目的 运用主成分分析（PCA）法和聚类分析法，研究秦艽不同配伍药对对风湿热痹类风湿关节炎（RA）大鼠的影响，探讨病证-药性-疗效之间的关系。方法 80只SD大鼠随机分为对照组、模型（CⅡ）组、病证模型组、雷公藤多苷组、单味秦艽组、秦艽+威灵仙（秦-威）组、秦艽+桑寄生（秦-桑）组和秦艽+防己（秦-防）组，采用天然牛Ⅱ型胶原（CⅡ）诱导加风湿热环境因素刺激制备风湿热痹RA大鼠模型，造模结束后各给药组分别ig给予制备的对应药液（25 mg/kg）。实验中测量大鼠体质量、足跖肿胀度、关节炎指数评分、痛阈值、冷痛耐受值及压力耐痛值，酶联免疫法测定各组大鼠血清中类风湿因子（RF）、C-反应蛋白（CRP）、肿瘤坏死因子-α（TNF-α）、白细胞介素-1β（IL-1β）、前列腺素E₂（PGE₂）水平。运用SPSS 21.0软件对数据进行PCA和聚类分析，设定主成分特征值大于1，聚类分析采用系统聚类法。结果 PCA筛选出2个主成分，二者携带原变量信息的89.6%，第1主成分反映除体质量之外其他变量的信息，第2主成分主要反映体质量这一变量的信息；以主成分因子得分与其权重（方差贡献）乘积之和相加，得出各组的总因子得分（F值，F值越大，病情越严重），病证模型组F值最大，对照组F值最小，给药组中秦-防组F值明显小于秦-威组、秦-桑组。聚类分析将8个组按对照组、模型组、给药组聚为3类，与理论实际相符合；给药组中秦-防组单独聚为一类，其余给药组聚为一类。结论 对于风湿热痹型RA，平寒相配疗效优于平温、平平相配，实验结果符合中医临床“疗热以寒药”的治疗原则；PCA法和聚类分析法操作简单、结果可靠，可以用于评价动物模型及药物疗效。

Objective To study the effects of Gentianae Macrophyllae Radix in cold and heat compatibility in rheumatic fever arthralgia rheumatoid arthritis (RA) model rats by principal component analysis and cluster analysis. Methods Eighty SD rats were randomly divided into blank group, collage Ⅱ model group, rheumatic fever arthralgia model group, tripterygium group, single Qinjiao group, Qinjiao-Weilingxian (Clematidis Radix et Rhizoma) group, Qinjiao-Sangjisheng (Taxilli Herba) group, and Qinjiao-Fangji (Stephaniae Tetrandrae Radix) group; rheumatic fever arthralgia model were induced by collage Ⅱ and being exposed in rheumatic fever environment, each administration group was gavaged with 15 mL/kg corresponding drug solution after modeling. The weight, paw thickness, arthritis index, pain threshold, cold pain tolerance time and pressure pain tolerance value were measured. The serum contents of RF, CRP, TNF-α, IL-1β, and PGE₂ were detected by ELISA. Results Principal component analysis selected two principal components, which carried 89.6% of the original information, the first principal component reflected all information in expect weight, the second principal component mainly reflected the weight. The total factor score F (the greater the F value, the more severe the disease) was obtained by the sum of the principal component factor score and its weight (variance contribution) product. The F value of rheumatic fever arthralgia model group was the highest and the score of the blank group was the lowest, the F value of Qin-Fang group was significantly lower than those of Qin-Wei group and Qin-Sang group. Cluster analysis divided eight groups into three categories according to the blank, model and administration, which was consistent with the theory; In the control group, Qin-Fang group was clustered into one group, and the other groups were clustered into one group. Conclusion For the rheumatic fever arthralgia RA, the combined effect of mild and cold traditional Chinese medicine (TCM) is better than the combined effect of mild and warm TCM and the combined effect of mild TCMs. Experimental results are consistent with the principles of TCM clinical treatment of "treating the hot diseases should use the cold medicine"; Principal component analysis and cluster analysis are simple and reliable, and can be used to evaluate the animal model and drug efficacy.

国家自然科学基金资助项目（81360648）