目的   建立适用于乙型脑炎减毒活疫苗（乙脑疫苗）生物学关键质量指标（critical quality attribution，CQA）的警戒限和行动限范围，开展趋势分析以及时发现异常趋势，指导疫苗生产。 方法   根据CQA的数据分布及特点，采用休哈特控制图法及阈值法建立限度范围，比较两种方法的结果，并采用适宜的限度范围对近3年的数据进行统计分析。 结果  乙脑疫苗的病毒滴度等生物学CQA数据为非正态分布，采用休哈特控制图建立的行动限范围有超出法规限的情况；而采用阈值法建立的行动限范围则均在法规限以内。采用阈值法建立的限度分析近3年乙脑疫苗各生产阶段生物学指标的状态，各指标超出行动限的点数均小于总点数的1%，表明乙脑疫苗各生产阶段处于稳定状态。 结论  阈值法为建立限度范围的适用方法；趋势分析中应结合数据本身特性与分布，选用适合的方法来建立限度标准，才能更为科学地将趋势分析工具用于指导实际生产。

 Objective  To establish alert and action limit ranges suitable for biological critical quality attribution (CQA) of a live attenuated Japanese encephalitis (JE) vaccine for conducting trend analysis to find abnormal trends in time and guiding vaccine manufacture.  Methods  According to data distribution and characteristics of CQA, the alert and action limit ranges for biological CQA of JE vaccine were established with Shewhart chart and cutoff value approach. The results of two methods were compared and the suitable limit values were used to statistically analyze the latest three year data.  Results  The biological CQA, ie. virus titer, was non-normally distributed, and some action limits established by Shewhart method exceeded regulatory range, while all of the limit ranges developed by cutoff value approach were within the scope of regulatory range. The biological CQAs of JE vaccine produced in the last three years were analyzed using the limit range obtained by cutoff value approach. The CQAs out of action limit were less than 1% of total number, indicating that JE vaccine production was in a stable state.  Conclusions  Cutoff value approach is suitable for establishing the limit ranges. In trend analysis, a suitable method should be selected  to develop the limits with consideration of data characteristics and distribution. Thus, the trend analysis tool can be used scientifically to guide the actual production.