AOAC 2018 Methods

(e) Recovery, repeatability, and intermediate reproducibility precisions. – Precision data generated on the four selected samples are summarized in Table 2018.03I for esters forms and Table 2018.03J for free forms. Precision was assessed at levels close to LOQ, on unspiked samples for bound forms and spiked samples at 5 µg/kg for free forms. Samples were analysed in duplicate, at six different occasions, each time on a different day involving different analysts, with independent preparation of reagents and calibration curve. Repeatability and intermediate reproducibility are below target of 22% and 44% respectively. For recovery assessment, selected powdered and liquid infant formula were spiked at 25 µg/kg and 3 µg/kg respectively for bound forms, and spiked at 5 µg/kg for MCPD free form on both powdered and liquid formula. Average recoveries are within the target values of 70 to 125%, at low levels. In the absence of highly contaminated infant and nutritional formula (in liquid and powder form), spiking experiments were performed to assess recovery, repeatability and intermediate reproducibility at high levels (around 70% of the upper level of the calibration curve). One infant formula powder sample (IF powder 2) was spiked at 1800 µg/kg, and one adult nutritional sample (Adult Nut 4 liquid) was spiked at 300 µg/kg for bound forms. Both samples were spiked at 700 µg/kg for free 3-MCPD and 2-MCPD. Average recoveries are within 91% - 107%. Repeatability and intermediate reproducibility are below 4.6% and 6.9% respectively and fulfill method performance requirements (RSD r below 22%). (f) QC Material. – In absence of reference material for Infant formula and adult nutrition, two samples from FAPAS Proficiency Testing that cover the analytes in the scope of this method were included as QC material and analyzed with each analytical batch during the validation. The purpose was not to replace accuracy evaluation of the method, but to demonstrate that part of the method are controlled. For the bound form, usage of an Oil QC sample demonstrate reliability of glycidyl ester conversion into MBPD-ester, transesterification step as well as accuracy of calibration curve. Similarly, for the free form, usage of a soya sauce QC sample allow to control transfer of free form from the extraction solvent to derivatization solvent and accuracy of calibration curve. Results are shown in the Table 2018.03K . (g) SPIFAN II kit samples. – Five samples from SPIFAN II kit were analyzed in duplicate. In absence of composition list and shelf life, samples were analyzed with the SPE cleanup to avoid artefact from mono- and diglycerides. Results details are sown in Table 2018.03L .

References: (1) Leigh, J.K., & MacMahon, S. (2016) J. Agric. Food Chem. 64 , 9442−9451 (2) Zelinkova, Z., Giri, A., & Wenzl T. (2017) Food Control 77 , 65-75

(3) Official Methods and recommended practices of the AOCS (2013) AOCS Official Method Cd 29a-13 (4) Wenzl, T., Samaras, V., Giri, A., Buttinger, G., Karasek, L., & Zelinkova, Z., (2015) EFSA supporting publication 12 , EN-779, 78 J. AOAC Int . (future issue) AOAC SMPR 2017.017 (MCPD) J. AOAC Int . 101 , 324(2018) DOI: 10.5740/jaoacint.SMPR2017.017