An ion-pair high-performance liquid chromatographic (HPLC) method has
been developed and validated for the estimation of Class III caramel a
dded to foods. The method is based on measurement of a component of th
e caramel that has been found in all Class III caramels analysed to da
te and is considered to be a characteristic marker substance. The abun
dance of this component is correlated to the amount of ammonia incorpo
rated into the caramel. Since the nitrogen content of Class III carame
ls from different manufacturers and different formulations varied over
a 2.3-fold range, whilst still described as Class III caramels, this
variability in the composition of the food additive limits the HPLC me
thod quantitatively to a relative standard deviation of 38%. A variety
of beers, biscuits, gravy powders, savoury spreads, confectionery pro
ducts and baked foods were tested using the method. There were no fals
e positives and no false negatives. The limit of detection of the meth
od was 0.1 g litre(-1) for beers and 0.3 g kg(-1) for solid foods. For
biscuit samples supplied by manufacturers with a known caramel conten
t, analysis found 38-92% of the caramel concentration declared by the
manufacturer. For beers, the values were 34-78%. Since the analytical
recovery from biscuits and beers spiked with caramel and analysed imme
diately was typically 100 +/- 10%, this is taken as evidence that the
marker component is only partially stable during the manufacture, pack
aging and storage of the foods. Indirect support for this postulate wa
s the finding that 3-year-old caramels (ex. 1990) contained on average
only 34% of the marker substance compared to fresh (ex. 1993) caramel
s after allowance was made for the nitrogen content. More controlled s
tability trials are underway. It is concluded, therefore, that the HPL
C method provides a simple, robust and semiquantitative measure of the
presence of Class III caramel in foods. Crown copyright (C) 1996 Publ
ished by Elsevier Science Ltd.