Single Laboratory Validation and Uncertainty Estimation of a HPLC Analysis Method for Deoxynivalenol in Noodles

• Published in: Han'gug sigpum wi'saeng anjeonseong haghoeji Vol. 26; no. 2
• Main Authors: Ok, H.E., Korea Food Research Institute, Seongnam, Republic of Korea, Chang, H.J., Korea Food Research Institute, Seongnam, Republic of Korea, Kang, Y.W., National Institute of Food and Drug Safety Evaluation, Cheongwon, Republic of Korea, Kim, M.H., National Institute of Food and Drug Safety Evaluation, Cheongwon, Republic of Korea, Chun, H.S., Korea Food Research Institute, Seongnam, Republic of Korea
• Format: Journal Article
• Language: Korean
• Published: 01.06.2011
• Subjects: Dried noodle; HPLC; Ramyeon; Uncertainty; VOMITOXIN; VOMITOXINA; VOMITOXINE
• ISSN: 1229-1153

An isocratic high performance liquid chromatography (HPLC) method for routine analysis of deoxynivalenol in noodles was validated and estimated the measurement uncertainty. Noodles (dried noodle and ramyeon) were analyzed by HPLC-ultraviolet detection using immunoaffinity column for clean-up. The limits of detection (LOD) and quantification (LOQ) were 7.5 ㎍/kg and 18.8 ㎍/kg, respectively. The calibration curve showed a good linearity, with correlation coefficients r² of 0.9999 in the concentration range from 20 to 500 ㎍/kg. Recoveries and Repeatabilities expressed as coefficients of variation (CV) spiked with 200 and 500 ㎍/kg were 82±2.7% and 87±1.3% in dried noodle, and 97±1.6% and 91±12.0% in ramyeon, respectively. The uncertainty sources in measurement process were identified as sample weight, final volume, and sample concentration in extraction volume as well as components such as standard stock solution, working standard solution, 5 standard solutions, calibration curve, matrix, and instrument. Deoxynivalenol concentration and expanded uncertainty in two matrixes spiked with 200 ㎍/kg and 500 ㎍/kg were estimated to be 163.8±52.1 and 435.2±91.6 ㎍/kg for dried noodle, and 194.3±33.0 and 453.2±91.1 ㎍/kg for ramyeon using a coverage factor of two which gives a level of statistical confidence with approximately 95%. The most influential component among uncertainty sources was the recovery of matrix, followed by calibration curve.