Determination of α-oxo-aldehydes and furfurals using HPLC-PDA-ELSD method in sugar-containing food products

Abstract

Four food toxins, α-oxo-aldehydes (glyoxal and Methylglyoxal) and furfurals (furfural and 5-hydroxymethlfurfural) were simultaneously retained on a Phenomenex Luna phenyl hexyl column (150×4.6mm×5 µm,100 Å) using ammonium acetate / acetonitrile (28:72, v/v) as mobile phase at 1 mL/min. The detector setting was tuned at 280 nm to quantify furfurals, whereas α-oxo-aldehydes were quantified using Evaporating light scattering detector (ELSD). The extraction of selected analytes from aqueous and lipid-based media was optimized using liquid-liquid extraction (LLE) method. A sum of 24 different brands (n = 20 in each group) of glucose containing preparations were collected, extracted, and screened for the presence of selected analytes. The observed level of selected analytes was divided into non-detectable, moderately high and above limit samples. The machine learning model viz. decision tree and multilayer perceptron models was optimized and validated for correlation of selected dependent variables with independent variables. The occurrence of 5-HMF depended on state and age of sample, whereas sugar concentration was responsible for furfural presence. The samples with higher 5-HMF levels were categorized as being viscous/semisolid and were preserved with potassium sorbate. About 77.3 % of the 480 samples studied indicated a considerable amount of methylglyoxal, and the food colors were classed as responsible for development of it followed by mixed fruit flavor and vanilla taste flavor as food additives.

Keywords

• Artificial neural network (ANN)
• Α-oxo-aldehydes
• Furfurals
• ELSD