近红外光谱法同时快速测定五种抗结核药物的主成分含量

Application of Near Infrared Spectroscopy in Rapid and Simultaneous Determination of Essential Components in Five Varieties of Anti-Tuberculosis Tablets

20世纪80年代以来,全球的结核病形势急剧恶化,利福平片、异烟肼片、吡嗪酰胺片、异福片和异福酰胺片是目前比较有效的抗结核药物。文章采用近红外光谱技术（NIRS）结合偏最小二乘法（PLS）建立了同时测定上述五种抗结核药物中的利福平（RMP）、异烟肼（INH）和吡嗪酰胺（PZA）含量的定量分析模型。实验结果表明,采用原始光谱建立模型测定以上抗结核药物中各组分含量时,测定RMP含量的最有效的波长范围为1981～2195nm,测定INH含量的最有效波长范围分别是1540～1717nm和2086～2197nm,测定PZA含量的最有效波长范围分别是1 460～1 537nm,1956～2022nm和2268～2393nm,最优模型测定RMP,INH和PZA含量的交互验证均方根误差（RMSECV）分别为0.0494,0.0257和0.0307,模型对预测集样品中的RMP,INH和PZA含量预测,预测均方根误差（RMSEP）分别为0.0349,0.0166和0.0134,预测相关系数（rp）分别达到0.9864,0.9989和0.9993,表明模型预测准确可靠,该研究对抗结核药物生产中质量检验和在线检测有重要的意义。

Since 1980s, tuberculosis has become increasingly serious. Rifampicin tablets, isoniazide tablets, pyrazinamide tablets, rifampicin and isoniazide tablets and rifampicin isoniazide and pyrazinamide tablets are currently relatively efficacious antituberculosis drugs. In the present paper, near infrared spectroscopy （NIRS） with partial least squares （PLS） was applied to the simultaneous determination of rifampicin （RMP）, isoniazide （INH） and pyrazinamide （PZA） contents in 5 varieties of anti-tuberculosis tablets. As the results showed, all of the models for the determination of RMP, INH and PZA contents applied the original NIR spectra. The most efficacious wavelength range for the determination of RMP contents was 1 981-2 195 nm, it was 1 540-1 717 nm and 2 086-2 197 nm for the determination of IblH contents, and it was 1 460-1 537 nm, 1 956-2 022 nm and 2 268-2 393 nm for determination of PZA contents. The root mean square error of the calibration set obtained by cross-validation （RMSECV） of the optimum models for the quantitative analysis of RMP, INH and PZA contents was 0. 049 4, 0. 025 7 and 0. 030 7, respectively. Using these optimum models for the determination of RMP, INH and PZA contents in prediction set, the root mean square error of prediction set （RMSEP） was 0. 018 2, 0. 016 6 and 0. 013 4, respectively. The correlation coefficient （rp） between the predicted values and actual values was 0. 986 4, 0. 998 9 and 0. 999 3, respectively. These results demonstrated that this method was precise and reliable, and is significative for in situ measurement and the on-line quality control for anti-tuberculosis tablets production.