近红外漫反射光谱法测定注射用头孢他啶中头孢他啶、水分和助溶剂精氨酸的含量

A near-infrared diffuse reflectance spectroscopy method for rapid determination of ceftazidime,water content and arginine in ceftazidime for injection

目的:利用近红外漫反射光谱(NIRDRS)分析技术和化学计量学方法对注射用头孢他啶中头孢他啶、水分和助溶剂精氨酸进行无损、快速定量分析。方法:以全国不同企业生产的注射用头孢他啶为分析对象,为了扩大浓度范围,通过减压干燥和定量添加精氨酸或碳酸钠的方法制得了实验室制备样品,用光纤测定近红外漫反射光谱,采用偏最小二乘法(PLS)建立模型。结果:注射用头孢他啶精氨酸定量分析模型由20批样品经内部交叉验证建立,13批样品用于外部验证,浓度范围为4.69%～32.74%,相关系数为0.9948,交叉验证均方差(RMSECV)为0.0109,外部验证均方差为(RMSEP)为0.0106;注射用头孢他啶水分定量模型由21批样品经内部交叉验证建立,21批样品用于外部验证,浓度范围为3.13%～14.70%,相关系数为0.9993,交叉验证均方差(RMSECV)为0.00314,外部验证均方差为(RMSEP)为0.00188;注射用头孢他啶定量模型由25批样品经内部交叉验证建立,22批样品用于外部验证,浓度范围为35.96%～84.59%,相关系数为0.9836,交叉验证均方差(RMSECV)为0.0249,外部验证均方差为(RMSEP)为0.0234。结论:建立的定量模型能对注射用头孢他啶进行准确、快速定量分析,方法可靠简单,可用于药品的现场快速分析。

Objective：To develop a near-infrared diffuse reflectance spectroscopy（NIRDRS） quantitative method for rapid and nondestructive determination of ceftazidime,water content and arginine in ceftazidime for injection.Methods：The diffuse reflectance near-infrared spectra of ceftazidime for injection from different pharmaceutical factories in China were measured intactly by a fiber optic probe.In order to expand the concentration range,we prepared samples by adding Na2CO3 or arginine.The calibration model was developed using partial least squares（PLS） algorithm.Results：For the quantitative model of arginine content,20 batches of ceftazidime for injection samples were used as correlation set,and 13 batches as validation set.The concentration range was 4.69%-32.74%,and the correlation coefficient of cross validation was 0.9948,the root mean square error of cross validation（RMSECV） was 0.0109 and the root mean square error of prediction was 0.0106; For the quantitative model of water content,21 batches of ceftazidime for injection samples were used as correlation set,and 21 batches as validation set.The concentration range was 3.13%-14.70%,and the correlation coefficient of cross validation was 0.9993,the root mean square error of cross validation（RMSECV） was 0.00314 and the root mean square error of prediction was 0.00188;For the quantitative model of ceftazidime content,25 batches of ceftazidime for injection samples were used as correlation set,and 22 batches as validation set.The concentration range was 35.96%-84.59%,and the correlation coefficient of cross validation was 0.9836,the root mean square error of cross validation（RMSECV） was 0.0249 and the root mean square error of prediction was 0.0234.Conclusion：The established NIR method is reliable for predicting the ceftazidime,it is fast and exact and can be used in the mobile lab.