Full text of DOI:10.1111/j.1365-2621.2003.tb07030.x

JFS: Sensory and Nutritive Qualities of Food
Effects of Freezing on Quality of Cold-
smoked Salmon Based on the Measurements
of Physiochemical Characteristics
A.M.B. RØRÅ AND O. EINEN
ABSTRACT: How the quality of cold-smoked Atlantic salmon (Salmo salar) is affected by freezing the raw fish as
whole fish, as gutted fish, and as fillets before smoking, and by freezing the finished product after smoking was
studied. Freezing before smoking resulted in increased product yield and water content, but softer texture and
increased K-value. Freezing reduced the content of astaxanthin but increased the lightness and the color inten-
sity of the flesh. Gaping increased when the fish was frozen as fillets before smoking. Freezing only after smok-
ing led to fewer changes in quality than freezing before smoking, whereas refreezing the finished products had
little additional effect on quality.
Keywords: cold-smoked salmon, freezing, texture, color, yield
Introduction
mentation, but usually a deep red/orange
OLD-SMOKED SALMON IS A LIGHTLY PRE- color is preferred (Moe 1990; Gormley 1992).
short-term freezing on salmon quality are
not known. Considerable attention has been
given to chemical and microbiological qual-
ity indicators to predict the shelf life of frozen
fish, also in combination with smoking (Tru-
elstrup Hansen and others 1996). However,
the ways in which freezing affects other im-
portant quality characteristics such as the
yield, color, and texture of cold-smoked salm-
on have been given little attention.
C
served fish product of considerable A relationship between the concentration of
economic importance throughout the world; carotenoids and red color intensity in raw,
40% to 50% of Norwegian Atlantic salmon boiled, and smoked Atlantic salmon is given
(Salmo salar) reaches the final consumer as by Skrede and Storebakken (1986). Besides
a cold-smoked product (Borch and Aaker initial concentration of carotenoids in the
1997). Fish are smoked in industrialized muscle, the color of the smoked product
countries to enhance the flavor and texture, also depends on textural properties, chem-
and this form of processing only to a certain ical composition, and the liquid-binding
degree protects against microbiological, en- ability of the muscle. Soft texture, fillet gap-
The objective of this study was to evalu-
ate how freezing the fish before and/or after
smoking affected central quality character-
istics of the finished cold-smoked salmon.
zymatic, and chemical degradation.
ing, and liquid loss are serious problems for
A successful smoked product depends the smoking industry. Though some effects
on high yield throughout the process (Rørå of pH and muscle composition can be
and others 1998), on desirable and good found, little is known about what initiates
quality, and on an acceptable shelf life these problems in smoked salmon.
Materials and Methods
(Dobbs and others 1992). Product yield is
Most sizes of farmed Atlantic salmon are
Fish material
important for the economic output, and now available throughout the year, but
both starvation (Lie and Huse 1992; Einen smokehouses may choose to freeze their fish
and others 1998) and feed ration level (Ein- either before or after processing for reasons
en and others 1999) before harvest have of tradition, price, or logistics. Furthermore,
been associated with reduced fillet yield of the product might be frozen once or twice in
Atlantic salmon. Furthermore, Rørå and oth- the retail store or the home freezer before
ers (1998) showed that smoking yield in- consumption. Freezing, frozen storage, and
creased with increasing fat content of the thawing are all processes that can cause
raw material. Cold-smoked salmon is pro- changes in the properties of fish flesh; and
cessed in many ways, and the techniques apart from the primary quality of the flesh,
used for salting, drying, and smoking affect the characteristics of the freezing process,
Gutted Atlantic salmon were obtained
from a feed ration experiment of duration 110
d at AKVAFORSK’s Research Station (Ekkilsøy,
Norway). Einen and others (1999) have de-
scribed the details of this experiment. The
experimental fish were kept in 6 net pens in
seawater and fed 3 different ration levels,
namely 0%, 75%, and 100% of voluntary feed
intake, with 2 net pens per ration level. The
fish were fed a commercial diet (ROYAL 2 +
manufactured by T. Skretting) containing 955
g/kg dry matter, 456 g/kg protein, 323 g/kg fat,
3 g/kg crude fiber, 103 g/kg carbohydrate, 51
mg/kg astaxanthin, and 25.0 MJ/kg gross en-
ergy. The ration levels (RLs) are denoted by
RL = 0, RL = 75, and RL = 100, respectively. Fif-
teen fish were sampled from each of the 6 net
pens for use in the experiment with different
freeze/thaw treatments.
both yield and general quality.
including storage and thawing, affect the
The color of the flesh is one of the most quality of the product at consumption. Ear-
important sensory attributes of farmed At- lier studies have shown the effects of frozen
lantic salmon (Simpson 1978; Christiansen storage on sensory attributes, color, and
and others 1995). Fading of color, or “de-pig- chemical composition of salmon (Refsgaard
mentation,” consequently has a severe neg- and others 1998; Sheehan and others 1998).
ative effect on the marketability of the salm- The changes have been explained as arising
on. Smoked salmon is sold on markets that from biochemical changes that occur during
have different preferences for flesh pig- long storage periods, and the effects of
© 2003 Institute of Food Technologists
Further reproduction prohibited without permission
Vol. 68, Nr. 6, 2003—JOURNAL OF FOOD SCIENCE 2123

Freezing of cold-smoked salmon . . .
Freeze treatments
filleted-frozen fish were smoked and packed pan) with light source D. Four measure-
in exactly the same manner, using the same ments were carried out directly on each fil-
The salmon were treated in 3 groups. Fish
in the first group were filleted and smoked
fresh; fish in the second group were filleted,
frozen as plastic-packed fillets, thawed, and
then smoked; whereas fish in the third group
were frozen as whole, gutted fish packed in
plastic, thawed, filleted, and then smoked.
The smoked fillets were cold-stored for 1 wk
before analysis (left-side fillets), or frozen,
thawed, and then cold-stored for 1 wk before
analysis (right-side fillets). The experiment
thus covered 36 experimental groups with 5
smoked fillets in each group: 6 different
freeze/thaw treatments × 3 feed ration
level × 2 net pens per ration level.
Whole fish or fillets to be frozen were
packed individually in plastic bags, frozen
by cold air in a climate chamber that could
be regulated from –50 °C to + 50 °C 0.5 °C,
and stored at –30 °C for 6 d before thawing
and subsequent smoking. Freezing and
thawing was recorded by temperature sen-
sors in the center of 2 fish per treatment. A
core temperature of –30 °C was reached in
both whole and fillet frozen fish after 30 h,
but the fillets frozen fish reached a core
temperature of –10 °C after 8 h, whereas the
whole-frozen fish reached the same tem-
perature after 16 h. When thawing, the fil-
lets passed –25 °C after 6 h, whereas the
whole fillets used 26 h to reach the same
temperature; both groups were completely
thawed after 36 h. After smoking, the left-
side fillets were frozen in vacuum packs in a
freezing room at a constant temperature of
–20 °C and stored frozen for 7 to 14 d before
being thawed (at 4 °C) and analyzed.
experienced trimmer for all the fish.
let. L* describes the lightness of the sample,
a* the intensity of red color (a* > 0), and b*
the intensity of yellow color (b* > 0). The color
Yield and liquid-holding capacity
Both round weight and gutted weight intensity (c) was calculated as c = (a*2 + b*2)1/
were recorded at slaughter, and fillet weight 2. Furthermore, the hue angle (hab) was cal-
was recorded after brine injection and after culated as hab = tan–1 (b*/a*), where hab = 0°
smoking. After smoking, all fish were for a red hue and hab = 90° for a yellow hue.
packed with a 10- × 10-cm absorbent pad Astaxanthin from pooled samples was de-
having a dry weight of 4.2 g (NKL, Oslo, Nor- termined by high-performance liquid chro-
way) between the skin side of the fish and matography using Hewlett Packard Series
the vacuum pack. The absorbent pad ab- 1050 Instrument (Palo Alto, Calif., U.S.A.) as
sorbed all liquid leakage from the smoked described by Bjerkeng and others (1997).
salmon during storage and during freezing. All samples were analyzed isocratically on a
The vacuum packs were opened and the Spherisorb S5CN-4800 nitrile column (Hich-
wet absorbent pad was weighed. The absor- rom Ltd., Theale, Berkshire, UK; 250-mm
bent pad was then dried at 60 °C for 24 h to length; 4.6-mm internal dia; 5-␮m particle
distinguish between water and other mate- size) using 20% acetone in hexane as mobil
rial. Loss in the vacuum pack was multiplied phase.
by 2, divided by the round weight, and ex-
Texture analysis
pressed as percentage loss.
Pooled samples of muscle from 5 smoked
Fillet gaping was evaluated according to
fillets, taken directly under the dorsal fin, a scale from 0 to 5 of gaping severity
were collected from each of the 36 experi- (Andersen and others 1994) on which score
mental groups, finely chopped, and ana- 0 describes fish with no gaping and score 5
lyzed for liquid-holding capacity. The sam- describes fish with extreme gaping with the
ples (15 g) were weighed and placed in a tube fillet falling apart. Two judges with several
with a weighted filter paper (V1) (Schleicher years experience in gaping evaluation per-
& Schuell GmbH, Dassel, Germany). The formed the scoring.
tubes were centrifuged at 500 × g for 10 min
Fillet texture was measured by compres-
at 10 °C, as described by Hermansson sion using a Texture Analyser, TA-XT2 (Sta-
(1986), and the wet paper was weighted (V2) ble Micro Systems, Surrey, England) with a
before drying at 50 °C to constant weight 12.5-mm cylindrical probe. Texture profile
(V3). The percentage liquid loss was calculat- analyses were performed using a test speed
ed on a wet weight basis as 100 × (V1
–
of 2 mm/s and 60% penetration on the fil-
V2) × S–1, where S = weight of muscle sample, lets and 90% penetration on 2-cm-thick cut-
water loss as 100 × (V2 – V3) × S–1, and fat loss lets. The force needed to create a break in
as 100 × (V3 – V1) × S–1, respectively.
the muscle was denoted as the breakpoint
value. Four measurements were taken from
each fillet and from each cutlet, and the
Filleting and smoking
Chemical analysis
Two of the groups were sent in a refriger-
ated truck directly to the smokehouse,
whereas the third group was sent to the lab-
oratory of AKVAFORSK, Ås, Norway. The iced,
gutted fish were there frozen on the third
day after slaughter. The 2 fish groups that
were sent directly to the smokehouse were
stored on ice at 0 to 3 °C for 5 d after slaugh-
tering before processing. One of these 2
groups was filleted by machine and trimmed
according to standard B (NSA 1996), which
includes skin but where visible dorsal and
ventral fat depots, bones, and peritoneum
are removed before the fillets were frozen.
The other group was machine-filleted, hand-
trimmed, and salted by automatic injection
brining using a 25% wt/wt brine and an injec-
tion pressure of 1 bar. The fillets were then
cold-smoked for 2.5 h at 23 °C and at a rela-
tive humidity of 69% in a smoke oven sup-
plied with smoke generated from beech
wood, and then chilled to 4 °C and vacuum-
packed. After thawing, the whole-frozen and
Dry matter content was determined on an mean was used in further calculation.
individual basis following heating at 105 °C
Statistical analysis
The effects of ration level, with the net pen
for 24 h, whereas the rest of the chemical
analyses were performed on pooled homoge-
nates of the 5 fillets in each group. Protein as the unit of observation, and the effects of
content was analyzed as Kjeldahl- N*6.25 freezing before and after smoking were ana-
(Kjeltec Autoanalyser, Tecator, Sweden) and lyzed by ANOVA. Means were ranked by the
fat by ethyl-acetate extraction (NSA 1994).
Student-Newman-Keuls test with the signif-
The K-value was determined using the icance level set at 5%. Principal component
test strips Fresh Tester FTP II (Transia, analysis (PCA) was performed including all
Netherlands), which is a simple enzymatic the variables measured on an individual
colorimetric method. This method gives the basis. The variables were weighed by 1/SD
K-value as 100% × (Inosine + Hypoxan- when used in the PCA analysis. Numbers in
thine)/(IMP + Inosine + Hypoxanthine) as the text are given as mean SD.
defined by Karube and others (1984). Salt
Results
was measured with a Chloride Analyser 926
(Corning, Halstead, U.K.).
CA ANALYSIS OF ALL DATA SHOWED THAT THE
Pmain principal component (PC1), ex-
plaining 31% of the variation, grouped the fil-
lets according to the degree of freezing (Fig-
ure 1a). The fresh smoked fillets were totally
separated from the other groups, whereas
Color analysis
Instrumental color analyses (CIE 1976
L*a*b*) were performed using a Minolta
Chroma Meter CR-200 (Minolta, Osaka, Ja-
2124 JOURNAL OF FOOD SCIENCE—Vol. 68, Nr. 6, 2003
JFS is available in searchable form at www.ift.org

Freezing of cold-smoked salmon . . .
Table 1—Yield and liquid-holding capacity of Atlantic salmon fillets frozen before and/or after smoking
Treatment before smoking
Fresh
Fillet frozen
Fresh Frozen
4380
Whole frozen
ANOVAa
Treatment after smoking Fresh
Frozen
Fresh
Frozen Root MSE
Pbefore
Pafter
PInteraction
Gutted weight (g)
4314
63.3
—
—
—
—
4474
65.1
58.1
0.2
—
—
—
0.2
4.7
876.9
2.83
3.01
0.09
1.81
1.23
0.495
0.002
0.001
0.921
0.001
0.001
—
—
—
0.263
0.001
0.036
—
—
—
0.562
0.164
0.059
Yield after brining (%)
Yield after smoking (%)
Loss in vacuum pack (%)
—
—
0.2
5.3
36.1
64.4
54.4
0.2
57.5
0.2
2.8
0.2
3.4
Liquid loss from salmon muscle (%) 4.4
(%) of which was water
2.9
36.8
34.8
34.1
34.5
34.7
a
Results from ANOVA where root MSE is the square root of the mean square error, and P
of freezing before and after smoking and the interaction, respectively.
, P
, and P
denote the significance levels for effects
Interaction
before
after
the fresh refrozen fillets were grouped closer group (3.1 2.0%). The amount of water lost those only cold-stored (3.4% 1.5% liquid
to the rest of the material, but still clearly sep- was approximately 2% higher in the fresh loss, 35.5% of which was water).
arated. The second principal component smoked groups than it was in the other
Chemical composition
(PC2), explaining 24% of the variation, sepa- groups (Table 1). Fillets frozen after smoking
rated the fillets according to feed ration lev- lost more liquid in the form of fat (4.4% 1.6%
The mean protein content of the fish ma-
el. The starved fish (RL = 0) were clearly sep- liquid loss, 35% of which was water) than terial was 19.4% (range, 18.2% to 20.8%), and
arated from the other ration level groups
(RL = 75 and RL = 100). The third principal
component, explaining 15% of the variation,
separated the fillets according to refreezing
after smoking (Figure 1b). The plots indicate
the relative importance of the pre-slaughter
treatment of the fish material and freezing
before and/or after smoking on the ultimate
quality of the smoked fillets.
Yield and liquid-holding capacity
Different feeding intensities before
slaughter gave a range in round weight from
2660 g to 6240 g. However, the fish were se-
lected so there were no significant differenc-
es in average weights between the different
freezing treatments (Table 1). The mean
yield after brine injection was 64.2% 2.9%
and after smoking was 56.6% 3.3%. The
yield after brine injection was significantly
higher when using whole-frozen fish than it
was when using fresh or frozen fillets before
smoking (P < 0.05). The yield after smoking
was significantly different between all 3
freezing treatments used before smoking.
There were no significant differences due to
freezing before or after smoking in the
amounts of liquid absorbed in the vacuum
packs (Table 1). However, there was a signif-
icant interaction between feed ration level
and treatment before smoking. The liquid
loss was higher at RL = 0 than it was at
RL = 75% and at 100% when the fish were fro-
zen as fillets before smoking, but the liquid
loss was lower at RL = 0 than it was at
RL = 75% when the fish were frozen whole or
kept fresh before smoking (Figure 1a). The
mean weight of liquid loss from salmon mus-
cle after centrifugation was 3.9% 2.0%, and
35.2% of this was water (Table 1). The liquid
loss was higher in the fresh smoked group
(4.8% 2.0%) than it was in the whole-frozen
group (3.8% 1.7%) and in the fillet-frozen
Figure 1—Principal component analysis (PCA) biplots of (a) PC1 (31%) and PC2
(24%), and (b) PC1 and PC3 (15%) of 36 groups of smoked Atlantic salmon.
The first denotation gives the feed ration level (0%, 75%, 100%); the second
gives the treatment before smoking (0 = fresh, 1 = fillet frozen, 2 = whole fro-
zen); and the third gives the treatment after smoking (0 = fresh, 1 = frozen).
Packloss = loss in vacuum packs.
JFS is available in searchable form at www.ift.org
Vol. 68, Nr. 6, 2003—JOURNAL OF FOOD SCIENCE 2125

Freezing of cold-smoked salmon . . .
Table 2—Composition of fillets frozen before and/or after smoking (n = 30)
Treatment before smoking
Fresh
Fillet frozen
Fresh Frozen
66.5 75.0
Whole frozen
ANOVAa
Pafter
Treatment after smoking Fresh
Frozen
Fresh
Frozen Root MSE
Pbefore
PInteraction
K-value (%)
Salt content (%)
Dry matter content (%)
59.3
3.1
36.8
72.5
3.1
36.1
67.2
2.7
34.5
75.0
2.9
34.7
6.04
0.24
1.23
0.001
0.001
0.001
0.001
0.001
0.036
0.034
0.115
0.059
2.6
2.9
34.1
34.8
a
Results from ANOVA where root MSE is the square root of the mean square error, and P
of freezing before and after smoking and the interaction, respectively.
, P
, and P
denote the significance levels for effects
Interaction
before
after
the mean fat content was 12.7% (range, (L*), redness (a*), and yellowness (b*) in- of the same group of Atlantic salmon docu-
10.8% to 14.6%). creased significantly due to freezing before mented several changes in the raw fillet
Fish smoked fresh had a lower K-value and after smoking (Table 3). The color inten- characteristics with reduced feed ration (low-
(65.9%) than whole-frozen fish (71.1%) and sity (c) also increased significantly after the er contents of fat, water-soluble protein, and
fillet-frozen fish (70.9%) (Table 2). There freezing treatments, whereas the hue (hab
)
astaxanthin; higher content of collagen and
were significant interactions both between did not change significantly. Freezing before higher pH). The study also showed that re-
feed ration level and freezing before smok- smoking had a significant effect on L*: fish duced feed ration affected the sensory qual-
ing, and between freezing before and after frozen whole (50.8) and fish frozen as fillets ity of smoked fillets (increased hardness and
smoking. Fish at RL = 100 had a lower K-val- (49.0) had higher L* values than fillets that redness, and reduced fattiness and juici-
ue than fish at RL = 75 and RL = 0 when were smoked fresh (45.6). Freezing before ness). The results presented here show that
smoked fresh, but this was not the case smoking, either as whole fish or as fillets, freezing before and/or after smoking also
when the fish were frozen (as fillet or whole) also increased significantly the a* and b* affects the yield and quality of smoked
before smoking (Figure 1a). Freezing after values (Table 3). Furthermore, the freezing salmon. The interactions between fish raw
smoking increased the K-value from 64.3% of fillets after smoking increased the aver- material for smoking and freezing (before
to 74.2% (Table 2).
The smoked fillets in this study were brine ue from 24.0 to 28.1 in fillets that were ei- of husbandry and feeding practices.
injected, and the analyzed salt content var- ther stored fresh or frozen after smoking, The higher yields and higher water con-
age a* value from 8.7 to 10.5 and the b* val- and/or after smoking) stress the importance
ied from 2.1% to 3.3% with an average of respectively. The relative increases in a* and tent (that is, lower dry matter) of the frozen
2.9%. There were significant differences in b* values were much larger than the increase fillets indicate increased uptake of brine
average salt content for freeze treatments in L*, indicating a more reddish and a more into the muscle due to freezing. Freezing
(Table 2). The salt content depended on the yellowish color impression of the fillets fro- and drying cause major changes in muscle
treatment before smoking, and fish smoked zen before or after smoking.
fresh had a higher salt content (3.1%) than
pH and the ionic strength, which profound-
ly influence the swelling of myofibrils and
thus the ability of the muscle to bind water
Gaping and texture evaluation
whole-frozen fish (2.8%) and fillet-frozen
fish (2.8%). Fish refrozen after smoking had
Gaping scores were generally low with a (Wilding and others 1986; Fennema 1990).
a higher salt content (3.0%) than those not mean of 0.9 (on a scale from 0 to 5) for all the The nature of the changes depends on the
frozen (2.8%). Fish smoked fresh also had a smoked fillets. Fish frozen as fillets before freezing method used. Although the prefro-
higher dry matter content (36.5%) than smoking had a higher gaping score (1.3) than zen groups contained more water, there was
whole-frozen fish (34.6%) and fillet-frozen fish smoked fresh (0.7) and than fish whole- no indication of a looser bondage of this
fish (34.4%). Fillets frozen after smoking had frozen before smoking (0.6) (Table 4). Freez- water when the fillets were stored in vacu-
a significantly lower dry matter content than ing the fillets after smoking did not signifi- um packs. It is generally accepted that my-
fish that were not refrozen.
cantly affect the gaping scores. There was an osin, in particular, undergoes aggregation
interaction between feed ration levels and reactions during frozen storage and that
freeze treatment before smoking: the higher these reactions lead to muscle toughening
Color characteristics
Freezing before smoking had a significant gaping scores were less pronounced at and drip loss during thawing (Mackie 1993).
effect on all color characteristics. Astaxanthin RL = 75 than they were at RL = 0 and at The frozen storage periods were short in the
content was reduced from fresh smoked to RL = 100 for fish frozen as fillets before experiments reported here, probably negli-
fillet-frozen and whole-frozen fish (Table 3). smoking compared with those that were fro- gible, whereas the freezing and thawing
There was no significant difference between zen whole or kept fresh until smoking. were relatively slow and the flesh was ex-
fish frozen whole and fish frozen as fillets The breakpoint measured on fillets (on posed to the critical temperatures of –1 °C to
before smoking. There was also an interac- average, 10.7N) showed significant effects –6 °C for a relatively long period. Generally,
tion effect with pre-slaughter treatment of both freezing before and after smoking slow thawing (at 5 °C) will give higher liquid
showing that starved fish, having the high- (Table 4), whereas the breakpoint measured loss than fast thawing (at 25 °C in water) (Bi-
est astaxanthin content when smoked as on cutlets (on average, 5.2N) was not signif- linski and others 1977; Nilsson and Ek-
fresh, ended with astaxanthin content in icantly affected by freezing after smoking. strand 1995). Only extracellular water will be
line with the 2 other ration levels (RL = 75 and However, for the breakpoint measured on freed by the low-pressure method used in
100) when the fish was frozen before smok- cutlets, there were significant interaction ef- the present study (Offer and Trinick 1983),
ing. There was on average significant lower fects between freezing before and after and loss of cellular water must involve move-
astaxanthin in fillets that were frozen after smoking (Table 4).
smoking (5.6 mg/kg) compared with fillets
ment of water from intracellular to extracel-
lular locations. The greater loss of liquid
during the centrifugation test from fillets
Discussion
stored fresh after smoking (6.1 mg/kg).
The CIE (1976) color values for lightness
A previous study (Einen and others 1999) frozen after smoking than from those ana-
2126 JOURNAL OF FOOD SCIENCE—Vol. 68, Nr. 6, 2003
JFS is available in searchable form at www.ift.org

Freezing of cold-smoked salmon . . .
Table 3—Color characteristics of fillets frozen before and/or after smoking (n = 30)
Treatment before smoking
Treatment after smoking Fresh
Astaxanthin (mg/kg) 7.0
Fresh
Fillet frozen
Fresh Frozen
5.7 5.3
Whole frozen
ANOVAa
Pafter
Frozen
Fresh
Frozen Root MSE
Pbefore
PInteraction
5.9
5.8
5.4
0.90
0.001
0.001
0.061
Instrumental color measurements (CIE 1976)
Lightness (L*)
Redness (a*)
44.3
7.6
21.4
22.7
70.5
46.8
10.1
26.9
28.7
69.5
47.9
9.4
25.6
27.3
69.9
49.9
11.0
29.3
31.3
69.5
50.1
9.2
25.1
26.7
69.9
51.6
10.5
28.1
30.0
69.5
1.84
1.43
1.80
1.93
2.55
0.001
0.001
0.001
0.001
0.899
0.001
0.001
0.001
0.001
0.134
0.336
0.080
0.006
0.004
0.824
Yellowness (b*)
Color intensity (c)
Hue (hab
)
a
Results from ANOVA where root MSE is the square root of the mean square error, and P
of freezing before and after smoking and the interaction, respectively.
, P
, and P
denote the significance levels for effects
Interaction
before
after
Table 4—Gaping score and instrumental texture measurements of fillets frozen before and/or after smoking (n = 30)
Treatment before smoking
Fresh
Fillet frozen
Fresh Frozen
1.6 1.1
Whole frozen
ANOVAa
Treatment after smoking Fresh
Frozen
Fresh
Frozen Root MSE
Pbefore
Pafter
PInteraction
Gaping score (0 to 5)
0.7
0.7
0.7
0.5
0.97
0.001
0.166
0.406
Breakpoint (N)
Fillet
Cutlet
11.9
6.5
11.9
6.1
9.3
5.0
10.4
4.8
9.4
3.9
11.2
4.6
1.24
0.44
0.001
0.001
0.007
0.787
0.069
0.006
a
Results from analyses of variance (ANOVA) where root MSE is the square root of the mean square error, and P
significance levels for effects of freezing before and after smoking and the interaction, respectively.
, P
, and P
denote the
Interaction
before
after
lyzed when fresh is probably due to gener- ments were stable during frozen storage.
al muscle degradation and not to freeze- The pigments of salmonids are sensitive to
al, indicating that astaxanthin is broken
down.
denaturation of proteins.
light, heat, and oxygen (Burton and Ingold
The level of astaxanthin was lower in fil-
lets frozen at any stage of the process, but
the instrumental color readings showed that
the fillets had a higher color intensity. This
agrees with results presented by No and
Storebakken (1991), who reported that fro-
zen storage resulted in extensive color
changes of rainbow trout fillets, causing
them to become lighter, more red, and more
yellow. We found that harder texture was
correlated with a darker color (a lower value
of L*) (r = –0.56, P = 0.001). This supports
the theory of Francis and Clydesdale (1975),
which states that visual color differences
are due to changes in light absorption and
light scattering, which can be caused by
freeze-denaturation.
Adenosine 5-triphosphate (ATP) degen- 1984; Burton 1989; Krinsky 1989), and they
eration occurs rapidly at subzero tempera- can function as antioxidants. Jensen and
tures down to –6 °C and will accelerate at others (1988) claimed that astaxanthin pro-
temperatures just below 0 °C (Cappeln and tects against the very early stages of lipid
others 1999). This finding agrees with our oxidation, whereas ␣-tocopherol is more
observations concerning the effects of freez- important as an antioxidant at more ad-
ing on the K-value. The ATP degeneration is vanced stages of lipid oxidation. Thus, raw
due to both an increase in enzyme activity material characteristics, individual antioxi-
and to increased concentrations of soluble dant status, freezing and thawing time, tem-
reactants or catalysts in the unfrozen water perature, and storage and package condi-
phase (Dyer 1968; Behnke and others tions will all affect the pigment stability
1973). The K-value does not account for bac- during freezing. The relative loss of astax-
terial growth and the risk of food-borne anthin was higher in the RL = 0 group, indi-
pathogens, which are the limiting factors for cating that muscle structure also might have
the shelf life of cold-smoked salmon (Rørvik an effect on pigment stability.
and Yndestad 1991; Jemmi 1993), and thus
Sheehan and others (1998) reported that
the impact of freezing before and after carotenoid concentration was not signifi-
smoking on the shelf life and freshness re- cantly reduced in Atlantic salmon muscle
quires further documentation of microbio- due to smoking. They found only small, in-
Fillet gaping is the term used to describe
the phenomenon in which the connective
tissue of the fish fails to hold the muscle
blocks together. Slits appear across the sur-
face of the fillets, making mechanical skin-
ning and slicing difficult. Gaping scores
were low in the samples we studied, but they
were significantly higher in the smoked fil-
lets that had been frozen as fillets before
smoking. However, the effect of fillet freez-
ing was lower in the RL = 75 group, indicating
a greater connective tissue strength in this
group. Gaping has often been associated
with low final pH (Lavèty and others 1988;
Andersen and others 1994) and with the
strength of the connective tissue (Lavèty
and others 1988; Mackie 1993). Freezing and
thawing cause changes in the salt concentra-
logical parameters.
significant reductions in fish fed on astax-
Astaxanthin levels were reduced by freez- anthin (from 9.4 to 8.5 mg/kg) and in fish
ing both before and after smoking (both on fed on canthaxanthin (from 10.6 to 9.3 mg/
a wet weight basis and dry matter basis). kg). In contrast, Choubert and others
Similarly, carotenoids are degraded in fro- (1992) found that the smoking process in-
zen stored rainbow trout muscle (Chen and duced an increase in carotenoid concentra-
others 1984; Pozo and others 1988; Ander- tion in rainbow trout muscle. The fillets in
son and others 1990; Ingemansson and oth- the rainbow trout study lost 12% of the raw
ers 1993), and in salmon (Lusk and others material muscle water weight, which might
1964; Sheehan and others 1998). However, explain the increase in carotenoid concen-
Christiansen and others (1995) found only tration during smoking. We have shown that
minor decreases in pigment concentration the astaxanthin content was reduced by
of salmon frozen stored for 180 d at –18 °C, freezing on a dry matter basis, and that the
whereas No and Storebakken (1991) and reduction cannot be fully explained by dif-
Scott and others (1994) found that the pig- ferences in water content of the raw materi-
JFS is available in searchable form at www.ift.org
Vol. 68, Nr. 6, 2003—JOURNAL OF FOOD SCIENCE 2127

Freezing of cold-smoked salmon . . .
Bjerkeng B, Reftie S, Fjalestad KT, Storebakken T,
Rødbotten M, Roem AJ. 1997. Quality parameters
of the flesh of Atlantic salmon (Salmo salar) as af- Lusk G, Kare M, Goldblith SA. 1964. Astacene pig-
fected by fat content and full-fat soybean meal as
partial substitute of fish meal in the diet. Aquacul-
ture 157:297–309.
composition of Atlantic salmon (Salmo salar).
Fiskeridirek Skrifter Ernær 5:11–6.
tion and in the pH of the muscle (Sigurgisla-
dottir and others 2000). However, we sug-
gest that the increased gaping that we ob-
served in the fillet-frozen group is related to
the mechanical stress that the deboned fil-
lets are subjected to during freezing. The
amount of damage occurring during freez-
ing depends on the amount of ice crystal
formation, and rapid freezing is generally
recommended (Nicholson 1973). The
amount of gaping arising during freezing
depended on the amount of gaping ap-
pearing in the raw material before freezing.
Smoked fillet seemed more resistant to fur-
ther gaping during freezing, and this may
have been due to the binding effects of salt
and to the surface drying during smoking.
Freezing before smoking resulted in a
softer texture, which is in accordance with
data from Dunajski (1979). These groups
also had a lower dry matter content, but the
correlation between texture and dry matter
content was low, indicating that the texture
softening was related to muscle degenera-
tion. Freezing and thawing affects the mem-
brane structure of the muscle cells (Fennema
1990), and refreezing resulted in a significant
difference in texture when measuring on the
surface of the fillets but not on the cutlets.
ment loss occurring in freeze-dried shrimp and
salmon during storage. Food Technol 18:157–8.
Mackie IM. 1993. The effects of freezing on flesh pro-
teins. Food Rev Int 9:575–610.
Borch OD, Aaker H. 1997. Markedsdifferensiert
produktkvalitet i lakseoppdrett (Market-differen- Moe NH 1990. Key factors in marketing farmed
tiated product quality in salmon farming). NF rep. salmon. Proc Nutr Soc New Zealand 15:16–22.
6/97. Nordland Reseach Inst., Bodø (In Norwegian). Nicholson FJ. 1973. The freezing time of fish. Torry
110 p.
advisory note 62. Torry Research Station, Aber-
deen. Edinburgh: HMSO Press. 8 p.
Nilsson K, Ekstrand B. 1995. Frozen storage and thaw-
ing methods affect biochemical and sensory at-
tributes of rainbow trout. J Food Sci 60:627–30, 635.
Burton GW. 1989. Antioxidant action of carotenoids.
J Nutr 119:109–11.
Burton GW, Ingold KU. 1984. ␤-carotene: an unusual
type of lipid antioxidant. Science 224:569–73.
Cappeln G, Nielsen J, Jessen F. 1999. Synthesis and No HK, Storebakken T. 1991. Color stability of rain-
degradation of adenosine triphosphate in cod (Ga-
dus morhua) at sub-zero temperatures. J Sci Food
Agric 79:1099–104.
Chen HM, Meyers SP, Hardy RW, Biede SL. 1984. Col-
or stability of astaxanthin pigmented rainbow trout
under various packaging conditions. J Food Sci [NSA] Norwegian Standards Assn. 1996. Forslag til
49:1337–40.
Choubert G, Blanc JM, Courvalin C. 1992. Muscle car-
otenoid content and color of farmed rainbow trout
fed astaxanthin and canthaxanthin as affected by
cooking and smoke-curing procedures. Int J Food Offer G, Trinick J. 1983. On the mechanism of water-
Sci Technol 27:277–84.
Christiansen R, Struksnaes G, Estermann R, Torris-
sen O-J. 1995. Assessment of flesh color in Atlantic Pozo R, Lavety J, Love RM. 1988. The role of dietary ␣-
salmon, Salmo salar L. Aquac Res 25:311–21.
[CIE] Commission Intl. de l’Eclairage. 1976. Official
recommendations on uniform color space, color
difference equations and metric color terms. CIE Refsgaard HHF, Brockhoff PB, Jensen B. 1998. Senso-
publ. 15. Suppl. 2. Colorimetry. Paris: Commission
Intl. de l’Eclairage.
Dobbs KL, Brodsky MH, Warburton DW. 1992. A retail
survey of smoked ready-to-eat fish to determine Rørå AMB, Kvåle A, Mørkøre T, Rørvik K-A, Steien
their microbiological quality. J Food Prot 55:208–10.
Dunajski E. 1979. Texture of fish muscle. J Texture
Stud 10:301–18.
bow trout fillets during frozen storage. J Food Sci
56(4):969–72, 984.
[NSA] Norwegian Standards Assn. 1994. Atlantic
salmon—colour and fat measurement. NS9402.
Oslo: Norwegian Standards Assn.
norsk standard for trimmings—og skinningsgrad-
er av laksefilet (Proposal on Norwegian standard
for trimming and skinning grades of salmon fil-
let). NS9601. Oslo: Norwegian Standards Assn.
holding in meat: the swelling and shrinking of
myofibrils. Meat Sci 8:245–81.
tocopherol (vitamin E) in stabilising the canthax-
anthin and lipids of rainbow trout muscle. Aquac-
ulture 73:165–75.
ry and chemical changes in farmed Atlantic salm-
on (Salmo salar) during frozen storage. J Agric Food
Chem 46:3473–9.
SH, Thomassen MS. 1998. Process yield, color and
sensory quality of smoked Atlantic salmon (Salmo
salar) in relation to raw material characteristics.
Food Res Int 31:601–9.
Dyer WJ. 1968. Detoriation and storage life of frozen
fish. In: Hawthorne J, Rolfe EJ, editors. Low temper- Rørvik LM, Yndestad M. 1991. Listeria monocytoge-
ature biology of foodstuff. Oxford: Pergamon Press
Ltd. p 429–47.
Einen O, Mørkøre T, Rørå AMB, Thomassen MS. 1999. Scott TM, Rasco BA, Hardy RW. 1994. Stability of krill
Feed ration prior to slaughter—a potential tool for
managing product quality of Atlantic salmon
(Salmo salar). Aquaculture 178:149–69.
Einen O, Waagan BO, Thomassen MS. 1998. Starva-
tion prior to slaughter in Atlantic salmon (Salmo Sheehan EM, O’Connor P, Sheehy DJA, Buckley DJ,
salar). I. Effects on weight loss, body shape, slaugh-
ter- and fillet-yield, proximate and fatty acid
composition. Aquaculture 166:85–104.
Fennema OR. 1990. Comparative water holding
properties of various muscle foods. J Muscle Foods Sigurgisladottir S, Ingvarsdottir H, Torrissen OJ,
nes in foods in Norway. Int J Food Microbiol 13:97–
104.
Conclusions
REEZING BEFORE AND/OR AFTER SMOKING
meal, astaxanthin and canthaxanthin color in cul-
tured rainbow trout (Oncorhynchus mykiss) fillets
during frozen storage and cooking. J Aqua Food
Prod Tech 3:53–63.
F
affected the yield and the quality of
cold smoked salmon. Product yield in-
creased when the raw material for smoking
was frozen and thawed before it was
smoked, due to increased water content in
the final product. Despite the fact that the
pigment concentration of astaxathin in the
fillets decreased during freezing, the color
intensity (c) and the lightness (L*) of the fil-
lets increased.
FitzGerald R. 1998. Stability of astaxanthin and
canthaxanthin in raw and smoked Atlantic salm-
on (Salmo salar) during frozen storage. Food Chem
63:313–7.
1:363–81.
Cardinal M, Hafsteinsson H. 2000. Effects of freez-
ing/thawing on the microstructure and the texture
of smoked Atlantic salmon (Salmo salar). Food Res
Int 33:857–65.
Francis FJ, Clydesdale FM. 1975. Food colorimetry:
theory and applications. Westport, Conn.: The AVI
Publishing Co. 170 p.
Gormley TT. 1992. A note on consumer preference Simpson KL. 1978. The recovery of protein and pig-
of smoked salmon color. Irish J Agric Food Res
31:199–202.
Hermansson AM. 1986. Water- and fat holding. In:
Mitchell JR, Ledward DA, editors. Functional prop-
erties of food macromolecules. London: Elsevier
Ltd. p 273–314.
Ingemansson T, Pettersson A, Kaufman P. 1993. Lipid
hydrolysis and oxidation related to astaxanthin
content in light and dark muscle of frozen stored Truelstrup Hansen L, Gill TA, Røntvedt SD, Huss HH.
rainbow trout (Oncorhynchus mykiss). J Food Sci
58:513–8.
Jemmi T. 1993. Listeria monocytogenes in smoked
fish: an overview. Arch Lebensm Hyg 44:10–3.
Jensen C, Birk E, Jokumsen A, Shibsted LH, Bertelsen
G. 1998. Effects of dietary levels of fat, ␣-tocopherol
ments from shrimp and crab meals and their use in
salmonid pigmentation. In: Muzzarelli RAA, Pariser
ER, editors. Proceedings of the first international
conference on chitin/chitosan. Cambridge, Mass.:
MIT Sea Grant Program. 254 p.
Skrede G, Storebakken T. 1986. Instrumental color
analysis of farmed and wild Atlantic salmon when
raw, baked and smoked. Aquaculture 53:279–86.
Our results show that the effects of freez-
ing on smoked fillet yield and quality also
depend on the treatment of the raw material
before slaughter, in this case feed ration lev-
el. The effects of freezing may be different if
other salting methods than brine injection
and other smoking conditions than those
used in the present study are used.
1996. Importance of autolysis and microbiological
activity on quality of cold-smoked salmon. Food
Res Int 29:181–8.
Wilding P, Hedges N, Lillford P. 1986. Salt-induced
swelling of meat. The effect of storage time, pH,
ion-type and concentration. Meat Sci 18:55–75.
References
Andersen UB, Stromsnes AN, Steinsholt K, Thomas-
sen MS. 1994. Fillet gaping in farmed Atlantic
salmon (Salmo salar). Norw J Agric Sci 8:165–79.
Anderson HJ, Bertelsen G, Christophersen AG, Ohlen
A, Shibsted LH. 1990. Development of rancidity in
salmonid steaks during retail display. Z Lebensm
Unters F A 191:119–22.
and astaxanthin on color and lipid oxidation dur- MS 20020687 Submitted 12/4/02, Revised 2/21/03,
ing storage of frozen rainbow trout (Onchorhyn-
chus mykiss) and during chill storage of smoked
trout. Z Lebensm Unters F A 207:189–96.
Karube I, Matsouoki H, Suzuki S, Watanabe E, Toyama K.
1984. Determination of fish freshness with an en-
zyme sensory system. J Agric Food Chem 32:314–9.
Krinsky NI. 1989. Antioxidant functions of caro-
tenoids. Free Radical Biol Med 7:617–35.
Accepted 6/4/03, Received 6/4/03
The authors would like to express their gratitude to Troll
Salmon, Norway, for allowing us to use their plant. This work
was supported by the Norwegian Research Council.
Behnke JR, Fennema O, Cassens RG. 1973. Rates of
post mortem metabolism in frozen animal tissues.
J Agric Food Chem 22:5–11.
Bilinski E, Jonas REE, Lau YC, Gibbard G. 1977. Treat-
ments before storage affecting thaw drip forma-
tion in Pacific salmon. J Fish Res Board Can
34:1431–4.
Authors Rørå and Einen are with AKVAFORSK,
Inst. of Aquaculture Research Amer. Somoa., PO
Box 5010, NO-1432 Ås-NLH, Norway. Direct in-
quiries to author Rørå (E-mail: mia.rora@
akvaforsk.nlh.no).
Lavèty J, Afolabi OA, Love RM. 1988. The connective
tissues of fish. IX. Gaping in farmed species. Int J
Food Sci Technol 23:23–30.
Lie Ø, Huse J. 1992. The effect of starvation on the
2128 JOURNAL OF FOOD SCIENCE—Vol. 68, Nr. 6, 2003
JFS is available in searchable form at www.ift.org