目的 建立一测多评（quantitative analysis of multi-components with a single-marker，QAMS）技术与化学模式识别及加权逼近理想解排序法（technique for order preference by similarity to ideal solution，TOPSIS）模型相结合的润燥止痒胶囊（Runzao Zhiyang Capsules，RZC）综合质量评价方法。方法 以Epic Polar C₁₈柱为色谱柱，乙腈-无水乙醇（80∶20）与0.1%磷酸为流动相进行梯度洗脱，检测波长为320、280、220 nm，二苯乙烯苷为内参物，采用QAMS法同时检测15批RZC中焦地黄苯乙醇苷A₁、毛蕊花糖苷、焦地黄苯乙醇苷B₁、二苯乙烯苷、虎杖苷、桑皮苷A、桑辛素M、二氢桑色素、槐果碱、苦参碱、氧化槐果碱、槐定碱和氧化苦参碱含量。通过对13个成分含量检测数据进行化学模式识别分析，挖掘RZC质量差异性化学成分。同时以化学模式识别分析中变量重要性投影（variable importance projection，VIP）值为权重，建立加权TOPSIS模型，对不同批次RZC进行质量差异性评价。结果 RZC中13个成分在各自范围内线性关系良好（r＞0.999），平均加样回收率为96.97%～100.12%，相对校正因子耐用性良好，外标法实测值与QAMS法计算值无明显差异。化学模式识别显示15批样品聚为3类，氧化苦参碱、二苯乙烯苷、桑皮苷A、苦参碱、氧化槐果碱、虎杖苷和毛蕊花糖苷是RZC质量差异性化学成分。加权TOPSIS模型显示15批RZC欧氏贴近度0.122 0～0.728 0，提示产品质量存在一定的批间差异，质量优劣排序与化学模式识别聚类结果基本一致。结论 建立的QAMS与化学模式识别及加权TOPSIS模型相结合方法，可用于RZC质量差异性评价，为提升RZC质控标准和确保临床疗效一致性提供参考依据。

Objective To establish a comprehensive quality evaluation method for Runzao Zhiyang Capsules (RZC, 润燥止痒胶囊) by quantitative analysis of multi-components with a single-marker (QAMS) combined with chemical pattern recognition and weighted technique for order preference by similarity to ideal solution (TOPSIS) model. Methods HPLC was used with Epic Polar C₁₈ column as chromatographic column. The mobile phase was acetonitrile-anhydrous ethanol (80:20)-0.1% phosphoric acid (gradient elution). The detection wavelengths were 320, 280 and 220 nm. A QAMS method was developed for simultaneous determination of jionoside A₁, verbascoside, jionoside B₁, 2,3,5,4'-tetrahydroxyl-diphenylethylene-2-O-β-D-glucoside, polydatin, mulberroside A, moracin M, dihydromorin, sophocarpine, matrine, oxysophocarpine, sophoridine and oxymatrine in 15 batches of RZC by using stilbene glycoside as internal reference. The chemical pattern recognition was used to analyze the detected content data, and the chemical composition of the quality difference of RZC was excavated. The variable importance for projection value in the chemical pattern recognition analysis was used as the weight, and the weighted TOPSIS model was further used to analyze the detected data, and the quality difference evaluation of different batches of RZC was carried out. Results The 13 components in RZC had good linear relationship in their respective ranges (r > 0.999). The average recovery rate was 96.97%—100.12%. The relative correction factor had good durability. There was no significant difference between the measured value of external standard method and the calculated value of QAMS. Chemical pattern recognition showed that 15 batches of samples were clustered into 3 categories. Oxymatrine, 2,3,5,4'-tetrahydroxystilbene-2-O-β-D-glucoside, mulberroside A, matrine, oxysophocarpine, polydatin and verbascoside were the chemical components with different quality of RZC. The weighted TOPSIS model showed that the euclidean closeness of 15 batches of RZC ranged from 0.122 0 to 0.728 0, suggesting that there were some differences in product quality between batches, and the quality ranking was basically consistent with the results of chemical pattern recognition clustering. Conclusion The established method of QAMS combined with chemical pattern recognition and weighted TOPSIS model can be used for the quality difference evaluation of RZC, and provide reference for improving the quality control standard of RZC and ensuring the consistency of clinical efficacy.

R283.6

陕西省科技计划项目（重点研发计划）（2024SF-YBXM-277）;陕西省科技计划项目（自然科学基础研究计划）（2023-JC-QN-0805）