近红外光谱法检测复方甘草酸苷制剂的质量  

Quality study of Compound glycyrrhizin preparation by Near-infrared spectroscopy

作　　者：童一洋 郭志永 闻宇 李杰 薛洪宝 李楠楠 殷晓建 TONG Yiyang;GUO Zhiyong;WEN Yu;LI Jie;XUE Hongbao;LI Nannan;YIN Xiaojian(School of Public Foundation,Bengbu Medical University,Bengbu,Anhui,233030 P.R.China;Department of Production Technology,Anhui Hongye Pharmaceutical Co.,Ltd.,Bengbu,Anhui,233045 P.R.China;Fifth Precinct of Medical Products Administration in Anhui Province,Bengbu,Anhui,233000 P.R.China)

机构地区：[1]蚌埠医科大学公共基础学院,安徽蚌埠233030 [2]安徽宏业药业有限公司生产技术部,安徽蚌埠233045 [3]安徽省药品监督管理局第五分局,安徽蚌埠233000

出　　处：《华西药学杂志》2024年第3期327-332,共6页West China Journal of Pharmaceutical Sciences

基　　金：安徽省高校学科(专业)拔尖人才学术资助项目(编号:gxbjZD2022041);安徽省新时代育人质量工程研究生教育项目(编号:2022lhpysfjd061,2022sshqygzz033)。

摘　　要：目的采用近红外光谱技术结合化学计量学快速检测复方甘草酸苷制剂的质量。方法利用标准正态变量变换和二阶导数进行预处理;分别采用连续投影算法、遗传算法和竞争自适应重加权算法对特征波长进行筛选;对筛选后的特征波长建模,并比较模型效果。结果竞争自适应重加权算法筛选后的波长建模效果最好,甘草酸苷、盐酸半胱氨酸、甘氨酸的含量预测模型的测定系数分别为0.9501、0.9522、0.9549,校正标准偏差分别为0.482、0.502、0.544,预测标准偏差分别为0.689、0.732、0.832。预测值可显著解释测量值。结论竞争自适应重加权算法的预测效果较好,结果可靠,可用于快速检测复方甘草酸苷制剂的质量。OBJECTIVE To rapidly detect the quality of Compound glycyrrhizin preparations by Near-infrared spectroscopy and chemometric methods.METHODS This study focuses on Compound glycyrhizin preparations and employs standard normal variate transformation and second-derivative preprocessing.The research involves a combination of techniques,including the successive projection algorithm,the genetic algorithm,and the competitive adaptive reweighted sampling.These methods are applied for feature wavelength selection and modeling,enabling a comparative assessment of model performance.RESULTS The competitive adaptive reweighted sampling demonstrated superior performance in feature wavelength selection,resulting in highly accurate models for the prediction of glycyrrhizin,cysteine hydrochloride,and glycine content.The determination coefficients for these models were 0.9501,0.9522,and 0.9549,while the root mean square error of calibration values were 0.482,0.502,and 0.544,and the root mean square error of prediction values were 0.689,0.732,and 0.832,respectively.External validation results further confirmed the significant explanatory power of the predictive models in relation to the measured values.CONCLUSION The competitive adaptive reweighted sampling demonstrates superior predictive performance,yielding reliable results suitable for the rapid quality assessment of Compound glycyrrhizin preparations.

关 键 词：复方甘草酸苷 近红外光谱 光谱预处理 连续投影算法 遗传算法 竞争自适应重加权算法 质量控制 

分 类 号：R917[医药卫生—药物分析学]