比较国家标准GB/T 20752-2006、GB/T 21311-2007以及农业部783号公告-1-2006中硝基呋喃类代谢物测定方法的优劣，确定硝基呋喃类代谢物测定的关键影响因素。动物源性食品样品不经甲醇溶液清洗，添加盐酸溶液和2-硝基苯甲醛溶液，置于37 ℃水浴中衍生反应4 h，调节pH为7.4，离心后取上清液过HLB固相萃取柱，收集洗脱液。采用超高效液相色谱-串联质谱法测定样品溶液中4种硝基呋喃类代谢物的含量。以Waters BEH C₁₈色谱柱为固定相，以不同体积比的含2.0 mmol·L^-1乙酸铵的0.05%（体积分数）甲酸溶液和2.0 mmol·L^-1乙酸铵甲醇溶液的混合液为流动相进行梯度洗脱，串联质谱分析中采用电喷雾正离子源和多反应监测模式。采用内标法定量。4种硝基呋喃类代谢物的线性范围均为0.50~20.0 μg·L^-1，检出限（3S/N）为0.05~0.25 μg·kg^-1。以空白样品为基体进行加标回收试验，所得回收率为94.0%~107%，测定值的相对标准偏差（n=6）为1.9%~3.4%。方法应用于能力验证样品（19-N866，19-P653）的分析，测定值与已知值相符。

By comparing the advantages and disadvantages of determination methods of the nitrofuran metabolites between GB/T 20752-2006, GB/T 21311-2007 and ministry of agriculture announcement No. 783-1-2006, the key influencing factors of nitrofuran metabolites determination were determined. Hydrochloric acid solution and 2-nitrobenzaldehyde solution were added to the sample of animal derived food without washing of methanol solution, and then the sample solution was derived in water bath at 37 ℃ for 4 h. pH of the sample solution was adjusted to 7.4. After centrifugation, the supernatant was taken, and then passed through HLB solid phase extraction column. The eluate was collected. UHPLC-MS/MS was applied to the determination of 4 nitrofuran metabolites in the sample solution. Waters BEH C₁₈ chromatographic column was used as stationary phase, and the mixture of 0.05% (volume fraction) formic acid solution containing 2.0 mmol·L^-1 ammonium acetate and 2.0 mmol·L^-1 ammonium acetate methanol solution mixed in different ratios was used as mobile phase in gradient elution. ESI⁺ and multiple reaction monitoring mode were adopted in MS/MS. Internal standard method was used for quantitative. Linearity ranges of the 4 nitrofuran metabolites were found in the same range of 0.50-20.0 μg·L^-1 with detection limits (3S/N) in the range of 0.05-0.25 μg·kg^-1. On the base of blank sample, test for recovery was made by standard addition method; values of recovery found were in the range of 94.0%-107%, with RSDs (n=6) of determined values in the range of 1.9%-3.4%. The proposed method was applied to the analysis of proficiency testing samples (19-N866, 19-P653), giving results in consistency with the known values.