2.6. Sample preparation for analysis of vitamin D3

The current analytical methods for vitamin D analysis are time consuming, labor intensive, require experienced analysts, and have only been validated for a few materials. The official methods available are relatively similar and involve saponification and extraction, clean‐up steps and separation using high‐performance liquid chromatography, and detection with diode array, with relative standard deviations between 10% and 15% (Byrdwell, DeVries, Exler, Harnly, Holden, & Holick, 2008). In this study, analysis of vitamin D3 was performed using Vitakit D™ (SciMed Technologies, Canada, USA), which is a competitive enzyme immunoassay kit. The enzyme‐linked immunosorbent assay (ELISA) can detect vitamin D3 between 0.125–0.75 IU/ml, where none of the samples in this study fell outside the detectable range, and the intra assay relative standard deviations for the ELISA was 6.8%.

Each of the food samples were diluted with deionized water to a fat content of 1–3%, and then mixed and homogenized using a domestic blender (Multi‐quick, Braun 300, Havant, UK), and three independent replicates of 1 g (wet weight) were weighed prior to the addition of 0.55 g of potassium hydroxide (laboratory reagent grade; Fisher Scientific) into 15 ml centrifuge tubes. The tubes were gently mixed and left uncapped for 2 min in the dark. The tubes were then capped and incubated in the dark for 4 min, followed by 2 min of vigorous shaking; this step was repeated twice. 2 ml of hexane (high‐performance liquid chromatography grade; Fisher Scientific) was then added to the tubes, which were then capped and shaken vigorously for another 2 min in the dark. Centrifugation at 3500 RCF for 10 min at room temperature was then performed. 200 μl of the upper organic phase was transferred to an amber screw cap glass vial.

10 μl of calibrators, extracted samples and controls were pipetted into the ELISA plate accordingly. The plate was shaken for 8 min on a plate shaker (180 ± 10 rpm) to evaporate the hexane. 60 μl of assay buffer was added to each well and mixed gently for 30 s. A lid was placed over the plate and shaken for 5 min (180 ± 10 rpm). 60 μl of anti‐vitamin D3 conjugate with horseradish peroxidase diluted in conjugate diluent was added to each well and gently mixed for 20 s. The plate was covered and shaken for 10 min in the dark (180 ± 10 rpm). A microplate washer (Labtech, LT‐3000, East Sussex, UK) was used to wash the plate six times with 380 μl/well of distilled water. After washing, the plate was tapped against absorbant paper until no trace of water was visible on the paper. 60 μl of substrate was added to each well and gently mixed for 10 s. The plate was then incubated in the dark for 5 min. Finally, 60 μl of stopping solution (0.2 M H₂SO₄) was added to the plate and gently mixed for 10 s. A Microplate Reader (Thermo Fisher Scientific, Multiskan Ascent, MA, USA) was used to measure the absorbance at 450 nm immediately.