Chinese herbal medicine for constipation: zheng-based associations among herbs, formulae, proprietary medicines, and herb–drug interactions

The review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We used the following databases to search the conventional medicine literature: PubMed, Ovid, Evidence-Based Medicine Reviews (EBMR), and Embase. The following databases were used to search the traditional CM literature: SinoMed, Chinese National Knowledge Infrastructure (CNKI), Chinese biomedical literature (CBM) CD, and China Journals Full-text Database. From the electronic database records and bibliographic references, we identified relevant primary sources and secondary sources (such as textbooks, review articles, and meta-analyses) as follows. We selected all EBM reviews, including Cochrane DSR, ACP Journal Club, DARE, CCTR, CMR, HTA, and NHSEED from inception to April 2014; EMBASE from 1980 to April 2014; EMBASE Classic from 1947 to 1979; PubMed from inception to April 2014; Ovid MEDLINE(R) from 1950 to April 2014; Ovid OLDMEDLINE(R) from 1948 to 1965; SinoMed from 1978 to April 2014; China Journals Full-text Database from 1994 to April 2014 and CBM disc from 1979 to April 2014. The search strategy was (1) (constipation) OR (chronic constipation) OR (functional constipation); (2) (herb*) OR (herbal medicine) OR (traditional Chinese medicine) OR (Chinese medicine) OR (Complementary medicine) OR (Naturopathy); (3) (case*) OR (clinical observation*) OR (clinical trial) OR (clinical study); (1) AND (2) AND (3) (*was used for truncation).

We included interventions using single herbs, CM classical formulae, CM new formulae, and Chinese herb-derived products and combination products. The clinical study types included were quasi- or randomized controlled trials, observational clinical studies, case series or case reports, and other types of appropriate research methods. We included studies on subjects with constipation of all ages and both sexes, studies using objective measures from laboratory or imaging techniques, and studies using measurement from nursing staff, patients, or other informants.

Two authors (LLDZ and GZ) independently searched the databases and selected relevant publications. If the two authors disagreed about a study’s eligibility, they would check the study against the selection criteria, discuss its eligibility, and make a further decision (ZXB). One author (LLDZ) extracted the data and the other (GZ) checked the extracted data. For each study, the following information was extracted: study design, sample size, mode of recruitment, sampling and diagnostic procedure, inclusion and exclusion criteria, and participants’ characteristics (including age, sex, and duration of constipation). CM patterns, CM treatment principles, treatment regimen, and CM treatment outcomes were recorded.

The methodological quality of relevant studies was assessed using the Jadad scale (Additional file 1; [11]). The Jadad scale evaluates a study in terms of the description of randomization, blinding, and dropouts. The scale ranges from 1 to 5; randomized controlled trials with a score between 3 and 5 are regarded as better quality trials. Points were awarded if the study was described as randomized (1 point), had an appropriate randomization method (1 point), was described as double-blind (1 point), used an appropriate blinding method (1 point), and had a description of withdrawals and dropouts (1 point) [11].

Classification based on keyword co-occurrence was conducted on the 18,272 items of constipation literature downloaded from SinoMed [12]. We also applied a dictionary-based data slicing algorithm constructed on the principle of keyword co-occurrence. We filtered the downloaded data using CM associated keywords, such as “Chinese herbal medicine,” “Chinese patent medicine,” and “CM syndrome/zheng,” which were obtained from textbooks and the Internet. The keyword co-occurrence classification was a good supplement to the literature search, as it provided insights into the quantitative relationship between the individual herbs and formulae used to treat constipation. We used a wheel-shaped network to indicate the association between different types of single herbs and their zheng indications [10]. The wheel-shaped network was a visualized graph that showed the frequencies and correlations among the same categories [10].

All Chinese-to-English translations were deduced primarily from the World Health Organization (WHO) Evidence-Based Complementary and Alternative Medicine International Standard Terminologies on Traditional Medicine in the Western Pacific Region [13].

From the included articles, we retrieved 190 different CM zheng diagnoses from 485 individual studies. The most common pattern was dual deficiency of qi and blood (Qi Xue Liang Xu), which was diagnosed in 948 of the 15,740 subjects (frequency = 48, percentage among the top 10 diagnosis = 16.8 %); this was followed by dual deficiency of qi and yin (Qi Yin Liang Xu) (subjects = 795, frequency = 45, percentage among the top 10 diagnosis = 15.7 %), excessive heat and qi stagnation (Qi Zhi Shi Re) (subjects = 726, frequency = 41, percentage among the top 10 diagnosis = 14.3 %), yang deficiency of spleen and kidney (Pi Shen Yang Xu) (subjects = 636, frequency = 32, percentage among the top 10 diagnosis = 11.2 %), deficiency of qi and blood (Qi Xue Liang Xu) (subjects = 595, frequency = 26, percentage among the top 10 diagnosis = 9.1 %), cold–heat complex (Han Re Cuo Za) (subjects = 513, frequency = 21, percentage among the top 10 diagnosis = 7.3 %), liver depression and spleen deficiency (Gan Yu Pi Xu) (subjects = 495, frequency = 20, percentage among the top 10 diagnosis = 7.0 %), deficiency–excess complex (Xu Shi Jia Za) (subjects = 483, frequency = 20, percentage among the top 10 diagnosis = 7.0 %), dual yin deficiency of liver and kidney (Gan Shen Yin Xu) (subjects = 410, frequency = 17, percentage among the top 10 diagnosis = 5.9 %) and intestinal dryness and yin deficiency (Yin Xu Chang Zao) (subjects = 223, frequency = 16, percentage among the top 10 diagnosis = 5.6 %). Subjects diagnosed with the top 10 CM zheng accounted for 37 % of the 15,740 subjects (Table 1). Table 1 lists the therapeutic principles for the CM zheng.

We identified 296 herbs from 485 clinical studies and analyzed their relationships using the wheel-shaped network (Fig. 2). In this figure, red and green nodes represent different single herbal medicines. The edges represent co-occurrence frequency in clinical studies. The edge label numbers represent the number of clinical studies demonstrating a connection between two single herbal medicines. There was a high concentration of one classical Chinese herbal formula, Ma-Zi-Ren-Wan, whose composition of six herbs occupied 42.5 % (1754/4127) of the total frequency of usage. Based on this, the green nodes represent CHMs in the formula Ma-Zi-Ren-Wan and the red nodes represent other CHMs used in clinical prescriptions [10]. Node size was calculated with the formula

where node_frequency is the literature record number of the associated keyword calculated in text mining, LOG is the logarithm calculation with a base set to 10, and Degree (node) is the number of connections/edges each node has with other nodes. The edge line width was calculated through LOG (edge_co-occurrent) where edge_co-occurrent was the number of node/keyword pairs that co-occurred in the associated literature.

Table 2 lists the top 10 most frequently used herbs and their actions. The most frequently used herbs were further categorized and analyzed to determine their combinations in classical formulae. The most frequently used classical formulae were Ma-Zi-Ren-Wan [14] and Zeng-Ye-Tang [15].

Among the 289 studies on Chinese herbal formulae, the most frequently used formulae based on zheng diagnosis were Ma-Zi-Ren-Wan and its modifications (frequency among all the studies, n = 75, percentage among the top 10 formulae, 33.07 %). This was followed by Bu-Zhong-Yi-Qi-Tang (n = 56, 22.58 %), Ji-Chuan-Jian (n = 51, 20.56 %), Zeng-Ye-Tang (n = 40, 13.84 %), and Ba-Zhen-Tang (n = 26, 10.48 %). The five most frequently used CM zheng-based Chinese herbal formulae and their indications are summarized in Table 3.

After the classical herbal decoctions, the next most frequently used clinical treatments for constipation were proprietary CHMs, because of their standard quality control and more convenient administration. We analyzed the most commonly used proprietary CHMs and their dosage (Table 4). The manufacturers of the proprietary CHMs are also shown, to indicate the quality and composition of the medicines. The most commonly used proprietary CHM was Run-Chang-Wan (frequency, n = 87), followed by Ma-Ren-Ruan-Jiao-Nang (n = 62), Ma-Ren-Run-Chang-Wan (n = 52), Liu-Wei-An-Xiao-Jiao-Nang (n = 50), Fu-Fang-Lu-Hui-Jiao-Nang (n = 35), Si-Ni-San (n = 32), Liu-Wei-Neng-Xiao-Jiao-Nang (n = 27), and Bu-Zhong-Yi-Qi-Wan (n = 17).

Few clinical studies reported the combined use of CM and Western medicine for constipation. Table 5 summarized the herb–drug interactions identified. Although most studies did not mention the adverse effects associated with combined administration of herbs and drugs, or proprietary CHMs and drugs, there were 43 reported trials of the integrative use of single herbs or proprietary CHMs. Among these studies, the most frequent adjunctive use was Da Huang (Radix et Rhizoma Rhei) with sodium bicarbonate tablets (n = 23), followed by Fan Xie Ye (Folium Sennae) with lactulose oral solution (n = 8), Ma-Ren-Ruan-Jiao-Nang with lactulose oral solution (n = 6) and Liu-Wei-An-Xiao-Jiao-Nang with mosapride citrate tablets (n = 6).