Understanding and reducing oxidative flavour deterioration in foods - Oxidation in Foods and Beverages and Antioxidant Applications - page 124

substances (TBARS) as well as instrumental methods such as HPLC, GC-MS,
NIR, FTIR and DSC. In lipid oxidation studies on meat products there seems to
be a fairly good correlation between TBARS and sensory data, although in many
cases no attempts have been made to statistically evaluate the correlations
(Eckert et al. 1997; Murano et al. 1998; Winne and Dirinck 1997). However,
newer data indicate that even in meat, TBARS may be an unreliable method (see
further discussion in Section 6.3.6 below; Summo et al. 2010).
PV and AV are often used to evaluate the lipid oxidation in oils and the
author's experience is that AV, at least to some extent, correlates with sensory
data in oxidized oils. In contrast, findings from fish oil enriched spreads clearly
showed that the correlation between sensory and PV and AV data was very poor
(Jacobsen 1999b) (Fig. 6.1(a) and (b)). It is not surprising that PV did not
correlate with sensory data since lipid hydroperoxides are not responsible for
off-flavour formation. Moreover, since lipid hydroperoxides are broken down to
secondary oxidation products during the later stages of oxidation a low PV does
not necessarily indicate that the lipids are not oxidized. However, it was
surprising that AV in some samples was high despite a high flavour acceptability
score. This finding could indicate that the sensitivity and specificity of this
method is too low to detect the changes in concentrations of volatiles responsible
for the off-flavour formation. Therefore, AV cannot be recommended for
measurement of lipid oxidation in complex food systems.
Good correlations between sensory data and data from headspace GC analysis
on a range of different oxidized products such as boiled potatoes (Blanda et al.
2010), fish (Milo and Grosch 1995), fish oil enriched milk (Venkateshwarlu et al.
2004a), mayonnaise and dressing (Hartvigsen et al. 2000; Let et al. 2007) have
been extensively reported in the literature and headspace GC analysis can
therefore be recommended as one of the best methods to chemically assess
Fig. 6.1 (a) Peroxide values, and (b) Anisidine values versus overall flavour
acceptability obtained from storage experiment with fish oil enriched spreads. Reprinted
from Jacobsen, C. (1999b). Sensory impact of lipid oxidation in complex food systems.
Fett/Lipid 12, 484±492 with permission from Wiley.