Storage lives of frozen fish
S. Balamurugan asked for '... any information on shelf life study of frozen
raw fish & fishery products (Raw Fish fillet, breaded fish fillet, marinated
fish fillet, breaded shrimp, marinated shrimp, breaded squid and crab meat,
raw surimi and surimi based products)'. Other than for raw fillets I do not
think that this information is available, at least, not in the open
literature. Anyway, storage lives of products is a matter for the Quality
Assurance programme of the company making the products and the company
should make their own determinations of storage lives.
I should justify these opinions and following is an essay on the subject.
Though I am long retired from active involvement in practical work my
comments are based on previous long experience of measuring and
investigating changes in properties of frozen fishery products,
predominately sensory properties, during storage.
The basic procedure for determination of storage life as described in the
scientific and technical literature is to store the product under specified
conditions, evaluate some property and on the basis of the rate of change of
this property declare an end to storage life. I do not have a copy of the
current issue of the IIR's 'Red Book', 'Recommendations for the Processing
and Handling of Frozen Foods', but an earlier edition had two criteria for
storage life, (or 'shelf life' if you like):
1. High Quality Life (HQL) defined as the elapsed time between freezing of
an initially high quality product and the moment when 70% of experienced
tasters successfully distinguish the product from the control in a
triangular or duo-trio test; and
2. Practical Storage Life (PSL) defined as the period of frozen storage
after freezing of an initially high quality product during which the
organoleptic quality remains suitable for consumption or for the process
intended.
These two criteria do not necessarily, indeed almost certainly will not,
lead to the same value of storage life, and require different experimental
procedures for their evaluation.
HQL
This definition specifies the experimental procedures that may be used, the
triangular and the duo-trio comparison tests. In the triangular test the
assessor is presented simultaneously with three samples, two of which are
identical, either the reference product or the test, and the other is
different. The assessor is required to choose the odd sample. In the
duo-trio test the reference sample is first presented to the assessor
followed by two samples, one of the test material and one of the reference.
The assessor then has to select which is the reference. In the duo-trio test
there are two possible answers so the probability of selecting the correct
one by guessing is 0.5; in the triangle test there are three possible
selections the probability of guessing correctly is 0.33. The definition of
HQL requires that a proportion of 0.7 of the panel at least must be
successful in the test, a proportion which is between the proportions of
correction judgements for a probability of 0.5 of detection in the two
tests. (This proportion of correct selections seems to be based on the
procedures used at the US Department of Agriculture's Western Regional
research laboratories in their studies of changes in frozen stored fruits).
The two tests differ in power, that is in sensitivity to a given size of
sensory difference between the products, which can give rise to differences
in the statistical significance of the outcomes of applying the two tests in
the same situation. However the criterion for HQL does not require a
statistically significant result to meet the criterion of 70% correct. Using
a panel of 10 assessors, 7 correct identifications would be significant at
the p=0.05 level in the triangle test, but not in th duo-trio test. Other
points to note are that these tests are very sensitive, the selection is
made on any sensory characteristic whether or not that characteristic is
relevant to the general consumer's appreciation of acceptability, and the
definition requires experienced assessors to be used who are likely to pick
up smaller differences than a consumer would.
There are practical difficulties in carrying out comparison tests such as
the triangle and the duo-trio tests in the case of fishery products. In both
tests two portions of one product are presented and it is crucial for the
correct operation of the tests that they be identical. This requirement is
almost impossible to meet in fish products. The differences in sensory
properties between parts of a single fillet within a fish or between fish
within a batch are well within the sensitivities of either test and unless
the samples from the batches are comminuted and thoroughly mixed the
panellists could well find a greater difference between the nominally
identical portions than between the reference and test products. However
comminution of fish alters both its flavour and texture and these effects
could obscure the changes induced by storage.
The two tests referred to are very sensitive and very small differences can
be detected. I can not recall any publications on determining HQL of frozen
fish products, but my own experience of using the triangle test with fishery
products and of changes in sensory properties of stored fish products lead
me to suppose that HQL's would be of the order of less than a week for
frozen white fish stored at -18ºC.
PSL
Practical Storage Life, as the wording itself suggests, is much more
relevant to the food technologist and food processor than is HQL, and is the
basis of storage lives quoted in various tables of recommended storage
lives. It is also more compatible with the EU definition of the length of
time a product can be stored under the stated conditions of storage while
retaining its intrinsic properties. The IIR definition requires that the
frozen product is of high quality at the time of freezing and presumably
would have the necessary intrinsic properties at that time. The IIR
definition is quite specific that sensory properties are to be evaluated,
though this is only implied in the EU definition. This does not rule out the
use of non-sensory tests but they must be considered as secondary measures
of quality and must be standardised against the results of sensory tests.
The definitions do not specify applicable experimental procedures, or even
the principles of applicable procedures, but there are several relevant
considerations to take into account when devising a procedure. The main ones
to my mind are: the sensory properties to be measured, the composition and
size of the panel that will evaluate the properties, the criterion for
deciding on the end of storage life, and the nature of the product to be
evaluated and conditions of storage. They first three are interrelated.
Odour, texture, and appearance change during frozen storage fishery
products. In lean fish a musty/cardboardy flavour develops which gets more
intense during storage. Fatty fish get progressively more rancid. In both
types of fish the texture becomes progressively firmer, tougher, more
fibrous, and drier. These attributes can be assessed and scored separately
as objective sensory properties and typically they would be evaluated by a
panel of expert assessors in a laboratory situation. The panel mean scores
can then be related to storage time and the effects of circumstance like
prior processing or of packaging can be determined. Objective sensory data
like this does not of itself provide an estimate of storage life; the data
has to be interpreted by the fish technologists against a criterion for end
of storage life that is compatible with the IIR or EU definitions. The
criterion in the EC Directive is difficult to apply because a frozen stored
product does not really necessarily lose its characteristic properties. So
far as flavour is concerned the intrinsic flavours are not lost but are
overlaid and eventually masked by cold storage and a rancid flavours
produced during storage. The texture of fish varies considerably both
between and within species and it is not easy to specify when the
progressive changes in texture that occur during frozen storage are no
longer 'intrinsic'. I am not aware of at any consumer trials to determine
the acceptability/unacceptability of cold storage and rancid flavours or of
changes in texture of frozen stored fish, but my opinion for what it is
worth is that consumers will probably tolerate a greater change in texture
than they would a change in flavour and that texture is not a limiting
factor for end of storage life. With regard to flavour I would suggest an
objective score corresponding to slight to moderate intensity of off flavour
would be a suitable criterion for end of storage life. At this level the off
flavour would be detectable by the majority of consumers, but would not be
strong enough to mask intrinsic flavours. However, objective sensory scaling
is just described has rarely been used in determination of storage lives of
frozen fish other than a few studies with oily fish when rancidity has been
measured and slight rancidity use as the criterion end of storage life.
In a few studies of frozen storage of fishery products hedonic scoring has
been used to monitor changes during storage. For example, to quote from a
published paper, 'Trained panels judged portions for texture, juiciness,
flavor and overall desirability on an intensity scale ranging from 9, "liked
extremely," to "disliked extremely". This is truly an hedonic scale, but I
would argue that such a scale should not be used by trained assessors.
Hedonic scales measure subjective immediate impressions which are personal
to particular person. It is not possible to train, in the sense used in
sensory evaluation, assessors to like or dislike products. In my view
hedonic scaling should be used only with large panels of untrained
assessors, but in all the studies I am aware of where hedonic rating has
been used in the context of determining storage lives of either chill stored
or frozen stored fishery products small laboratory panels have been used.
Potentially hedonic rating it suitable for determining storage lives because
the IIR criterion of 'suitable for consumption' can be defined on the scale
if you consider something that is liked is suitable for consumption and
something that is disliked is not.
Most publications on changes during frozen storage of fishery products that
I have examined, especially those oriented to measure effects of
temperature/processing/packaging on storage lives, have used a sort of
quality scale. For example, to quote from one paper, the products were
'rated for flavor and texture ... on a scale of one to five (1 = poor, 5 =
very good) by a six member panel with experience in evaluating seafoods.'.
It is assumed here that the qualifiers referred to quality, but in some
papers the word 'quality' is explicitly stated. Such scales require that the
assessors are aware of what attributes of the product contribute to quality
and the their characteristics and intensities that define levels of quality.
This would be an objective of selection, briefing, and training of the
assessors. Quality scales can be considered appropriate for estimating
storage lives because both the IIR and the EU criteria can be defined on
quality scales. A good quality product would retain its intrinsic
properties, would be suitable for consumption, and would be suitable for any
further processing operation. What could be debated is whether the
appreciation of quality by assessors in a laboratory situation is the same
as that in a commercial company.
Stored frozen products change during storage and it is assumed that these
changes are in the direction that consumers would appreciate as reduction in
the pleasantness of their eating experience, that is, the products are
reducing in 'quality'. The manufacturer of the product wants to ensure that
the customers have a pleasant experience when consuming the product and,
will make repeat purchases of the product. The quality managers are well
aware of this and arrange their quality assurance programme to ensure the
requisite quality in the product leaving the factory, and, in the context
we're discussing, the product has an appropriate shelf life date on the
packaging. Many processors though might not have the same opinion as to how
much change their consumers will accept, and could come to different
conclusions as to what would be the appropriate shelf life date. Processors
have to compromise; they do not want to put too short a shelf life date
because that might mean consumers have to dispose of out of date products
too often, but do not want to put a long a date which would mean that the
product was unpleasant at the time consumption. Perhaps it is as cynical
view, but the shelf life date could be defined as the longest date that the
manufacturer can put on the product without giving rise to an unacceptable
level of consumer complaints.
All this means that it is not easy to to come t a conclusion as to storage
lives of products under laboratory conditions; the criteria that a
laboratory might use to define the end of storage life might not be the same
that would be used in industry. Determining storage lives as an aid to date
stamping of products is really the responsibility of industry rather than of
a regulatory body. Frozen stored products would not become unsafe to eat as
a result of prolonged frozen storage and the role of the regulator is to
ensure that be product be fit, not necessarily pleasant, for consumption
which would imply a storage time much longer than that at which a
manufacturing company careful to safeguard its reputation for quality would
put on the package. The processor must also bear in mind that the date stamp
on a processed product starts from the time of production of the product,
not from the time the raw material was frozen. Consider the manufacture of
coated fish portion. The raw material might have been frozen at sea on a
factory vessel, shipped in reefer to a processing plant, converted to fille
blocks and stored in that form until transformed into the consumer product
perhaps many months later. The material will have suffered some cold storage
changes along this supply chain and the manufacturer should bear this in
mind when allocating a date stamp for the consumer product.
Peter Howgate
This archive was generated by hypermail 2b29 : Wed Feb 14 2007 - 01:45:32 PST