Aflatoxin Testing in Peanuts: A Proficiency Assessment Scheme for Chinese Analytic Laboratories

作者:Bao Lei*; Bao Zhenmin; Zhang Yibing; Liang Chengzhu; Lu Ning; Liu Xintong; Jin Ying; Ze Pei Jun; Ge Manli; Tian Ling
来源:Journal of AOAC International, 2009, 92(2): 481-486.

摘要

This national assessment program was established by the China National Accreditation Service for Conformity Assessment (CNAS) to evaluate the aflatoxin-testing proficiency of a cross-section of Chinese laboratories. The Shan Dong Inspection and Quarantine Bureau of China conducted the assessment according to ISO 13528:2005 (E) and the International Harmonized Protocol for Proficiency Testing. The 77 laboratories that participated in the study had either been previously accredited by CNAS or were candidates for CNAS accreditation. The analytic samples for this testing scheme were prepared from naturally contaminated peanuts and diluted to approximately 10 mu g/kg for aflatoxin B, and 18 mu g/kg for total aflatoxins. The S(s)/sigma(p) test (with a required result of S(s) <= 0.3 sigma(p)) was used to evaluate the homogeneity of the test samples; sample stability was confirmed with a t-test. The performance of each laboratory was designated by a z-score that was calculated using robust statistics. The robust mean of the participants' results in this study was nearly coincident with the median. A modified Horwitz equation was used to determine the standard deviation. The study compared analytic results obtained by 5 different methods: high-performance liquid chromatography (LC), enzyme-linked immunosorbent assay, thin-layer chromatography, fluorometry, and LC with tandem mass spectrometry. A satisfactory performance rating required z-scores between -2 and +2 for the target analytes. Of the 73 laboratories that reported results for aflatoxin B(1), 66 (90.4%) performed satisfactorily. Of 32 laboratories that reported total aflatoxins (B(1) + B(2) + G(1) + G(2)), 30 (93.8%) performed satisfactorily. Laboratories whose performance ratings were questionable or unsatisfactory were re-evaluated in a second interlaboratory comparison.