USING DATA MINING APPROACHES TO SELECT ACUPOINTS IN ACUPUNCTURE AND MOXIBUSTION FOR KNEE OSTEOARTHRITIS

Background: Acupuncture and moxibustion are traditional Chinese medicine therapies commonly used to treat knee osteoarthritis (KOA). Although acupoint selection affects the effectiveness of acupuncture and moxibustion, the basic rules of acupoint selection are little understood and there is a lack of guidelines regarding prescription. In this study, we used data mining approaches to investigate the principles of acupoint selection and provide a framework for formulation prescription in acupuncture and moxibustion for clinical treatment of KOA. Materials and Methods: PubMed, Cochrane Library, Science Citation Index, Wanfang database, VIP database, and China National Knowledge Infrastructure were searched for randomized controlled clinical trials published in English or Chinese from January 1, 2009 to October 1, 2015 evaluating the effect of acupuncture and moxibustion on KOA. Databases were established. Frequency statistics and association rule were used to extract and analyze the data. Results: A total of 876 acupuncture prescriptions and 122 acupoints were included in the analysis. Acupoints were concentrated in acupoints of fourteen meridians. The most frequently used acupoints were Dubi (ST35), Neixiyan (EX-LE4), Yanglingquan (GB34), Xuehai (SP10), Liangqiu (ST34), Zusanli (ST36), Yinlingquan (SP9), and Ashi point. The most frequently used meridian was Stomach Meridian of Foot-Yangming. Acupoints were concentrated mainly in the lower limbs. 42 acupoint pairs occurred frequently, and the top acupoint pairing was Dubi (ST35) and Neixiyan (EX-LE4). Conclusion: Acupoint selection and formulation prescription should focus on locally affected areas, and follow the theory of meridians, which helps establish guidelines for acupuncture and moxibustion in KOA patients.