目的 基于质量源于设计（QbD）理念和序贯设计法优化丹参川芎嗪注射液前处理过程中的水沉工艺。方法 研究联用了鱼刺图法及基于全局式序贯设计的中心复合实验法作为优化方法，在基础实验设计中关键参数维度的轴向方向增加实验点，再进行新的多元线性回归拟合，最终得到了水沉工艺的设计空间并对其进行了验证，同时给出了较为稳健且全面的操作空间。结果 根据设计空间，推荐操作空间为水沉体系pH值为3.1～3.4时，加水倍量3.25～5.00，静置时间为7～17 h，静置温度为7℃。结论 首次在中药制药过程工艺优化中引入序贯设计法。相较于样本量巨大的均匀空间网格设计、正交设计等实验设计方法而言，序贯设计在实现与前者同样的样本对可行域覆盖程度的基础上，能够有效降低工艺优化实验的工作量，避免样本与人力物力的浪费，同时保证较高的模型预测性能。

Objective To optimize the water-precipitation in the pretreatment manufacture process of Salviae Miltiorrhizae (Danshen) and Ligustrazine Hydrochloride Injection (SMLHI) based on the quality by design (QbD) concept and the sequential design. Methods A design space was obtained by the global-type sequential design methodology combined with the fishbone-diagram risk analysis and central composite experiment based on global type sequential design method. The preliminary optimized zone of critical process parameters was extended in axial direction on the basis of the statistical result gained from the first central composite design. Then an updated multivariate linear regression model was fitted and a design space was delineated and verified. Results According to the design space, the recommended operation space of the water-precipitation was as follows:when the pH value of the material was between 3.1 and 3.4, 3.25-5.00 times of water adding into the solution was recommened, the standing time was 7-17 h and the standing temperature was 7℃. Conclusion This study introduces the sequential design into the process optimization of traditional Chinese medicine pharmaceuticals. Compared with the experimental design methods such as uniform space grid design and orthogonal design with a large number of samples, while samples cover the same feasible zone in a continuous operation form, sequential design methodology can effectively reduce the workload of the optimization experiment, avoid the waste of manpower and resources as well as guarantee the prediction ability of the fitting model.

国家“重大新药创制”科技重大专项（2018ZX09201011-002）