Food Lipids: Chemistry, Nutrition, and Biotechnology

G. Mathematical Modeling
The pathways and relationships involved in lipid oxidation are complex. Consequently,
there is a need to have methods to quantitatively link product composition
to oxidative stability. Mathematical modeling is a tool that allows the synthesis of
data from one or many experiments into an integrated system from which quantitative
changes in many components may be calculated. Critical to any mathematical model
is the endpoint selected for shelf life. Oftentimes, this endpoint is arbitrarily selected,
whereas it should be based on consumer acceptability.
Aside from studies exploring variations in moisture and oxygen concentrations
[278], few studies have attempted to model lipid oxidation, and even fewer focus
on lipid oxidation in muscle foods. Kurade and Baranowski [279] reported that the
shelf life of frozen and minced fish meat might be predicted by measuring total iron
and myoglobin levels and the time for extracted lipids to gain 1% weight. Using
these three variables, deviation from the actual ‘‘shelf life’’ measured 7.38%. On the
other hand, the estimate for frozen shelf life of fish samples obtained by Ke et al.
[280] gave an average deviation of 17% when only the last variable had been used.
In either case, these models are flawed as predictors of the contribution of membrane
lipids to oxidative stability. From isolated model system studies, it is known that the
membrane environment is a major factor in the high susceptibility of phospholipids,
and extraction of the phospholipids eliminates that factor. Experimental support was
provided by Ke et al. [280] who observed that the polar lipids oxidized more slowly
than the neutral lipids.
Using a slightly different approach to model lipid oxidation, Tappel et al. [281]
incorporated into their model some of the major chemical features associated with
lipid oxidation, including peroxidizability of polyunsaturated lipids, activation of
inducers and their initiation of lipid peroxidation, concurrent autoxidation, inhibition
of lipid peroxidation by vitamin E, reduction of some of the hydroperoxides by
glutathione peroxidase and formation of thiobarbituric acid–reactive substances
(TBARS). The equations used to model the reactions were first brought into agreement
with published information on these reactions by the determination of kinetic
factors: activation degradation factor, inducer loss factor, antioxidant use factor, autoxidation
factor, and hydroperoxide reduction factor. Subsequently, when the simulation
program was applied to tissue slice and microsomal peroxidizing systems,
the results of the simulation were in agreement with experimental data.
Babbs and Steiner [282] also used a computation model based on reactions
involved in initiation, propagation, and termination. Their model incorporated 109
simultaneous enzymatic and free radical reactions, and rate constants were adjusted
to account for the effects of phase separation of the aqueous and membrane lipid
compartments. Computations from this model suggested that substantial lipid peroxidation
occurred only when cellular defense mechanisms were weakened or overcome
by prolonged oxidative stress. Consequently, while this model was developed
to understand the contribution of free radical reactions to disease states of living
organisms, useful insights may also be gleaned from it or similar models in understanding
the oxidative stability of muscle food systems. In this manner, the variability
that is inherent in the composition of foods may be factored into shelf life predictions,
and conditions for optimal stability may be derived.
Copyright 2002 by Marcel Dekker, Inc. All Rights Reserved.