目的 建立黄芪丹参煎液的HPLC指纹图谱及多指标成分定量分析方法，为黄芪丹参煎液质量控制提供科学依据。方法 采用HPLC-UV法，ODS Hypersil DIM色谱柱（250 mm×4.6 mm，3 μm）；以乙腈-0.1%甲酸溶液为流动相，梯度洗脱，柱温25℃，体积流量0.8 mL/min，进样量20 μL，检测波长254 nm，建立HPLC指纹图谱；采用LC-MS法，建立黄芪甲苷、黄芪皂苷III、黄芪皂苷II、毛蕊异黄酮葡萄糖苷、毛蕊异黄酮、芒柄花苷、丹酚酸B、原儿茶醛、咖啡酸、迷迭香酸、紫草酸11个成分含量测定方法，并采用主成分分析（PCA）和正交偏最小二乘判别分析（OPLS-DA）对测定结果进行识别。结果 10批黄芪丹参煎液的指纹图谱中有12个共有峰，通过与化学对照品比对，指认出7个色谱峰，分别为1号峰咖啡酸、3号峰毛蕊异黄酮葡萄糖苷、6号峰迷迭香酸、8号峰紫草酸、9号峰毛蕊异黄酮、10号峰丹酚酸B、11号峰芒柄花苷，10批样品的相似度≥ 0.995。黄芪甲苷、黄芪皂苷III、黄芪皂苷II、毛蕊异黄酮葡萄糖苷、毛蕊异黄酮、芒柄花苷、丹酚酸B、原儿茶醛、咖啡酸、迷迭香酸、紫草酸线性关系良好，r²值均大于0.995；回收率为89.3%~110.9%。通过PCA可将10批黄芪丹参煎液分成2类，并可看出不同产地、不同批次药材之间的质量存在差异，进一步采用OPLS-DA筛选出导致质量差异的两个物质丹酚酸B和芒柄花苷。结论 根据黄芪丹参煎液指纹图谱的建立，结合质谱进行定量，能够为其质量控制提供更可靠的参考。

Objective To establish the Huangqi-Danshen Decoction (HDD) fingerprint and multicomponent quantitative method, and provide a scientific basis for the quality standard of HDD. Methods The HPLC fingerprint was performed on ODS Hypersil DIM column with acetonitrile-0.1% formic acid as mobile phase at gradient elution. The column temperature was 25℃; The flow rate was 0.8 mL/min, the volume was 20 μL, and the detection wavelength was 254 nm. At the same time, LC-MS was used to establish content determination method for 11 constitutes:astragaloside, astragaloside III, astragaloside II, calycosin 7-O-beta-D-glucoside, calycosin, ononin, salvianolic acid B, protocatechuic aldehyde, caffeic acid, rosmarinic acid, and lithospermic acid; The samples were identified by PCA analysis and OLPS-DA analysis. Results There were 12 common peaks in the fingerprints of 10 batches of HDD. By comparing with the chemical reference, seven peaks were confirmed, they were:caffeic acid (peak 1), calycosin 7-O-beta-D-glucoside (peak 3), rosmarinic acid (peak 6), lithospermic acid (peak 8), calycosin (peak 9), salvianolic acid B (peak 10), and ononin (peak 11); Ten batches of samples have a high similarity (no less than 0.995). Astragaloside, astragaloside III, astragaloside II, calycosin 7-O-beta-D-glucoside, calycosin, ononin, salvianolic acid B, protocatechuic aldehyde, caffeic acid, rosmarinic acid, and lithospermic acid had good linearity with r² ≥ 0.995 and recovery was at range of 89.3%-110.9%. Through principal component analysis, 10 batches of HDD can be divided into two categories. It can be seen that there are differences in the quality between different origins and different batches of medicinal materials. Finally, OPLS-DA was used to screen out two substances that cause quality differences:salvianolic acid B and ononin. Conclusion According to the establishment of the fingerprint of HDD and the quantification by LC-MS, it can provide a reliable reference for its quality control.

国家自然科学基金项目（81603437）；国家自然科学基金项目（81804052）；深圳市科技计划项目（ZDSYS201606081515458）；深圳市科技计划项目（JCYJ20170307154652899）