目的 首次提出质量综合评价指数（F_q）的概念，并以当归Angelicae sinensis药材为例，通过多项检测指标的整合分析，建立一种药材品质综合评价与等级划分的方法。方法 采用《中国药典》2020年版方法对当归药材中总灰分（x₁）、酸不溶性灰分（x₂）、水分（x₃）、浸出物（x₄）、挥发油（x₅）和阿魏酸（x₆）6项指标进行测定。利用近红外光谱技术（NIRS）构建上述检测指标与光谱之间的拟合模型，并开展多批次样品的系统检测建立数据库。通过定义F_q为各项指标的含量（x_i）与其多批次均值（X_i）偏离度（K_i）的加合，并基于其对药材品质的贡献引入权重分配系数（W_i），通过正交优化确立综合评价公式。结果 所建立的上述6项指标的近红外预测模型效果良好，并应用其完成了对200份当归样本的测定，建立了数据样本库。各检测指标经偏离度计算及权重分配优化，最终确定当归药材的综合评价指数公式。经统计所收集当归样品的F_q值呈高斯分布，按照正态分布概率将样品平均分为5个等级区域，其不合格样品与优质样本可以得到明显区分，等级评价效果较好。结论 该方法基于现行的《中国药典》监管体系实现了对当归药材品质的快速综合评价与等级区分，为中药材的科学监管研究提供了新的解决方案。

Objective Based on the complexity of Chinese medicinal ingredients, and the fragmentation of quality evaluation and indexes for Chinese medicinal materials, the concept of quality comprehensive evaluation index factor (F_q) was proposed for the first time in this paper. Taking Angelicae Sinensis Radix (ASR) as an example, through the integration analysis of several testing indexes, an integrated evaluation method of quality evaluation and a grade identification system were developed. Methods According to the method of Chinese Pharmacopoeia (2020) for ASR, six indicators including total ash content (x₁), acid insoluble ash (x₂), water (x₃), extract (x₄), naphtha (x₅), and ferulic acid (x₆) were measured. In addition, the fitting model between the near infrared spectroscopy (NIRS) and the content with detection index was established respectively. Then a novel database was established for the systematic quality evaluating of multiple batches of samples. F_q value was defined as the sum of the deviation (K_i), which indicated each testing index (x_i) and the mean (X_i) of multi-batch. In addition, based on its contribution to the quality of medicinal materials, the weight allocation coefficient (W_i) was introduced and optimized by orthogonal optimization, and the comprehensive evaluation formula was finally established. Results The results showed that the NIR prediction model of the above six indexes had a good desired effect, and then applied to detect of 200 of batches ASR medicinal materials. After deviation degree calculation and weight allocation optimization of each index, the comprehensive evaluation index formula of ASR was finally determined. According to the statistics, the distribution of F_q values of multi-batch samples was gaussian distribution. By the normal distribution probability, the samples were divided into five grade regions on average, among which the unqualified samples and superior samples could be precisely controlled, and the grade classification effect was satisfactory. Conclusion Based on the current pharmacopoeial supervision system, this method can quickly evaluate the quality of ASR, carry out grade appraisal, and provide a new solution for the scientific supervision of Chinese medicinal materials.

R286

天津市科技援助项目（18YFYZCG00060）