Objective: To explore the characteristics and rules of acupoint sensitization phenomena based on knee osteoarthritis (KOA), one of the clinical dominant diseases of acupuncture-moxibustion.
Methods: In combination with literature and expert experiences, the acupoints with the highest use frequency in treatment of KOA were screened, e.g. Heding (EX-LE 2), Liangqiu (ST 34), Mingmen (GV 4), Neixiyan (EX-LE 4), Ququan (LR 8) and Dubi (ST 35). In 814 patients with KOA and 217 healthy subjects, the acupoint temperature, mechanic pain threshold and pressure pain threshold were detected separately. Using machine learning method, the sensitization was judged at each acupoint.
Results: Compared with healthy subjects, the acupoint temperature was increased and the mechanic pain threshold and pressure pain threshold were reduced in KOA patients (P<0.05). Besides, the cut-off value was presented to distinguish whether the acupoint was sensitized or not. The results of machine learning showed that the highest prediction accuracy of acupoint sensitization was 86.7% (Shenshu [BL 23]) and the lowest one was 73.9% (Heding [EX LE 2]). The prediction accuracy at the third clinical stage trial was higher, the highest was 93.3% (Ququan [LR 8]) in KOA patients.
Conclusion: It is confirmed that the acupoint sensitization reflects the characteristics of disease and is correlative with the conditions of illness, which may provide the reference for the auxiliary diagnosis and condition assessment of KOA.
目的：基于针灸临床优势病种膝关节骨关节炎（KOA）探究穴位敏化现象的特征与规律。方法：结合文献和专家经验，筛选出治疗KOA使用频率最高的穴位，包括：鹤顶、梁丘、命门、内膝眼、曲泉、犊鼻等，对814例KOA患者和217例健康受试者的穴位温度、机械痛阈、压痛阈变化进行检测，采用机器学习的方法判别穴位是否发生敏化。结果：与健康受试者比较，KOA患者穴位温度升高、机械痛阈和压痛阈降低（P<0.05），且存在区分穴位是否发生敏化现象的界值。机器学习结果显示，穴位敏化预测准确率最高为86.7%（肾俞），最低为73.9%（鹤顶）；对KOA患者临床3期预测准确率较高，最高为93.3%（曲泉）。结论：本研究证实了穴位敏化现象反映疾病的特征，且与疾病的病情存在关联，可为KOA辅助诊断和病情评估提供参考。.
Keywords: acupoint sensitization; acupoint temperature; knee osteoarthritis; machine learning; mechanic pain threshold; pressure pain threshold.