目的 应用Box-Behnken设计-响应面法，结合基准关联度和层次分析法（analytic hierarchy process，AHP）-熵权法优化经典名方小承气汤的提取工艺。方法 以芦荟大黄素和大黄酚、厚朴酚、辛弗林含量、出膏率及指纹图谱相似度为关键质量属性（critical quality attributes，CQAs），以加水倍量、提取时间、提取次数为关键工艺参数（critical process parameters，CPPs），采用单因素实验确定各因素水平，Box-Behnken设计优化提取工艺参数并进行验证。最后，计算不同提取参数的各评价指标下各样品与基准样品的基准关联度，采用AHP-熵权法确定各种评价指标的权重系数，进行综合评分，预测最佳提取工艺。结果 方差分析结果显示，2次项回归模型的方差显著（P＜0.01），且失拟值不显著，表明所建模型具有统计学意义，各因素与响应值之间的关系可以用所建模型进行函数化。确定最佳提取工艺为全方饮片加6倍量水，提取30 min，提取1次。平行3次的验证试验综合评分均值为95.37，RSD为2.31%，且符合基准样品质控标准。结论 基于质量源于设计（quality by design，QbD）理念、基准关联度和AHP-熵权法筛选的经典名方小承气汤提取工艺参数稳健可靠，同时为其他经典名方的制备工艺开发提供了参考。

Objective The extraction process of the classical prescription Xiaochengqi Decoction (小承气汤, XD) was optimized based on Box-Behnken experimental design, standard relation and analytic hierarchy process combined with entropy method. Methods With the contents of aloe emodin and chrysophanol, magnolol, synephrine as well as the paste-forming rate and fingerprint similarity as critical quality attributes (CQAs). The amount of water, extraction time and frequency were used as the critical process parameters (CPPs). The single factor experiment design was used to determine the level of each factor, and the Box-Behnken experimental design was used to optimize the extraction process. Finally, the standard relation between each samples and standard decoction of various indexes of different extraction parameters was calculated, while AHP-entropy method was applied to determine the weight of each index for comprehensive score. Results The results of variance analysis showed that the P value of the quadratic regression model was significant (P < 0.01), and the lack-of-fit value was not significant, indicating that the model was statistically significant. The relationship between various factors and the response values could be functionalized by the established models. Through further optimization of design space, the optimal extraction process of XD was 6-fold volume of water, extracted once and extracted for 30 min. The average comprehensive evaluation of three parallel verification tests was 95.37, RSD value was 2.31%, and meet the quality requirements of standard decoction. Conclusion The extraction process parameters of XD based on the quality by design (QbD), standard relation and analytic hierarchy process combined with entropy method was stable and reliable, which would provide guidance for the process development of other classical prescriptions.

国家自然科学基金资助项目(82173989)