luton; U.K.). The gas was collected
with a syringe after intrusion of the top foil.
The analyzer was calibrated against a certi-
fied gas mixture (O2:CO2:N2 1.1:44.1:54.8)
and air before each sampling.
pH measurements were made in tripli-
cate with a pH meter (Beckman 72, Dan
Meszansky, Oslo, Norway) on 25 g of homo-
genate of salmon muscle with 25 mL of 0.1 M
KCl in distilled water.
on a semiautomatic packaging machine Sensory evaluation
(Dyno VGA 462, Polimoon). Food grade CO2
and N2 (AGA, Oslo, Norway) was mixed in a
gas mixer (Witt KM 100-2m, Witt Gasetech-
nick, Witten, Germany) to obtain the gas
mixture, and the gas volume-to-product (g/
p) ratio was approximately 1 (100 mL gas/
100 g salmon).
The overwrap packages were expanded
polystyrene trays with built in exudate ab-
sorber between layers in the bottom of tray
(LinPac, Marietta, Ga., U.S.A.) and wrapped
with cling film with low barrier properties
(PVC-film, ITS Foil and Film Rewinding b.v.,
Apeldoorn, The Netherlands). Salmon pack-
ages of both types were stored in chill cabi-
nets (Porkka CM710, Huurre Group, Hollo-
la, Finland) set to –2 °C and +4 °C,
respectively, and evaluated regularly from
0 (day of packaging) and up to 24 d accord-
ing to the experimental design (Table 1).
Sensory evaluation was carried out on
both raw- and cooked-salmon samples using
a panel of 4 assessors trained according to
the ISO-standard (ISO 1993). All sensory
evaluations were carried out in randomized
order of coded samples. The odor of salmon
at package opening was evaluated in tripli-
cate with a quality scale from 10 to 1 (10: sea
fresh, typical of species; 1: faecal, indole and
skatole, adopted from Shewan and others
1953). Test pieces (15- to 20-mm-wide cut of
the stored samples) for cooked salmon eval-
uation were packaged in cook-plastic pouch-
es (PA/PE 20/50) under slight vacuum (95%)
and cooked in water vapor (80 °C in 10 min)
without any salt or spice addition. The sam-
ples were evaluated in duplicate using a de-
scriptive test (adopted from Shewan and oth-
ers 1953) with a scale from 10 (fresh sea
weedy odor, fresh sweet flavor and firm, suc-
culent texture) to1 (putrid odor and flavor;
sloppy, dry texture) for odor, flavor, and tex-
ture (firmness and juiciness). For both raw
and cooked salmon, a score of 5 was chosen
as the minimum acceptance level.
Formation of drip
The drip formed in the samples during
storage was measured gravimetric. The
mass of the drip (g) was divided by the ini-
tial mass of product (g) and reported as a
percentage (%).
Statistical analysis
Partial least squares (PLS) regression with
full cross-validation (leave-one-out) to find
correlation (R2), random mean square error
of prediction (RMSEP) and regression coeffi-
cients, and analyses of effects (multiple lin-
ear regression, MLR) to find the significant ef-
fects (Tukey) of the design variables on the
responses were performed using Unscram-
bler/Guideline+7.6 (Camo, Oslo, Norway).
Univariate analysis of variance were per-
formed with Minitab 13.3 (Minitab, Coven-
try, U.K.) using Tukey’s HSD test at level
P < 0.05 (95%) to obtain confidence intervals
for all pairwise differences between level
means of temperature and packaging type
on each sampling time (d of storage).
Microbiological examinations
Samples of 25 g were aseptically taken
from the salmon fillets and homogenized in
250 mL of 0.9% NaCl (w/v) and 0.1% pep-
1468 JOURNAL OF FOOD SCIENCE—Vol. 68, Nr. 4, 2003
JFS is available in searchable form at www.ift.org

Storage of superchilled MAP salmon . . .
Table 2—Regression coefficients (from PLS-2 model with 4 components) for
Results and Discussion
the microbial analysis, water loss, and sensory scores against design variables
and interactions. Significant effects (analysis of effect, MLR) of the design vari-
ables and interaction on the sensory score is marked with *** (P < 0.005), **
(P < 0.01), and * (P < 0.05).
HE GAS COMPOSITION IN THE MA
T
packages was measured to 61.1% 0.3%
CO2, 0.2% 0.1% O2 immediately after pack-
aging, and the remaining gas was N2. After
packaging, CO2 was dissolved into the prod-
uct. Seven d after packaging, the CO2 level
in the headspace of the chilled MA packages
was reduced to 34.0% 0.8% and in the su-
perchilled to 31.1% 0.5%, giving higher
concentration of dissolved CO2 in the super-
chilled salmon compared with the chilled.
Solubility of CO2 in water increases with
decreasing temperature (Carroll and others
1991); however, CO2 will not dissolve into ice
(Monnin and others 2001). The fact that
superchilled MA salmon dissolved more
CO2 (P < 0.01) compared with the chilled
salmon indicates little ice formation in the
Aerobic
plate
count
H2S-
producing
bacteria
Psychro-
trophic
bacteria
Drip
loss
Effect
Temperature (A)
MAP (B)
Storage time (C)
A*B
A*C
B*C
0.770***
–0.526***
0.268*
0.087
–0.043
–0.180
0.683
0.618***
–0.573***
0.332*
–0.249**
0.099
–0.245
0.517
0.893
0.810***
–0.442***
0.297*
0.199**
–0.032
–0.098
0.679
0.517**
0.171
0.498
–0.406*
–0.159
0.240
RMSEP
0.937
0.490
Correlation (R2)
0.923
0.937
Odor, raw Odor, cooked Flavor
Firmness
Juiciness
Temperature (A)
MAP (B)
Storage time (C)
A*B
–0.688***
0.448***
–0.458*
0.095
–0.260
0.092
–0.678***
0.293**
–0.543**
0.139
–0.285*
–0.027
0.784
–0.682***
0.304*
–0.524*
0.012
–0.338
–0.035
0.717
–0.572***
0.400**
–0.495*
0.210
–0.375
0.071
–0.599***
0.390**
–0.502*
0.202*
–0.347*
0.061
superchilled product. The CO2 level de- A*C
B*C
creased linearly from day 7 (31.1%) to day
RMSEP
0.430
0.935
0.572
0.826
0.478
0.856
Correlation (R2)
0.864
0.873
stored at chilled conditions was 4.4 0.2 °C
(4.0/5.1 °C min/max), and for the super-
chilled samples the mean temperature was
–1.9 0.1 °C (–2.0/–1.9 °C min/max) during
24 d of storage. No differences in tempera-
tures between the packaging methods were
observed.
Microbial analysis of the raw material af-
ter filleting and cutting showed that the
salmon fillets were produced at very hy-
gienic conditions and had a very high mi-
crobial quality (APC < 25 CFU/g salmon).
APC, H2S-producing bacteria, and psy-
chrotrophic bacteria counts increased dur-
ing storage (Figure 1a through 1c) for both
temperatures and packaging types, but
only negligible growth was observed for MA
packaged salmon stored at superchilled
temperature (CFU/g < 1000). Growth was
always higher in salmon exposed to air com-
pared with salmon in MA at same tempera-
ture and storage time.
The microbial counts had very good corre-
lations (PLS-regression) with the design
variable (R2 [APC] = 0.92, R2 [H2S] = 0.89,
and R2 [psychrotrophic] = 0.94) and a pre-
diction within the boundaries of the design
levels, with less than 0.7 log CFU/g accura-
cy can be performed (RMSEP: 0.7, 0.5, and
0.7 for APC, H2S-producing, and psy-
chrotrophic counts, respectively).
PLS-regression coefficients and analysis
of effects (Table 2) showed the highly sig-
nificant effect of the storage temperature
and packaging method (P < 0.005) on micro-
bial growth for all 3 analyses. Not surpris-
ingly, a significant effect of storage time
24 (23.8%) of superchilled storage and even
faster for the chilled-MA packages to 15%
after 14 d, probably caused by permeation
of CO2 out of the HDPE trays. According to
Henry’s law, some of the dissolved CO2
would be released from the product when
the partial pressure of CO2 in the packages
atmosphere decreases due to the perme-
ation of CO2. Using a packaging material
with better barrier properties would proba-
bly enhance the effect of MAP, since this
would maintain a higher partial pressure of
CO2 inside the package, and the inhibiting
effect on many spoilage microorganisms is
dependent on the amount of dissolved CO2
in the product (Devlieghere and Debevere
2000). Use of a larger g/p ratio would also in-
crease the effect of MAP, since this would in-
crease the dissolvement of CO2. The O2 in
the MA packages never exceeded 1% during
storage, and the O2 permeating through the
packaging materials would probably be con-
sumed by aerobic bacteria inside the pack-
ages.
Atmosphere composition inside the over-
wrap packages was close to air composition
throughout the storage time, 20.9% O2 and
0.0% CO2 during 21 d of superchilled air
storage and during 10 d of chilled air storage,
confirming good exposure of the salmon to
air through the PVC film. For the chilled air
salmon, a decrease in O2 to about 17% and
an increase in CO2 level to about 2% of the
packages atmospheres were observed, coin-
ciding with high microbial activity (de-
scribed subsequently).
Figure 1—Growth of (a) aerobic plate
count (APC), (b) H2S-producing bacte-
ria, and (c) psychrotrophic bacteria in
salmon fillet during storage. Legend ᭺:
chilled, overwrap (air); ᭹: chilled, MAP;
ᮀ: superchilled, overwrap (air); and ᭿:
superchilled, MAP.
The average temperature in the samples
JFS is available in searchable form at www.ift.org
Vol. 68, Nr. 4, 2003—JOURNAL OF FOOD SCIENCE 1469

Storage of superchilled MAP salmon . . .
Table 3—Sensory evaluation of salmon fillets packaged in MA or air and stored
at superchilled or chilled conditions. Mean results of 2 (3 for raw odor) repli-
cates times 4 judges.
(P < 0.05) also was observed; however, this
effect is relatively smaller than the 2 other
experimental variables. The only significant
interaction effect was the temperature–
packaging method (AB) effect on H2S-pro-
ducing bacteria and psychrotrophic bacte-
ria. The combination of CO2 and
superchilling had a synergistic effect on the
inhibition of psychrotrophic growth, where-
as the superchilled and overwrap packaging
had a synergistic effect on H2S-producing
counts. The psychrotrophic counts for MAP
salmon at 4 °C was about 20% higher than
APC, while for the other samples the psy-
chrotrophic counts was very similar to APC,
indicating an enhanced growth of a psy-
chrotrophic bacteria, in the chilled MAP
salmon. The specific spoilage organism
(SSO) of MAP salmon has recently been
found to be P. phosphoreum (Emborg and
others 2002), the same SSO as previously
found for MAP cod (Dalgaard and others
1993; Dalgaard 1995). P. phosphoreum has
been shown to grow on PCA-added salt in
previous trials at our institute, but the colo-
nies from this experiment were not charac-
terized. Assuming the psychrotrophic count
includes P. phosphoreum, then the low num-
bers indicated a very sparse growth under
MA at –2 °C during 24 d of storage. Charac-
terization of specific strains at superchilled
temperatures should be included in further
studies. The superchilled salmon exposed
to air had relatively lower counts of H2S-pro-
ducing bacteria when compared with APC
than chilled salmon exposed to air. S. putre-
faciens may not contribute to all black colo-
nies on iron agar, although this bacteria has
often been shown to be the main H2S pro-
ducer on fresh fish stored in ice (Gram and
others 1987; Dalgaard 1995); but if so, these
results indicate an inhibitive effect on S.
putrefaciens at superchilled conditions.
Water loss or drip formation from the sam-
ples was significantly increased by temper-
ature (P < 0.01) and temperature–MAP in-
teraction (P < 0.05) (Table 2). The drip was
highest in the chilled overwrap samples (4%
to 5%), and lowest in the superchilled over-
wrap (about 2%). The 2 MA packages had
almost the same loss of water (about 3%),
lower than in the chilled air but higher than
in superchilled air storage. It has been sug-
gested by several authors that dissolved
CO2 in the products could decrease the wa-
ter-holding capacity (Sivertsvik and others
2002). However, for salmon this effect is not
pronounced (Randell and others 1999): typ-
ical drip loss at about 2% under 60% CO2 at-
mosphere at 2 °C compared with about 1%
for vacuum packaging and about 0.5% in air
storage. Our results suggest that the storage
temperature is more important, and super-
Days of storage*
Packaging/
Parameter
storage
0
7
10
14
17
21
24
Odor, raw
Raw material
Air, superchilled
Air, chilled
MA, superchilled
MA, chilled
8.8
7.5ab 6.8b
6.6b 5.7b
8.4a 8.2a
7.0b 6.3b
6.8a 6.4b 6.2b 6.1b
3.9c ne ne ne
7.7a 7.6a 7.4a 7.8a
5.8b 5.8b 4.6c ne
Odor, cooked
Raw material
Air, superchilled
Air, chilled
MA, superchilled
MA, chilled
9.0
9.0
9.0
9.0
7.5
5.9
8.0
7.6
6.9a
3.8b
7.4a
6.8a
5.7ab 6.6a 6.4a 5.0b
2.6c ne
ne
ne
7.0a 7.3a 6.3a 7.0a
3.7bc 3.3b 3.1b Ne
Flavor, cooked Raw material
Air, superchilled
Air, chilled
7.4ab 7.0a
6.3b 4.0b
8.3a 7.8a
8.0a 6.8a
5.5
3.5
7.7
5.3
6.1b 6.0a 5.3b
ne ne ne
7.9a 7.0a 7.3a
3.3c 2.4b ne
MA, superchilled
MA, chilled
Raw material
Air, superchilled
Air, chilled
MA, superchilled
MA, chilled
Raw material
Air, superchilled
Air, chilled
Firmness
7.3
7.6
8.0
8.1
7.0ab 7.2
7.3ab 6.8
ne ne
8.0a 6.9
5.8b 5.5
7.1
ne
7.9
ne
5.2b
7.3a
7.1a
5.0
7.8
6.5
Juiciness
6.8
6.7
7.8
7.6
7.1
5.8
7.5
7.3
7.3a 7.0ab 6.5
5.0b ne ne
7.2ab 7.8a 6.8
5.5ab 6.1b 5.4
6.3
ne
6.4
ne
MA, superchilled
MA, chilled
*Equal or no lowercase letter within column and parameter = no significant difference (P > 0.05). ne = not
evaluated due to spoilage.
chilled storage at –2 °C does not lead to ex- superchilling and MAP on the sensory qual-
cessive drip caused by cell destruction due ity of salmon. Significant interaction effects
to freezing. However, the results from the were observed for the juiciness of salmon in
drip measurements had large variance, the temperature * MAP combination and for
and PLS regression against design variables juiciness and cooked odor in the tempera-
showed low correlation (R2 = 0.49) and a ture * storage combination, showing syner-
high RMSEP (0.93). No differences among gistic effects of these interactions.
design variables in salmon pH were ob-
Significant differences observed in the
served; the pH was 6.2 to 6.3 for all mea- individual sensory parameters within each
surements, except for a pH increase to 6.7 sampling days are shown in Table 3. The
after spoilage in the overwrapped salmon salmon exposed to air and stored at 4 °C
stored at 4 °C for 21 d.
had a shelf life of 7 d based on the sensory
The sensory scores also showed good cor- scores, and odor (cooked and raw) in partic-
relation with the design variables (R2 = 0.93 ular. MA-stored salmon at the chilled tem-
[raw odor], 0.86 [cooked odor], 0.87 [flavor], perature had a shelf life of 10 d, a 3-d in-
0.82 [firmness], and 0.86 [juiciness]). The crease as compared with overwrap package.
PLS model is able to predict sensory score The superchilled salmon had a considerable
within half-a-score accuracy for raw odor longer shelf life; both air and MA exposed
(RMSEP = 0.4), and a little lower accuracy for was acceptable after 21 d of storage, but the
cooked odor and flavor, RMSEP of 0.8 and MA superchilled salmon got higher scores
0.7, respectively. Sensory evaluation results throughout the storage period and at all
(Table 2 and 3) confirmed the effects ob- sampling days. The superchilled MA salm-
served for the microbial counts. The tem- on was evaluated of good quality after 24 d
perature effect was highly significant of storage, whereas the superchilled over-
(P < 0.005) for all sensory scores, and the wrap was not acceptable based on cooked
use of MAP significantly increased sensory odor scores.
scores (P < 0.005 for raw odor, P < 0.05 for
These shelf lives for the salmon exposed
cooked flavor, and P < 0.01 for the rest). Sen- to air is in good agreement with shelf lives
sory scores decreased as a function of stor- estimated using the square root model (Sto-
age time; however; the effect of using short rey 1985) and relative rate of spoilage com-
storage time was less important than using pared with spoilage at 0 °C:
1470 JOURNAL OF FOOD SCIENCE—Vol. 68, Nr. 4, 2003
JFS is available in searchable form at www.ift.org

Storage of superchilled MAP salmon . . .
evaluation of foods. Gaithersburg, Md.: Aspen Pub-
lishers, Inc. p 110-39.
Dalgaard P, Gram L, Huss HH. 1993. Spoilage and
shelf-life of cod fillets packed in vacuum or modi-
fied atmospheres. Int J Food Microbiol 19(4):283-
94.
Relative rate of spoilage =
(0.1 * storage temperature [°C] + 1)²
others 1997). At superchilled conditions, the
temperature in the products is several de-
grees lower than the minimum growth lim-
Using 14 d of shelf life for salmon at 0 °C, it of C. botulinum Type E, excluding and
this model gives a 6.8-d shelf life of air- health risks. Storage at chilled conditions
stored salmon at 4.4 °C, and a 21.3-d shelf (4 °C) makes growth possible, but time to
life at –1.9 °C, which is the same as our re- toxin formation in inoculated MA-packaged
sults. Since the superchilled MA salmon was salmon at this temperature is at least 50 d
not stored for longer than 24 d, the end of (Stier and others 1981; Garcia and Genigeor-
shelf life was not determined; but taking gis 1987; Reddy and others 1997). Because of
into account the low microbial levels and the 10-d shelf life observed for chilled MA
high sensory scores, at least on some days salmon in this study, there should be no risk
longer shelf life should be possible, but at a for toxin formation prior to spoilage at this
more mediocre quality. Again, using the temperature. However, toxin formation has
square-root model and shelf life of MA been found to precede spoilage in some cas-
salmon of 10 d at 4.4 °C, as observed in this es at abuse temperatures (Sivertsvik and
study, the shelf life at 0 °C would be about others 2002), so strict temperature control is
21 d, and 32 d at –1.9 °C. For MA-packaged vital both to ensure safety and shelf life. It is
salmon, shelf lives up to 34 d (frozen- also imperative to keep the superchilled
thawed product at 2 °C) (Emborg and oth- temperature as stable as possible since only
ers 2002) and even 55 d to 62 d have been re- 1 to 2 degrees decrease could lead to 30% to
ported (Clostridium botulinum challenge 50% freezing of the water in the product
test at 4 °C) (Reddy and others 1997). For (Huss 1995) and lead to severe tissue dam-
the air-exposed samples, an APC of log 6 age and excessive drip.
de la Hoz L, Lopez-Galvez DE, Fernandez M, Hierro
E, Ordonez JA. 2000. Use of carbon dioxide enriched
atmospheres in the refrigerated storage (2 °C) of
salmon (Salmo salar) steaks. Eur Food Res Technol
210(3):179-88.
Devlieghere F, Debevere J. 2000. Influence of dis-
solved carbon dioxide on the growth of spoilage
bacteria. Lebensm Wiss Technol 33(8):531-7.
Emborg J, Laursen BG, Rathjen T, Dalgaard P. 2002.
Microbial spoilage and formation of biogenic
amines in fresh and thawed modified atmosphere-
packed salmon (Salmo salar) at 2 °C. J Appl Micro-
biol 92(4):790-9.
Fletcher GC, Statham JA. 1988. Shelf-life of sterile
yellow-eyed mullet (aldrichetta forsteri) at 4 °C. J
Food Sci 53(4):1030-5.
Garcia G., Genigeorgis C. 1987. Quantitative evalu-
ation of Clostridium botulinum nonproteolytic
types B E and F growth risk in fresh salmon tissue
homogenates stored under modified atmospheres.
J Food Prot 50(5):390-7.
Graham AF, Mason DR, Maxwell FJ, Peck MW. 1997.
Effect of pH and NaCl on growth from spores of
non-proteolytic Clostridium botulinum at chill
temperature. Lett Appl Microbiol 24(2):95-100.
Gram L, Trolle G, Huss HH. 1987. Detection of specif-
ic spoilage bacteria from fish stored at low (0 °C)
and high (20 °C) temperatures. Int J Food Microbiol
4:65-72.
Haard NF. 1992. Technological Aspects of Extending
Prime Quality of Seafood: A Review. J Aqua Food
Prod Tech 1(3/4):9-27.
Hong LC, Leblanc EL, Hawrysh ZJ, Hardin RT. 1996.
Quality of Atlantic mackerel (Scomber scombrus
L.) fillets during modified atmosphere storage. J
Food Sci 61(3):646-51.
Huss HH. 1995. Quality and quality changes in fresh
fish. FAO Fisheries Technical Paper 348. Rome, It-
aly: Food and Agricultural Organization of the
United Nations. Available at <http://www.fao.org/
documents/>. Accessed 10/22/2002.
[ISO] International Organization for Standardiza-
tion. 1993. ISO 8586-1 Sensory analysis - general
guidance for the selection, training and monitor-
ing of assessors. 1st ed. Geneva, Switzerland: In-
ternational Organization for Standardization.
Killeffer DH 1930. Carbon dioxide preservation of
meat and fish. Ind Eng Chem 22(2):140-3.
Magnussen OM, Nordtvedt TS, Torstveit AK. 1998.
Use of partial freezing in the cold chain. In: IIR
Proceedings Series “Refrigeration Science and
Technology” 1998/6. Paris, France: Intl. Inst. of
Refrigeration. P 363-70
Monnin E, Indermuhle A, Dallenbach A, Fluckiger J,
Stauffer B, Stocker TF, Raynaud D, Barnola JM. 2001.
Atmospheric CO2 concentrations over the last gla-
cial termination. Science 291:112-4.
Pastoriza L, Sampedro G, Herrera JJ, Cabo ML. 1996.
Effect of carbon dioxide atmosphere on microbial
growth and quality of salmon slices. J Sci Food Ag-
ric 72:348-52.
Pearson SF. 1980. The application of engineering
science to fish preservation. Part 2. In: Connell JJ,
editor. Advances in fish science and technology.
Surrey, U.K.: Fishing News Books Ltd. P 184-91.
Penney N, Bell RG, Cummings TL. 1994. Extension of
the chilled storage life of smoked blue cod (Parap-
ercis colias) by carbon dioxide packaging. Int. J Food
Sci Technol 29:167-78.
CFU/g corresponds well with the sensory
Conclusions
spoilage time in this study (Figure 1a). Since
spoilage is affected both by the numbers of
specific spoilage organisms and their activ-
ity, the APC numbers cannot be directly re-
lated to time of spoilage without knowing the
species involved. When the concentration
of oxygen decreases and that of CO2 increas-
es during vacuum and MAP storage, this
gradually selects toward CO2-tolerant bac-
teria (Borch and others 1996). Dalgaard and
others (1993) reported P. phosphoreum as
the major spoilage bacterium of vacuum-
packaged and gas-packaged (in mixtures of
CO2 and N2) cod fillets stored at 0 °C. Fish
stored under anaerobic conditions or in
CO2-containing atmosphere often contain
lower number of psychrotrophic bacteria,
such as S. putrefaciens and Pseudomonas,
than aerobically stored fish. At these condi-
tions, the numbers of psychrotrophic bacte-
ria such as P. phosphoreum reach a level of
107 to 108 CFU/g at spoilage (Dalgaard and
others 1993). In contrast to Dalgaard (1995),
we did not observe high numbers of psy-
chrotrophic bacteria in superchilled MA-
packaged salmon. The increased shelf life of
superchilled salmon seems to be a suppres-
sion of SSOs in the product. Storage experi-
ments of more than 32 d should therefore be
carried out to confirm the estimates from
the square root model.
OTH MODIFIED ATMOSPHERE AND SUPER-
B
chilled conditions extended the shelf
life of salmon fillets, and when combined, a
synergistic effect was observed giving an
additional effect, maybe caused by in-
creased dissolvement of CO2 at the super-
chilled temperature. The combination of
MA and superchilling gave a product of high
quality in 24 d of storage with almost total
bacterial growth inhibition. The shelf-life
extension was at least 2.5 times that of tra-
ditional chilled MA salmon and at least 3.5
times that of chilled storage exposed to air.
No negative texture differences were ob-
served in the superchilled salmon, and the
increased drip loss was insignificant.
References
Ashie IN, Smith JP, Simpson BK. 1996. Spoilage and
shelf-life extension of fresh fish and shellfish. Crit
Rev Food Sci Nutr 36(1-2):87-121.
Borch E, Kant-Meuermans ML, BlixtY. 1996. Bacteri-
al spoilage of meat and cured meat products. Int J
Food Microbiol 33:103-20.
Carroll JJ, Slupsky JD, Mather AE. 1991. The solubil-
ity of carbon dioxide in water at low pressure. J
Phys Chem Ref Data 20:1201-9.
Chang KLB, Chang JJ, Shiau CY, Pan BS. 1998. Bio-
chemical microbiological and sensory changes of
sea bass (Lateolabrax japonicus) under partial
freezing and refrigerated storage. J Agric Food
Chem 46(2):682-6.
Rahman MS, Driscoll RH. 1994. Freezing points of
selected seafoods (invertebrates). Int J Food Sci
Technol 29:51-61.
Randell K, Hattula T, Skyttä E, Sivertsvik M, Bergslien
H. 1999. Quality of filleted salmon in various re-
tail packages. J Food Qual 22(5):483-97.
Reddy NR, Solomon HM, Yep H, Roman MG, Rhode-
hamel EJ. 1997. Shelf life and toxin development
by Clostridium botulinum during storage of mod-
ified-atmosphere-packaged fresh aquacultured
salmon fillets. J Food Prot 60(9):1055-63.
Shewan JM, Macintosh RG, Tucker CG, Ehrenberg
ASC. 1953. The development of a numerical scor-
ing system for the sensory assessment of the spoil-
age of wet white fish stored in ice. J Sci Food Agric
4:283-98.
Chang HD, Tao LC 1981. Correlations of enthalpies
of food systems. J Food Sci 46:1493-7.
Coyne FP. 1932. The effect of carbon dioxide on bac-
terial growth with special reference to the preser-
vation of fish. Part I. J Soc Chem Ind 51:19T-121T.
Coyne FP. 1933. The effect of carbon dioxide on bac-
terial growth with special reference to the pres-
ervation of fish. Part II. J Soc Chem Ind 52:19T-
24T.
Dalgaard P. 1995. Qualitative and quantitative char-
acterization of spoilage bacteria from packed fish.
Int J Food Microbiol 26(3):319-33.
The growth of psychrotrophic non-pro-
teolytic Clostridium botulinum Type E has
been a major concern for fish products pack-
aged in oxygen-free environments, mainly
because of their ability to grow and pro-
duce toxin under anaerobic conditions at
temperature as low as 3.0 °C (Graham and
Dalgaard P. 2000. Fresh and lightly preserved sea-
food. In: Man CMD, Jones AA, editors. Shelf-life
JFS is available in searchable form at www.ift.org
Vol. 68, Nr. 4, 2003—JOURNAL OF FOOD SCIENCE 1471

Storage of superchilled MAP salmon . . .
Sikorski ZE, Sun P. 1994. Preservation of seafood qual- Stammen K, Gerdes D, Caporaso F. 1990. Modified
Inst. of Refrigeration. p 239-43.
ity. In: Shahidi F, Botta JR, editors. Seafoods: chem-
atmosphere packaging of seafood. Crit Rev Food Sveinsdottir K, Martinsdottir E, Hyldig G, Jorgensen
istry processing technology and quality. Glasgow,
U.K.: Blackie Academic and Professionals. p 168- Stansby ME, Griffiths FP. 1935. Carbon dioxide in
95
Sci Nutr 29(5):301-31.
B, Kristbergsson K. 2002. Application of Quality
Index Method (QIM) scheme in shelf-life study of
farmed Atlantic salmon (Salmo salar). J Food Sci
67:1570-9.
handling fresh fish—Haddock. Ind Eng Chem
27(12):1452-8.
Sivertsvik M, Jeksrud WK, Rosnes JT. 2002. A review
of modified atmosphere packaging of fish and fish- Stier RF, Bell L, Ito KA, Shafer BD, Brown LA, Seeger MS 20020643 Submitted 11/11/02, Revised 12/2/02,
ery products - significance of microbial growth ac-
tivities and safety. Int J Food Sci Technol 37(2):107-
27.
Sivertsvik M, Nordtvedt TS, Aune EJ, Rosnes JT. 1999.
Storage quality of superchilled and modified at- Storey RM. 1985. Time temperature function inte-
mosphere packaged whole salmon. In: Proc. from
20th Intl. Congress of Refrigeration; Sydney, Aus-
tralia; 19-24 Sept. Paris, France: International In-
stitute of Refrigeration. Vol IV, p 2488-95.
ML, Allen BH, Porcuna MN, Lerke PA. 1981. Effect
of modified atmosphere storage on C. Botulinum
Toxigenesis and the spoilage microflora of salmon
fillets. J Food Sci 46:1639-42.
Accepted 1/13/03, Received 1/17/03
Authors Sivertsvik, Rosnes, and Kleiberg are with
NORCONSERV, Inst. of Fish Processing and Pres-
ervation Technology, Alex. Kiellandsgt. 2, P.O. Box
327, 4002 Stavanger, Norway. Direct inquiries to
author Sivertsvik (E-mail: morten.sivertsvik@
norconserv.no).
gration its realisation and application to chilled
fish. In: Proceedings from IIR Conference of Stor-
age lives of chilled and frozen fish and fish prod-
ucts; Aberdeen; 1-3 Oct. 1985. Paris, France: Intl.
1472 JOURNAL OF FOOD SCIENCE—Vol. 68, Nr. 4, 2003
JFS is available in searchable form at www.ift.org