Matrix-specific method validation for quantitative analysis of vitamin C in diverse foods

Vitamin C, assayed as total ascorbic acid (AA), was extracted from foods using HPLC with ultraviolet spectrophotometric detection, including treatment of the extract tris(2-carboxyethyl) phosphine (TCEP) to reduce any dehydroascorbic acid to AA. The method was validated for a variety of matrices including fruits and vegetables, fruit juice, dried spices, and high-starch and high-fat foods, using spike recovery, sequential extractions, analysis of available certified reference materials, and verification of AA peak purity. The limits of detection and quantitation were 0.06–0.09 mg and 0.2 mg AA per 100 g food, respectively. The average recovery of added AA from all matrices was 97–103%.The inter-day relative standard deviation (RSD) for matrices including orange juice, fortified cereal, a fruit and vegetable composite, and freeze-dried vegetables was 1.1–2.0% and 4.8%, and HORRAT values (RSD/predicted RSD) for a wide range of foods were <0.1–0.6. Results for certified reference materials, BCR®431 (freeze-dried Brussels sprouts), BCR®421 (milk powder) and VMA399 (dry breakfast cereal) (465 ± 4.6, 74.0 ± 1.1 and 70.5 ± 1.5 mg/100 g, respectively) were within the certified ranges. Without matrix-specific method adjustments to the method validated for orange juice, extraction problems and interferences in the AA peak for particular matrices lead to over- or underestimation of vitamin C in many foods (0.3–70 mg/100 g; 5.5–64%).