Effective detection and quantification of chemical adulterants in model fat-filled milk powders using NIRS and hierarchical modelling strategies.

Skimmed milk powder can be blended with indigenous vegetable oils to formulate fat-filled milk powder (FMP) to satisfy the growing demand of low-income consumers for dairy proteins in the developing countries where food control systems may be fragmented. Unscrupulous manufacturers may adulterate FMP with melamine or urea to give the false impression that it contains sufficient protein. This study investigated, for the first time, the efficacy of near infrared spectroscopy to detect and quantify melamine and urea (0.01-16.00%) in FMP formulated with 4 different vegetable oils (i.e., coconut, palm, soya-bean and sunflower). Multilevel analyses were able to detect, confirm and differentiate the adulterations with an efficiency ranging from 89.8 to 100.0%. The partial least square regression models yielded satisfactory predictions (R2p ≥0.96, RSR ≤0.19) at adulteration levels ≥1.00%. This study provides appropriate guidelines for cost-efficient screening of FMP products for adulterants to protect public health.

Chemometric studies of the effects of milk fat replacement with different proportions of vegetable oils in the formulation of fat-filled milk powders: Implications for quality assurance.

Bovine milk lipids can be replaced with cheaper indigenous vegetable oils to produce milk alternatives with healthier saturated/unsaturated fat balance for those in areas where milk supply is poor or even absent. A wide range of vegetable oils can be used, but their impacts when blended with skimmed milk powder to formulate filled milk powder (FMP) are still unknown. We investigate the baseline variances in 12 types of FMP produced onsite with 3 proportions (10%, 20% and 30%) of 4 different vegetable oils (i.e., coconut, palm, soya-bean and sunflower) using fatty acid- and near infrared spectra profiles. Chemometric analyses revealed 8 significant overlapping clusters of FMP types but 100% classification efficiency was achieved. Sunflower oil, and particularly soya-bean FMP types had statistically the lowest indices of atherogenicity and thrombogenicity. This appears to be the first chemometrics study of FMP; the spectral analytical models used may be effective for product monitoring.