Abstract: Background　Chronic pulmonary heart disease（PHD） is a common respiratory disease，which is the final stage of some chronic lung diseases. Chinese medicines have proved to be effective in treating PHD. However，due to complexity of PHD，the TCM syndrome differentiation for it is not unified，and the prescriptions are various，so it is of great significance to explore the characteristics of prescribing Chinese medicines and the rules of combination use of Chinese medicines for treating PHD. Objective　To explore the rules of prescribing Chinese medicines and the rules of combination use of these medicines for PHD using data mining technique. Methods　Studies about PHD treated with Chinese medicines published during 2000—2020 were searched in databases of CNKI，Wanfang Data，VIP，and SinoMed，and screened according to the inclusion and exclusion criteria. Data were extracted and input into a standardized database created from an Excel spreadsheet. Data mining technique was used to statistically analyze the use frequency，properties，taste，channel tropism，and effectiveness of Chinese medicines，rules of combination use and cluster analysis of core combinations for complex conditions. Results　A total of 166 articles were included，involving 197 prescriptions，and 202 herbs，with cumulative frequency of drug use of 2 944 times. The most frequently used medicines were Poria cocos （60.41%），Glycyrrhiza uralensis （59.39%），Astragalus membranaceus （58.84%），Salvia miltiorrhiza （49.23%），Lepidium seed （44.67%），Pinellia ternata （41.12%），and Atractylodes macrocephala Koidz （40.10%）. The commonly used categories classified by physiologic effect were tonifying deficiency （25.79%），antitussive and antiasthmatic （15.35%），promoting blood circulation and removing blood stasis （13.71%），expectorant （9.28%），and promoting diuresis for eliminating dampness （8.23%）. The drug properties were mainly warm （38.40%） and mild （22.73%）. The three primary tastes were bitter （31.84%），sweet （30.13%） and pungent （28.11%）. The three primary meridian tropisms of drugs were lung meridian（27.47%），spleen meridian（18.52%） and heart meridian（16.15%）. Nineteen dual combination therapies，227 triple combination therapies，and 160 quadruple combination therapies for pulmonary heart disease were identified，which mainly belong to the combination use of tonifying qi drugs，antitussive，antiasthmatic and expectorant drugs，and qi-regulating and spleen-strengthening drugs. Seven core prescriptions were excavated by cluster analysis. Conclusion　The mechanism of Chinese medicines treating PHD is mainly based on tonifying lung functions，strengthening spleen and cardiac functions，supplemented with antitussive，antiasthmatic，expectorant，blood stasis removing，and eliminating the pathogenic qi according to syndrome differentiation analysis. The commonly used drugs and rules of drug combinations for PHD obtained by data mining may contribute to clinical treatment of PHD using Chinese medicines，the development of new preparations，and drug combinations for PHD.

Key words: Pulmonary heart disease, Traditional Chinese medicine, Data mining, Medication rules

摘要： 背景　慢性肺源性心脏病是一种呼吸系统常见疾病，是某些慢性肺部疾病的终末阶段。中医药治疗肺源性心脏病的有效性已被证实，但该病的复杂性导致中医辨证分型不统一，方药复杂多变，因此进一步挖掘肺源性心脏病用药规律及组方特点意义重大。目的　基于数据挖掘方法探讨现代中医药治疗肺源性心脏病用药规律和组方特点。方法　计算机检索中国知网（CNKI）、万方数据知识服务平台（Wanfang Data）、维普网（VIP）、中国生物医学文献服务系统（SinoMed）中关于中医药治疗肺源性心脏病的研究，检索时限为2000—2020年。严格筛选纳入符合标准的文献，建立Excel规范化数据库，采用数据挖掘技术对药物的使用、性味归经及功效进行频次统计分析、关联规则分析和系统聚类分析。结果　共纳入文献166篇，包括197首方剂药物，202味中药，累计药物使用频次2 944次。用药频次较高的药物为茯苓（60.41%）、甘草（59.39%）、黄芪（58.84%）、丹参（49.23%）、葶苈子（44.67%）、半夏（41.12%）、白术（41.12%），常用的药物类型为补虚药（25.79%）、止咳平喘药（15.35%）、活血化瘀药（13.71%）、化痰药（9.28%）、利水渗湿药（8.23%），药性以温（38.40%）、平（22.73%）为主，药味以苦（31.84%）、甘（30.13%）、辛（28.11%）为主，药物归经以肺经（27.47%）、脾经（18.52%）、心经（16.15%）为主。经关联规则分析，得出药对关联规则19条，3味中药配伍关联规则227条，4味中药配伍关联规则160条，基本以补气药，止咳平喘化痰药和理气健脾药物之间相配伍。聚类分析挖掘出7个核心组方。结论　中医药治疗肺源性心脏病主要以补肺健脾强心为主，辨证予以止咳、平喘、化痰、祛瘀、祛邪等。数据挖掘得出的用药和组方规律对临床中医药辨证治疗肺源性心脏病具有一定的参考价值，为新药开发和新方组合提供思路和借鉴。

关键词: 肺心病, 中医药, 数据挖掘, 用药规律