近红外在线检测丹红注射液醇沉过程中多指标成分的含量变化

Online Monitor Multi-component Content of Danhong Injection in the Alcohol Precipitation Process Using Near-infrared Spectroscopy

目的运用近红外光谱分析技术(NIR)在线检测丹红注射液醇沉过程中多指标成分的含量变化,包括丹参素、原儿茶醛、迷迭香酸、丹酚酸B和乙醇浓度。方法以HPLC和气相色谱法分别测定4种指标成分和乙醇浓度的含量,作为参考值,并与NIR光谱数据关联,运用偏最小二乘回归(PLSR)法建立NIR定量校正模型,并将模型用于在线检测丹红注射液醇沉过程。结果模型测量校正集样本的丹参素、原儿茶醛、迷迭香酸、丹酚酸B和乙醇浓度的相对分析误差(RPD)分别为5.72%,7.46%,8.44%,4.41%和5.50%,相关系数(R2)分别为0.969 5,0.982 0,0.985 9,0.948 4和0.967 0,验证集相对偏差RSEP分别为5.35%,7.59%,6.42%,4.51%和2.30%。结论 NIR光谱分析技术可作为一种丹红注射液大生产中醇沉过程的在线分析方法。

OBJECTIVE To measure the changeable contents of multi-indicators in Danhong injection alcohol precipitation process, including danshensu, protocatechualdehyde, rosmarinic acid, salvianolic acid B and ethanol concentration by online near-infrared（NIR） spectroscopy analysis technique. METHODS HPLC and gas chromatography acquired the data of former four index components and ethanol concentration, respectively. Partial least squares regression（PLSR） was introduced to establish the multi-component models for online monitoring alcohol precipitation process of Danhong injection. RESULTS The residual predictive deviation（RPD） values of danshensu, protocatechualdehyde, rosmarinic acid, salvianolic acid B and ethanol concentration models were 5.72%, 7.46%, 8.44%, 4.41% and 5.50%, respectively; the values of correlation coefficient（R2） were 0.969 5, 0.982 0, 0.985 9, 0.948 4 and 0.967 0, respectively; the relative standart error（RSEP） values of validation sets were 5.35%, 7.59%, 6.42%, 4.51% and 2.30%, respectively. CONCLUSION NIR can be used as an online analytical method in Danhong injection alcohol precipitation process.