目的 基于质量源于设计(QbD)理念，优化板蓝根配方颗粒喷雾干燥工艺。方法 结合信息熵赋值法，以喷干粉得粉率及尿苷、腺苷、鸟苷和(R，S)-告依春含量作为关键质量属性(CQAs)，利用Plackett-Burman试验设计对进风温度、雾化压力、药液温度、进液速度和药液相对密度进行考察，筛选出关键工艺参数(CPPs)，再运用中心点复合设计(Center Point Composite design，CCD)试验对筛出的CPPs进行优化，在二次多项式回归模型的基础上建立板蓝根配方颗粒喷雾干燥工艺设计空间，并加以验证。结果 Plackett-Burman试验结果显示，药液相对密度和进液速度为该研究的关键工艺参数。CCD试验方差分析结果显示构建的模型具有显著性差异(P<0。01)，失拟项不显著(P>0。05)，表明模型预测能力良好。构建的CPPs优化设计空间为药液相对密度1。05-1。08，进液速度30%-40%。结论 基于QbD理念建立的板蓝根配方颗粒喷雾干燥工艺设计空间，可提高其工艺的稳定性，有利于保障产品质量的一致性。
Study on Optimization of spray Drying Process of Banlangen Formula Granules
Objective To optimize the spray drying process of Banlangen(Isatidis Radix)formula granules based on quality by design(QbD)concept.Methods Using powder yield and the contents of uridine,adenosine,guanosine,and(R,S)-goitron as the critical quality attributes(CQAs),Plackett-Burman design was used to screen out critical process parameters(CPPs)for inlet temperature,spray pressure,liquid temperature,pump speed,and liquid relative density.The central-composite design(CCD)test was used to optimize the CPPs,which were screened.Based on the quadratic polynomial regression model,the design space of spray drying process of Banlangen(Isatidis Radix)formula granules was established,and further validated by experiments.Results Plackett-burman test results show that liquid relative density and inlet velocity are the key parameters for the study.The variance analysis results of CCD test showed that the constructed model in a good prediction ability,since the P-values of model was less than 0.01 and P-values of items lack of fit was more than 0.05.The optimized design space of CPPs was the liquid relative density 1.05-1.08,and pump speed 30%-40%.Conclusion Based on the QbD concept,the design space for the spray drying process of Banlangen(Isatidis Radix)formula granules can improve the stability of its process and help ensure the consistency of product quality.
Quality by designBanlangen(Isatidis Radix)formula granulesSpray dryingPlackett-Burman experimental designCentral-composite design
NETL
NSTL
万方数据
