偏最小二乘法用于近红外漫反射光谱快速测定异福片  

Rapid determination of rifampicin and isoniazide tablets using near infrared diffuse reflectance spectroscopy

作　　者：逯家辉[1] 陈渝飞[1] 蒋朝军[1] 石健[1] 郭伟良[1] 滕利荣[1] 

机构地区：[1]吉林大学生命科学学院,长春130012

出　　处：《药物分析杂志》2007年第10期1518-1522,共5页Chinese Journal of Pharmaceutical Analysis

摘　　要：目的:采用近红外光谱技术,建立一种快速测定异福片中利福平和异烟肼含量的方法。方法:将光谱分别进行卷积平滑、一阶导数、二阶导数预处理,应用偏最小二乘法(PLS)建立定量分析模型,以校正集的交互验证均方根误差(RMSECV)及相关系数(r)为优化参数,选择最佳预处理方法和波长范围。通过留一交互验证法,以预测残差平方和(PRESS)为优化参数,选择最适主因子数。结果:校正集样品利福平和异烟肼的 r 分别为0.9940和0.9910,RMSECV 分别为0.00764和0.00445。对预测集样品利福平和异烟肼的预测均方根误差(RMSEP)分别为0.00397和0.00383;平均加样回收率分别为100.88%和100.87%;重复性试验 RSD 分别为0.672%和1.08%。结论:结果表明该方法预测精度高,且具有方便快捷、非破坏、无污染、可在线检测、重复性好等优点,可作为异福片原位质量检测和在线质量监控的方法予以推广。Objective：To establish a rapid method to determination of rifampicin and isoniazide in rifampicin and isoniazide tablets by near infrared technology. Methods： NIR spectra were pretreated by Savitzky - Golay, 1 st order derivative and 2nd order derivative respectively;and then were used to construct quantitative analysis models using PLS. Results：In the calibration models, the regression coefficient （r） in rifampicin determination was 0. 9940, the root mean square error of cross - calibration （RMSECV） was 0. 00764, the r in isoniazide determination was 0. 9910, and the RMSECV was 0. 00445. The samples of validation set were predicted by these models. The root mean square error of prediction （RMSEP） in rifampicin determination was 0. 00397. The RMSEP in isoniazide de- termination was 0. 00383. This result demonstrates that the method has high precision of prediction. The average re- coveries of rifampicin and isoniazide are 100. 88% and 100. 87% respectively, and the RSDs were 0. 672%, 1.08% in reappearance experiments. Conclusions ： With the advantages of facility, rapid, non - destruction, no pol- lution, on- line detection and good reappearance, the method must be popularized in situ measurement and in the on -line quality control for rifampicin and isoniazide tablets productions.

关 键 词：近红外漫反射光谱 偏最小二乘法 利福平 异烟肼 

分 类 号：R917[医药卫生—药物分析学]