Component analysis of Chinese medicine and advances in fuming-washing therapy for knee osteoarthritis via unsupervised data mining methods Academic research paper on "Medical engineering"

JTCM

Online Submissions:http://www.journaltcm.com info@journaltcm.com

JTradit Chin Med 2013 October 15; 33(5): 686-691 ISSN 0255-2922 © 2013 JTCM. All rights reserved.

INFORMATION AND EDUCATION

Component analysis of Chinese medicine and advances in fum-ing-washing therapy for knee osteoarthritis via unsupervised data mining methods

Jun Liu, Jianke Pan,Yanping Wang, Dingkun Lin, Dan Shen, Hongjun Yang, Xiang Li, Minghui Luo, Xuewei Cao

Jun Liu, Dingkun Lin, Xiang Li, Minghui Luo, Xuewei Cao,

Department of Orthopedics, the Second Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou 510120, China

Jianke Pan, Doctoral Degree Candidate, the Second School of Clinic Medicine, Guangzhou University of Chinese Medicine, Guangzhou 510405, China

Yanping Wang, Institute of Basic Research in Clinical Medicine, China Academy of Chinese Medical Sciences, Beijing 100700, China

Dan Shen, Hongjun Yang, Institute of Chinese Materia Medica, China Academy of Chinese Medical Sciences, Beijing 100700, China

Supported by Grant from the Administration of Traditional Chinese Medicine of Guangdong Province in China (No. 20131161) and the Specialized Research Fund for the Doctoral Program of Higher Education of China (No. 20124425110004)

Correspondence to: Prof. Xuewei Cao, Department of Orthopedics, the Second Affiliated Hospital of Guangzhou University of Chinese Medicine, Guangzhou 510120, China. ca-oxuewei@126.com Telephone: +86-15913145030 Accepted: June 20,2013

Abstract

OBJECTIVE: To analyze the component law of Chinese medicines in fuming-washing therapy for knee osteoarthritis (KOA), and develop new fum-ing-washing prescriptions for KOA through unsupervised data mining methods.

METHODS: Chinese medicine recipes for fum-ing-washing therapy for KOA were collected and recorded in a database. The correlation coefficient among herbs, core combinations of herbs, and

new prescriptions were analyzed using modified mutual information, complex system entropy cluster, and unsupervised hierarchical clustering, respectively.

RESULTS: Based on analysis of 345 Chinese medicine recipes for fuming-washing therapy, 68 herbs occurred frequently, 33 herb pairs occurred frequently, and 12 core combinations were found. Five new fuming-washing recipes for KOA were developed.

CONCLUSION: Chinese medicines for fum-ing-washing therapy of KOA mainly consist of wind-dampness-dispelling and cold-dispersing herbs, blood-activating and stasis-resolving herbs, and wind-dampness-dispelling and heat-clearing herbs. The treatment of fuming-washing therapy for KOA also includes dispelling wind-dampness and dispersing cold, activating blood and resolving stasis, and dispelling wind-dampness and clearing heat. Zhenzhutougucao (Herba Speranskiae Tuber-culatae), Honghua (Flos Carthami), Niuxi (Radix Achyranthis Bidentatae), Shenjincao (Herba Lycopo-dii Japonici), Weilingxian (Radix et Rhizoma Clemati-dis Chinensis), Chuanwu (Radix Aconiti), Haitongpi (Cortex Erythrinae Variegatae), Ruxiang (Olibanum), Danggui (Radix Angelicae Sinensis), Caowu (Radix Aconiti Kusnezoffii), Moyao (Myrrha), and Aiye (Folium Artemisiae Argyi) are the main herbs used in the fuming-washing treatment for KOA.

© 2013 JTCM. All rights reserved.

Key words: Knee osteoarthritis; Fuming-washing therapy; Drug prescriptions; Data mining

INTRODUCTION

To raise awareness for bone and joint diseases, the World Health Organization declared the first 10 years of the 21st century as the "Bone and Joint Decade." As one of the most common joint diseases, knee osteoarthritis (KOA) seriously influences patient quality of life and healthcare expenses. KOA is a worldwide health concern and is one of the foremost causes of chronic disability in seniors.1'2 As the world's population ages, this problem will spread and place a large burden on societies and health care systems. Traditional Chinese Medicine (TCM) has a variety of therapeutic methods available to treat KOA. Among them, fuming-washing therapy is the most commonly used therapeutic method and has a definite curative effect.3 Chinese medicine used for KOA can show how much experience a TCM practitioner has for treating KOA. Using data mining techniques to search for component laws, it can be possible to remove the limitation of dependence on personal experience in the clinic. Unsupervised techniques in data mining can find the core combinations of drugs from recipe data according to characteristics of the variables with no artificial interference. Because the recipe data are discrete, multi-factored, sparse, and non-linear, we used unsupervised data mining methods including modified mutual information methods, complex system entropy clusters, and unsupervised hierarchical clustering, to find a component law of Chinese medicine recipes and develop new recipes.4-8 In this study, the above-mentioned methods are used to analyze Chinese medicine fuming-washing recipes for treating KOA, so as to provide references for the clinical treatment of KOA.

MATERIALS AND METHODS

Source of recipes

A computer-based online search of relevant Chinese articles between 1990 and 2012 was conducted in the Chinese National Knowledge Infrastructure, Wanfang database, and VIP database. Meanwhile, relevant English articles between 1990 and 2012 were searched in PubMed with the key words: "knee osteoarthritis", "fuming" and "washing".

Screening of recipes

KOA is a degenerative disease and is an impediment disease in TCM, because of its pathogenic characteristics and clinical manifestations. Therefore, the Chinese medicines for fuming-washing therapy for indications of knee joint swelling, pain, numbness, aching, joint stiffness, joint deformity, and muscular atrophy were selected from the recipes. Recipes for knee trauma, tuberculosis, tumor, rheumatoid arthritis, and infectious diseases were excluded. Repetitive recipes were also deleted.

Construction of database

Microsoft Net framework 3.5 was used as the develop-

ment platform, and a computer program made Chinese herb names in the recipes binary variables. Therefore, clicking a herb from a certain recipe gave it a value defined as " 1" while the values of other herbs were automatically expressed as "0". From 1 English article and 344 Chinese articles, 345 Chinese herb recipes were collected. There were 213 commonly used Chinese herbs, and a corresponding database of fum-ing-washing recipes for KOA was constructed.

Analysis methods

The data were analyzed through modified mutual information methods that we developed. We used the complex system entropy cluster and unsupervised hierarchical clustering methods, found a quantitative description of herb correlation coefficient, extracted core combinations, and found new recipes.4-8 New recipes refer to those achieved using data mining, and those that did not exist in previous databases. For the mining methods, the core combinations of Chinese herbs for KOA were acquired by a complex system entropy cluster. Then, the core combinations were mined by unsu-pervised hierarchical clustering, which generated new recipes. The principles of the complex system entropy cluster and hierarchical clustering are briefly introduced as follows:

Complex system entropy cluster: the core formula of the modified mutual information method and the complex system entropy cluster method is as follows:8

H(Xi)-H(X/)-H(X,,Xj) Po(jJ) ^

(Xi,Xj)=

H(Xi)+H(Xj) -2H(Xi,Xj) H(Xj)

Po(iJ)< 5.

Here, Po (i, j) expresses the positive occurrence frequency of the two variables Xi and Xj, and 8 is a threshold. Selecting an appropriate threshold can not only separate the positive correlation and the negative correlation, but also avoid interfering with erroneous data. Hierarchical clustering: the core combinations of Chinese herbs, which occur in the original database and in certain compound recipes but cannot be called a new recipe, were attained by hierarchical clustering. This is based on the hypothesis that a new recipe is the recombination of a strong correlative core combination. Hierarchical clustering and correlation calculations were necessary, and both were involved in calculating the correlation between classes. If the combination attained by hierarchical clustering did not exist in the original database, and the mutual information needed to be improved, then the recipe is new.

Here, Pro (X, Y) >0 expresses the frequency of core combination X and Y. A frequency >0 means that it occurs in a certain compound prescription, so it is not a new recipe. Thus, it is necessary to decrease the correla-tivity, making the correlativity of the core combination in the new recipe reach the maximum. H (X) and H (Y) represent information entropy of X and Y, respectively. H (X, Y) expresses the combination entropy.

RESULTS

Higher use frequency herbs

A total of 68 herbs with a high use frequency (>20) in Chinese medicines for fuming-washing therapy for treatment of KOA were found by analyzing and screening 245 recipes in the original database (Table 1).

able 1 Frequently used herbs

Herb i

Use frequency

Use frequency

Zhenzhutougucao (Herba Speranskiae Tuberculatae) 342

Honghua (Flos Carthami) 329

Niuxi (Radix Achyranthis Bidentatae) 319

Shenjincao (Herba Lycopodii) 315

Weilingxian (Radix et Rhizoma Clematidis Chinensis) 307

Chuanwu (Radix Aconiti) 298

Haitongpi (Cortex Erythrinae Variegatae) 281

Ruxiang (Olibanum) 277

Danggui (Radix Angelicae Sinensis) 265

Caowu (Radix Aconiti Kusnezoffii) 257

Moyao (Myrrha) 243

Aiye (Folium Artemisiae Argyi) 243

Guizhi (Ramulus Cinnamomi) 243

Duhuo (Radix Angelicae Biserratae) 215

Huajiao (Pericarpium Zanthoxyli Bungeani) 201

Fangfeng (Radix SaposhniKOAviae) 192

Chuanxiong (Rhizoma Chuanxiong) 166

Mugua (Fructus Chaenomelis Speciosae) 157

Sumu (Lignum Sappan) 156

Wujiapi (Cortex Acanthopanacis Radicis) 149

Jixueteng (Caulis Spatholobi) 146

Xixin (Herba Asari Mandshurici) 140

Qianghuo (Rhizoma Et Radix Notopterygii) 122

Cu(Vinager) 109

Sangzhi (Ramulus Mori) 108

Baizhi (Radix Angelicae Formosanae) 106

Chishao (Radix Paeoniae Rubra) 100

Qinjiao (Radix Gentianae Macrophyllae) 86

Dahuang (Radix Et Rhizoma Rhei Palmati) 83

Sangjisheng (Herba Taxilli Chinensis) 78

Gusuibu (Rhizoma Drynariae) 77

Taoren (Semen Persicae) 73

Xuduan (Radix Dipsaci) 71

Liujinucao (Herba Artemisiae Anomalae) 65

Danshen (Radix Salviae Miltiorrhizae) 65

Kuanjinteng (Caulis Tinosporae Sinensis) 61

Gancao (Radix Glycyrrhizae) 59

Huangbo (Cortex Phellodendri) 58

Sanleng (Rhizoma Sparganii) 56

Cangzhu (Rhizoma Atractylodis) 54

Zelan (Herba Lycopi Hirti) 53

Yanhusuo (Rhizoma Corydalis Yanhusuo) 49

Ezhu (Rhizoma Curcumae Phaeocaulis) 49

Haifengteng (Caulis Piperis Kadsurae) 47

Jingjie (Herba Schizonepetae Tenuifoliae) 44

Qiannianjian (Rhizoma Homalomenae) 42

Dilong (Pheretima Aspergillum) 41

Duzhong (Cortex Eucommiae) 41

Jianghuang (Rhizoma Curcumae Longae) 40

Tubiechong (Eupolyphaga) 39

Baishao (Radix Paeoniae Alba) 39

Jiu (Liquor) 38

Fuling (Poria) 35

Mahuang (Herba Ephedrae Sinicae) 34

Yiyiren (Semen Coicis) 29

Fangji (Radix Stephaniae Tetrandrae) 29

Zexie (Rhizoma Alismatis) 27

Tufuling (Rhizoma Smilacis Glabrae) 26

Xuejie (Sanguis Draconis) 25

Ganjiang (Rhizoma Zingiberis) 25

Luoshiteng (Caulis Trachelospermi ^ Jasminoidis)

Muxiang (Radix Aucklandiae) 24

Liangmianzhen (Radix Zanthoxyli) 23

Rougui (Cortex Cinnamomi Cassiae) 22

Gouji (Rhizoma Cibotii) 21

Fuzi (Radix Aconiti Lateralis Preparata) 21

Maqianzi (Semen Strychni) 21

Yan (Salt) 20

Commonly used herb pairs

Thirty-three commonly used herb pairs with correlation coefficients over 0.015 were generated by analysis with the modified mutual information method (Table 2).

Core combinations

Twelve core combinations were found using the complex system entropy cluster method (Table 3).

New recipes

Five new recipes capable preventing and treating KOA

Table 2 Commonly used herb pairs

were developed by the unsupervised hierarchical clustering method (Table 4).

DISCUSSION

The aim of the study was to find the component law of Chinese medicines in fuming-washing therapy for KOA, and develop new fuming-washing prescriptions for KOA through unsupervised data mining methods. From the frequency analysis, we found that the main

Herb pair Correlation coefficient

Niuxi (Radix Achyranthis Bidentatae), Sangjisheng (Herba Taxilli) 0.037 723

Niuxi (Radix Achyranthis Bidentatae), Taoren (Semen Persicae) 0.033 588

Niuxi (Radix Achyranthis Bidentatae), Gusuibu (Rhizoma Drynariae) 0.031 851

Niuxi (Radix Achyranthis Bidentatae), Zhenzhutougucao (Herba Speranskiae Tuberculatae) 0.031 507

Niuxi (Radix Achyranthis Bidentatae) , Ruxiang (Olibanum) 0.026 259

Haitongpi (Cortex Erythrinae Variegatae), Sumu (Lignum Sappan) 0.021 813

Honghua (Flos Carthami), Moyao (Myrrha) 0.020 972

Duhuo (Radix Angelicae Pubescentis), Chuanxiong (Rhizoma Chuanxiong) 0.020 092

Ruxiang (Olibanum), Honghua (Flos Carthami) 0.019 995

Aiye (Folium Artemisiae Argyi), Caowu (Radix Aconiti Kusnezoffii) 0.019 721

Niuxi (Radix Achyranthis Bidentatae), Moyao (Myrrha) 0.019 238

Niuxi (Radix Achyranthis Bidentatae), Haitongpi (Cortex Erythrinae Variegatae) 0.018 815

Duhuo (Radix Angelicae Pubescentis), Duzhong (Corthex Eucommiae) 0.018 757

Caowu (Radix Aconiti Kusnezoffii), Moyao (Myrrha) 0.018 655

Weilingxian (Radix et Rhizoma Clematidis Chinensis), Mugua (Fructus Chaenomelis) 0.018 646

Danggui (Radix Angelicae Sinensis), Weilingxian (Radix et Rhizoma Clematidis Chinensis) 0.018 626

Aiye (Folium Artemisiae Argyi), Niuxi (Radix Achyranthis Bidentatae) 0.018 232

Ruxiang (Olibanum), Baizhi (Radix Angelicae Dahuricae) 0.018 176

Danggui (Radix Angelicae Sinensis), Jixueteng (Caulis Spatholobi) 0.018 139

Danggui (Radix Angelicae Sinensis), Shengdihuang (Radix Rehmanniae Recen) 0.018 030

Weilingxian (Radix et Rhizoma Clematidis Chinensis), Sumu (Lignum Sappan) 0.017 979

Huajiao (Pericarpium Zanthoxyli), Danggui (Radix Angelicae Sinensis) 0.017 849

Chuanwu (Radix Aconiti), Honghua (Flos Carthami) 0.017 723

Duhuo (Radix Angelicae Pubescentis), Chuanwu (Radix Aconiti) 0.017 125

Duhuo (Radix Angelicae Pubescentis), Caowu (Radix Aconiti Kusnezoffii) 0.017 064

Chuanwu (Radix Aconiti), Dilong (Pheretima) 0.016 808

Moyao (Myrrha), Sumu (Lignum Sappan) 0.016 670

Danggui (Radix Angelicae Sinensis), Gancao (Radix Glycyrrhizae) 0.016 143

Duhuo (Radix Angelicae Pubescentis), Fangfeng (Radix Saposhniviae) 0.015 595

Sangjisheng (Herba Taxilli), Fuling (Poria) 0.015 554

Huajiao (Pericarpium Zanthoxyli), Baizhi (Radix Angelicae Dahuricae) 0.015 463

Niuxi (Radix Achyranthis Bidentatae), Guizhi (Ramulus Cinnamomi) 0.015 380

Danggui (Radix Angelicae Sinensis), Sumu (Lignum Sappan) 0.015 000

Table 3 Core combinations

No Core combination

1 Ruxiang (Olibanum), Moyao (Myrrha), Zhenzhutougucao (Herba Speranskiae Tuberculatae)

2 Ruxiang (Olibanum), Moyao (Myrrha), Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Carthami), Shen-jincao (Herba Lycopodii), Niuxi (Radix Achyranthis Bidentatae)

3 Honghua (Flos Carthami), Niuxi (Radix Achyranthis Bidentatae) , Danggui (Radix Angelicae Sinensis)

4 Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Carthami), Shenjincao (Herba Lycopodii), Aiye (Folium Artemisiae Argyi)

5 Ruxiang (Olibanum), Moyao (Myrrha), Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Carthami), Shenjincao (Herba Lycopodii), Aiye (Folium Artemisiae Argyi), Niuxi (Radix Achyranthis Bidentatae), Haitongpi (Cortex Erythrinae Variegatae), Chuanwu (Radix Aconiti), Caowu (Radix Aconiti Kusnezoffii), Weilingxian (Radix et Rhizoma Clematidis Chinen-sis)

6 Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Carthami), Shenjincao (Herba Lycopodii), Haitongpi (Cortex Erythrinae Variegatae)

7 Ruxiang (Olibanum), Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Carthami), Shenjincao (Herba Lycopodii), Aiye (Folium Artemisiae Argyi), Niuxi (Radix Achyranthis Bidentatae), Danggui (Radix Angelicae Sinensis), Haitongpi (Cortex Erythrinae Variegatae), Chuanwu (Radix Aconiti), Weilingxian (Radix et Rhizoma Clematidis Chinensis)

8 Ruxiang (Olibanum), Honghua (Flos Carthami), Zhenzhutougucao (Herba Speranskiae Tuberculatae), Danggui (Radix Angelicae Sinensis), Shenjincao (Herba Lycopodii), Niuxi (Radix Achyranthis Bidentatae), Chuanwu (Radix Aconiti), Caowu (Radix Aconiti Kusnezoffii), Weilingxian (Radix et Rhizoma Clematidis Chinensis)

9 Ruxiang (Olibanum), Honghua (Flos Carthami), Zhenzhutougucao (Herba Speranskiae Tuberculatae), Shenjincao (Herba Lycopodii), Aiye (Folium Artemisiae Argyi), Niuxi (Radix Achyranthis Bidentatae), Haitongpi (Cortex Erythrinae Variegatae), Caowu (Radix Aconiti Kusnezofßi), Weilingxian (Radix et Rhizoma Clematidis Chinensis)

10 Zhenzhutougucao (Herba Speranskiae Tuberculatae), Niuxi (Radix Achyranthis Bidentatae), Chuanwu (Radix Aconiti), Caowu (Radix Aconiti Kusnezofßi)

11 Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Carthami), Niuxi (Radix Achyranthis Bidentatae), Shenjincao (Herba Lycopodii), Chuanwu (Radix Aconiti), Caowu (Radix Aconiti Kusnezoffii)

12 Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Carthami), Niuxi (Radix Achyranthis Bidentatae), Shenjincao (Herba Lycopodii), Weilingxian (Radix et Rhizoma Clematidis Chinensis)

|Table 4 New recipes

No New recipe

1 Dahuang (Radix Et Rhizoma Rhei), Huashi (Talcum), Huanglian (Rhizoma Coptidis), Bixie (Rhizoma Dioscoreae Septemlobae), Cheqianzi (Semen Plantaginis), Wushaoshe (Zaocys), Gouji (Rhizoma Cibotii), Dilong (Pheretima), Chuanxiong (Rhizoma Chuanxiong)

2 Guanghuoxiang (Herba Pogostemonis), Pugongying (Herba Taraxaci), Yuxingcao (Herba Houttuyniae), Zihuadiding (Herba Vi-olae), Sangbaipi (Cortex Mori), Buguzhi (Fructus Psoraleae), Xiangfu (Rhizoma Cyperi), Tubiechong (Eupolyphaga Seu Steleoph-aga)

3 Danggui (Radix Angelicae Sinensis), Duhuo (Radix Angelicae Pubescentis), Sangjisheng (Herba Taxilli), Wujiapi (Cortex Acan-thopanacis), Qianghuo (Rhizoma Et Radix Notopterygii), Chuanxiong (Rhizoma Chuanxiong), Aiye (Folium Artemisiae Argyi), Haitongpi (Cortex Erythrinae Variegatae), Honghua (Flos Carthami), Zhenzhutougucao (Herba Speranskiae Tuberculatae), Shenjincao (Herba Lycopodii Japonici)

4 Kuanjinteng (Tinospora Sinensis Merr), Gouteng (Ramulus Uncariae Cum Uncis), Wangbuliuxing (Semen Vaccariae), Rudijin-niu (Shinyleaf Pricklyash Root), Dahuang (Radix Et Rhizoma Rhei), Haifengteng (Caulis Piperis Kadsurae), Jianghuang (Rhizoma Curcumae Longae)

5 Ruxiang (Olibanum), Moyao (Myrrha), Chuanwu (Radix Aconiti), Caowu (Radix Aconiti Kusnezoffii), Guizhi (Ramulus Cin-namomi), Xixin (Herba Asari), Zhenzhutougucao (Herba Speranskiae Tuberculatae), Haitongpi (Cortex Erythrinae Variegatae), Weilingxian (Radix et Rhizoma Clematidis Chinensis)

Chinese medicines mainly consist of wind-dampness-dispelling and cold-dispersing herbs, blood-activating and stasis-resolving herbs, and wind-dampness-dispelling and heat-clearing herbs. Therefore, the treatment principle is to eliminate pathogenic factors. Eliminating the pathogenic factors includes dispelling wind-dampness and dispersing cold, activating blood and resolving stasis, and dispelling wind-dampness and clearing heat. The most frequently used herbs that eliminate pathogenic factors are: Zhenzhutougucao (Herba Speranskiae Tuberculatae), Honghua (Flos Car-

thami), Niuxi (Radix Achyranthis Bidentatae), Shenjincao (Herba Lycopodii Japonici), Weilingxian (Radix et Rhizoma Clematidis Chinensis), Chuanwu (Radix Aconiti), Haitongpi (Cortex Erythrinae Variegatae), Ruxiang (Olibanum), Danggui (Radix Angelicae Sinensis), Caowu (Radix Aconiti Kusnezoffii), Moyao (Myrrha), and Aiye (Folium Artemisiae Argyi). This phenomenon corresponds to modern study of syndromes. Liu et al9 found that "blood stasis," "cold," and "exogenous damp" were the common pathogenic factors. Wind-prevailing migratory arthralgia, cold-prevailing

agonizing arthralgia, dampness-prevailing fixed arthralgia, and heat-obstructing arthralgia were the basic patterns of osteoarthritis that frequently appeared in the study. In addition, study on osteoarthritis patients found that cold-dampness, Qi stagnation and blood stasis, and moist heat arthralgia spasms are common.10 These twelve herbs are considered the main herbs for treating KOA with fuming-washing therapy. Herb pairs are used jointly in formulas or prescriptions to produce desired therapeutic effects and reduce toxic or side effects. Herb pairs are the simplest and most common form of TCM compound recipe. By analyzing the internal relations among the 33 commonly used herb pairs, we found that the attributes of these herb pairs include mutual reinforcement, assistance, restraint, and suppression, which are in accordance with TCM clinical practice. Such as Niuxi (Radix Achyran-this Bidentatae) and Sangjisheng (Herba Taxilli), Duhuo (Radix Angelicae Pubescentis) and Chuanxiong (Rhizoma Chuanxiong), Duhuo (Radix Angelicae Pubescentis) and Fangfeng (Radix Saposhniviae), Danggui (Radix Angelicae Sinensis) and Jixueteng (Caulis Spatholobi), all of these herb pairs have the function of reinforcement or assistance, whereas Danggui (Radix Angelicae Sinensis) and Shengdihuang (Radix Rehmanni-ae Recen) have the function of restraint or suppression. From the herb compositions of the five recipes, recipes 1 and 2 clear heat and drain dampness, activate blood to relieve pain, and dispel wind to free the collateral vessels. Therefore, they are suitable for patients with retained damp heat toxin. Recipes 3, 4, and 5 dissipate cold and dispel dampness, activate blood to relieve pain, and dispel wind to free the collateral vessels. Therefore, they are suitable for patients with dampness-prevailing fixed arthralgia. These five new recipes conform to the theories of Chinese medicine in the prevention and treatment of KOA.

This study summarized the universal rule of Chinese medicines in fuming-washing therapy for KOA. In the future, we will try to further analyze the recipes using data mining methods, and new recipes will be tested in animal experiments or clinical trials to verify their efficacy.

REFERENCES

1 Woolf AD, Pfleger B. Burden of major musculoskeletal conditions. Bull World Health Organ 2003; 81(9): 646-656.

2 Guccione AA, Felson DT, Anderson JJ, et al. The effects of specific medical conditions on the functional limitations of elders in the framingham study. Am J Public Health 1994; 84(3): 351-358.

3 Ou ZX, Jin JC, Huang D, et al. Comparative study on effects of combined massage-smouldering-washing therapy and mini-invasive surgery in treating knee osteoarthritis. Zhong Guo Zhong Xi Yi Jie He Za Zhi 2008; 28(10): 925-928.

4 Yang HJ, Zhao YL, Tang SH, et al. Analysis of correlativi-ty between herbs in recipes for treatment of apoplexy by entropy method. Zhong Guo Zhong Yi Ji Chu Yi Xue Za Zhi 2005; 11(9): 706-709.

5 Tang SH, Chen JX, Yang HJ, et al. Research thinking of discovery of new chinese medicine recipes by complex system entropy cluster method. Shi Jie Ke Xue Ji Shu — ZhongYi YaoXian Dai Hua 2009; 11(2): 225-228.

6 Chen JX. Establishment of complex system model of chinese medicine syndromes and research on its correlativity with diseases. Graduate University of Chinese Academy of Sciences, Doctoral Dissertation 2008: 7.

7 Yang HJ, Chen JX, Tang SH, et al. New herb R&D of Traditional Chinese Medicine — role of data mining approaches. J Biol Syst 2009; 17(3): 329-347.

8 Tang SH, Chen JX, Li G, et al. Research on component law of Chinese patent medicine for anti-influenza and development of new recipes for anti-influenza by unsuper-vised data mining methods. J Tradit Chin Med 2010; 30 (4): 288-293.

9 LiuXQ, Yao GH, Lu M. Periodical literature about patho-genesis factor of knee osteoarthritis etiology and syndrome type analyzed. Jiangsu Zhong Yi Yao 2005; 26(10): 63-64.

10 Liu XQ, Yao GH, Li JB. A retrospective analysis of knee osteoarthritis of TCM syndrome diagnosis in Chinese medicine hospital medical records. Hunan Zhong Yi Xue Yuan Xue Bao 2004; 24(5): 30-32.