Abstract: A proficiency testing (PT) program for the determination of aflatoxin M1 (AFM1) in milk powder was designed and conducted so as to identify the interlaboratory differences and to understand, assess and improve the proficiency of the participant laboratories. Split-level test samples were prepared according to the Guidance on Statistical Treatment of Proficiency Testing Results and Proficiency Evaluation (CNAS?CL002), and the uniformity and stability of the samples were analyzed by one-way analysis of variance and t test, respectively. Robust statistical analysis of the proficiency testing results was conducted to evaluate laboratory proficiency through Z score. The homogeneity and stability of the samples met the requirements of proficiency testing. Of 49 laboratories that reported results for AFM1 in milk powder, 45 (91.8%) performed satisfactorily. The majority of the participant laboratories passed the proficiency testing with satisfactory results, indicating that the laboratory performance for the determination of AFM1 in milk powder was generally good. The outlier results found in a few laboratories was largely ascribed to experimental operation and data processing errors.

Key words: aflatoxin M1; determination; milk powder; laboratory proficiency testing; robust statistics

摘要： 设计并实施乳粉中黄曲霉毒素M1（aflatoxin M1，AFM1）的检测能力验证计划，了解和评价实验室检测乳粉中AFM1的检测能力和水平，识别实验室间存在的差异，促进实验室提高检测水平。参照CNAS—CL002《能力验证结果的统计处理和能力评价指南》制备分割水平对的样品，采用单因素方差分析和t检验对样品均匀性和稳定性进行分析。对能力验证结果进行稳健统计分析，通过Z值比分数评价实验室能力。结果表明：样品通过均匀性和稳定性的检验要求，满足能力验证计划要求；参加实验室49 家，满意结果数为45 家，满意率为9

关键词: 黄曲霉毒素M1；检测；乳粉；实验室能力验证；稳健统计