Pharmacological treatments of Chinese herbal medicine for irritable bowel syndrome in adults: A network meta-analysis of randomized controlled trials

Introduction Plenty of clinical studies have suggested the value of Chinese herbal medicine (CHM) for patients with irritable bowel syndrome (IBS), but their efficacy and safety have not been systematically concluded yet. This article aimed to compare and rank the therapeutic effect and safety of CHM with routine pharmacotherapies and placebo in the treatment of IBS. Methods Randomized controlled trials regarding CHM to treat IBS were searched in six databases from inception to Jan 31, 2020. A network meta-analysis was conducted to analyze the data of included publications. The quality assessment was assessed by Cochrane Handbook and GRADEpro software. The risk ratio was calculated for dichotomous outcomes while the standardized mean difference was used for continuous variables with 95% credible intervals. A Funnel plot was performed to evaluate publication bias. The surface under the cumulative ranking curve was conducted to rank the included interventions. Data were analyzed with STATA 15.0 and Review Manager 5.3. Result 3194 records were searched, and 28 eligible trials involving 3323 patients ere identified. Compared with conventional therapies and placebo, Jianpi-Chushi therapy showed significant improvement in adequate relief and IBS symptom severity scale; Shugan-Jianpi therapy showed the best efficacy in relieving the abdominal pain and abdominal distension; Wenshen-Jianpi therapy had a better effect on avoiding adverse effects and improving stool character. Conclusion This study confirmed that CHM could be beneficial for patients with IBS in relieving their clinical symptoms and should be recommended as alternative therapies. The quality of evidence in this study based on the GRADE system was “low”.


Introduction
Irritable bowel disease (IBS) is one of the most common chronic digestive disorders in the world, which is characterized by abdominal pain and discomfort, defecation as well as change in stool consistency and frequency [1]. According to epidemiological research, the incidence ranges from 19.58%~23.40% in China and 10~25% in North America and Europe [2,3]. According to the Rome IV criteria [4], IBS can be presented as 4 pattern subtypes: IBS with diarrhea (IBS-D), IBS with constipation (IBS-C), mixed IBS (IBS-M), and unclassified IBS (IBS-U).
The main pathogenesis of IBS has been conceptualized as a condition of visceral hypersensitivity (leading to abdominal discomfort or pain) [5], and gastrointestinal motor disturbances (leading to diarrhea or constipation) [6]. In addition, there is increasing evidence regarding the roles of mood and anxiety disorders, infection and immune activation, serotonin dysregulation, bacterial overgrowth, central dysregulation as well as brain-gut interaction, family genetics in the etiology of IBS [7][8][9][10][11][12]. Due to the diversity of pathogenesis, the main pharmacological treatments of IBS such as smooth-muscle relaxants, prokinetic agents, peripheral opioid agonist, antidiarrheal, antidepressants, and probiotics, can only achieve limited clinical benefits [13,14], and some of them may even cause a risk of cardiovascular events in long term use [15]. Therefore, it is necessary to look for more effective and safer alternative therapies.
Traditional Chinese medicine (TCM) has been used to treat symptoms associated with IBS for thousands of years in East Asia and may offer insights into a more targeted approach for therapeutic development [16]. Plenty of previous studies have evaluated the efficacy and safety of CHM (Chinese herbal medicine) formulae in the treatment of IBS-C and IBS-D [17][18][19][20], but these studies focused on pairwise comparisons between single formula and conventional medicines, and no comparison with different CHM formulae was conducted in the treatment of IBS on a large scale.
Therefore, a Bayesian network meta-analysis (NMA) which integrates direct evidence with indirect for multiple intervention comparisons was performed to compare and rank different CHM formulae with routine pharmacotherapies in the management of clinical symptoms in patients with IBS.

Methods
This study was performed in conformity to the Cochrane Handbook for the Systematic Review of Interventions and the Preferred Reporting Items for Systematic Review and Meta-Analyses [21]. The completed PRISMA checklist was presented as S1 File.

Data extraction and quality assessment
Two investigators independently selected the studies. The review of the selected studies, the extraction of the relevant information, and the assessment of the risk of bias tool were performed by two investigators. Relevant information was extracted from each included study: Study ID (first author and publication year), classification of disease and diagnostic criteria, the characteristics of participants (gender, age, and sample size), the course of disease, detailed of interventions (treatment and duration), primary outcomes (adequate relief, improvement of irritable bowel syndrome-severity scoring system (IBS-SSS)) and secondary outcomes (adverse effects, improvement of clinical symptoms). Any missing information will be acquired by contacting the corresponding author. The access to the included trials was displayed in S3 File.
The risk of bias of the included studies was evaluated with the Cochrane Collaboration Recommendations assessment tool [22]. Seven domains were assessed as low-risk, high-risk, or unclear-risk including random sequence generation, allocation concealment, blinding of participants and personnel, blinding (or) masking of outcomes assessors, incomplete outcome data, selective reporting, and other bias. Besides, the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) was performed to assess the quality of evidence as high, moderate, low, or very low quality.

Statistical analysis
An NMA with a Bayesian framework using Software for Statistics and Data Science (STATA, version 15.1) was conducted to assess the outcomes of different interventions. For continuous variables (IBS-SSS and the improvement of clinical symptoms), standardized mean differences (SMD) were calculated with a 95% confidence interval (CI). For dichotomous data (adequate relief, adverse effects), risk ratios (RR) were calculated with a 95% CI. Considering the diversity of interventions and potential heterogeneity among included studies, a random-effect model was applied in all meta-analyses. The consistency test results were judged by node-splitting analysis and an inconsistency model. When the p-value of the node-splitting analysis was greater than 0.05, a consistency mode was selected [23]; otherwise, an inconsistency model was used. Heterogeneity analysis was assessed through inconsistency index statistic (I 2 ). The I 2 value above 50% was considered as heterogeneity throughout the study. Additionally, sensitivity analysis was conducted to verify the robustness of the results and test the source of heterogeneity in each RCT. To summarize the probabilities for all interventions, the surface under the cumulative ranking curve (SUCRA) was selected to offer a summary statistic for the cumulative ranking [24]. Based on the definition, the larger SUCRA scores are, the more effective interventions are.

Study identification and selection
In total, 3194 citations (PubMed 43, Spring 643, EMBASE 73, CNKI 459, CBD 1933, WanFang 43) published from inception to January 31, 2020, were identified by the search. After removing duplicates and unrelated articles, 28 articles comprising 3323 patients were deemed eligible for further quantitative analyses . A flow diagram of the specific screening procedures is shown in Fig 1. The baseline characteristics of the studies were extracted in Table 2. The frequency of utilization of the included herbs is summarized in Fig 2 while the components of each formula are summarized in Table 3.

Quality assessment of included studies
We evaluated the quality of included studies with the Cochrane Collaboration Recommendations assessment tools [53]. All of the studies (28/28) described a random component in the sequence generation process such as a computer-generated random number or a random number table. Allocation concealment was performed using an appropriately sealed method in 17.9% (5/28) of the studies, while 82.1% (23/28) either did not describe concrete methods or used an inappropriate allocation concealment method. In performance bias, 35.7% (10/28) of the included trials reported the methods of blinding for both participants and personnel. In detection bias, 64.3% (18/28) of the outcome assessors in the studies either could not be blinded or were unclear. In attrition bias, all of the studies were deemed to have low-risk outcome data (i.e., reported drop out rates within the range of statistical estimations, provided detailed explanations of drop out rates, or performed intention-to-treat analysis). A detailed quality assessment is presented in Fig 3. (RR 1.62, 95% CI 1.30 to 2.02) and the differences were statistically significant. The efficacy of WJ (RR 1.70, 95% CI 1.39 to 2.17) and SJ (RR 1.52, 95% CI 1.30 to 1.78) were also better than placebo and rank 2 nd , 3 rd among all the therapies. The treatments were ranked as follow according to the SUCRA ( Fig 4B): JC > WJ > SJ > pinaverium > trimebutine > probiotics > antidiarrheal > placebo. The heterogeneity in Fig  Irritable bowel syndrome-severity scoring system (IBS-SSS). The improvement of IBS-SSS was reported in 10 studies with 7 treatments. The specific network is presented in Fig  7A. It is revealed from Table 5 that JC was better than trimebutine (SMD 2.93, 95% CI 1.16 to 7.39), antidiarrheal (SMD 14.01, 95% CI 3.42 to 57.50) in the improvement of IBS-SSS. The efficacy of SJ (SMD 8.25, 95% CI 3.34 to 20.35) and WJ (SMD 6.86, 95% CI 2.13 to 22.12) were also better than antidiarrheal. The SUCRA is presented in Fig 7B and  character, WJ was better than antidiarrheal (SMD 2.79, 95% CI 1.01 to 7.11) and trimebutine Adverse effects. There were 26 studies with 8 treatments that reported adverse effects. The most common side effects in the treatment groups were nausea and vomiting, constipation, and slight elevation of liver aminotransferases while abdominal pain and distension, nausea, and flatulence in the controlled groups. The specific network was presented in Fig 8D. The result in Table 7 indicated that there were no significant statistical differences among all the    therapies. Based on the SUCRA in Fig 9D, the treatments are ranked as follow: WJ > placebo > Pinaverium > JC > SJ > Probiotics > Trimebutine > Antidiarrheal. GRADE quality evidence. The application of the GRADE approach aims to provide ratings for the confidence in the estimates of effect for specific comparison [54]. There are five elements to downgrade the quality of evidence: risk of bias, inconsistency, indirectness, imprecision, and publication bias while three factors to upgrade: large effect, plausible confounding that would change effect, dose-response gradient. Based on these criteria, the evidence quality

Discussion
The treatments of IBS are largely based on its subtypes [1,55]. In this study, there were 26 trials focus on IBS-D [25, 27-39, 41-54], 1 trial on IBS-C [40], and 1 trial on IBS (contained both IBS-D and IBS-C) [26]. Limited by the treatments of the controlled group, we could only compare TCM with placebo, pinaverium, trimebutine, antidiarrheal, and probiotics. Pinaverium, trimebutine, and probiotics are universal therapies for all types of IBS in relieving abdominal pain while antidiarrheal suits patients with IBS-D.  This NMA systematically evaluated the AR, improvement of IBS-SSS, the improvement of clinical symptoms, and adverse effects after the application of CHM as compared to conventional pharmacological therapies for patients with IBS. In patients with IBS-D, JC performed the best in AR and the improvement of IBS-SSS compared with placebo and any other pharmacological treatments. WJ showed great improvement in improving stool character. SJ had better effects on relieving abdominal pain and abdominal distension. Similarity, in patients with IBS-C, JC also was more effective on adequate relief and in improving stool consistency compared to placebo [40]. There was no difference between CHM and other therapies in adverse effects. In conclusion, CHM could be more beneficial to patients with IBS in decreasing their clinical symptoms and improving their quality of life, which provided more suggestions and guidance in clinical decisions.

PLOS ONE
Pharmacological treatments of Chinese herbal medicine for irritable bowel syndrome in adults As is mentioned before, the etiologies of IBS are diverse, so it is hard to treat it from one single dimension. The core principle of diagnosing and treating disease in TCM is a treatment based on syndrome differentiation (TCM jargon: bian zheng lun zhi") [56]. According to TCM theory, syndrome (TCM jargon: Zheng) is a presentation of the pathological changes of a certain disease course, revealing the location, cause, and nature of a disease, the correlation between pathogenic factors and health factors, and the body's ability to resist disease, and thus is a precondition and fundamental for diagnosis and treatments [57]. Under the principle of "bian zheng lun zhi", the CHM formulae, composed of many different herbs, take the basic prescription as the core and add or delete some drugs on the condition of patients' symptoms. Therefore, the effective substance of CHM formulae is multi-component, and its functions are multi-target, multi-pathway, and multi-effects. A review study involved 67784 IBS participants found out that the major syndromes of IBS patients were the syndrome of liver stagnation and spleen deficiency, spleen-stomach weakness, and spleen-kidney yang deficiency [58].

PLOS ONE
Pharmacological treatments of Chinese herbal medicine for irritable bowel syndrome in adults Therefore, based on the syndrome differentiation, the treatment of CHM formulae was concluded as Shugan-Jianpi therapy, Jianpi-Chushi therapy, and Wenshen Jianpi therapy.
Shugan-Jianpi therapy mainly consists of herbs such as Atractylodes macrocephala Koidz. (Bai zhu), Bupleuri Radix (Chai hu), Paeoniae Radix Alba (Bai shao), which can influence the expression of transient receptor potential vanilloid-1 and Calcitonin Gene-Related Peptide (CGRP) in the colon tissue of the rat model with visceral hypersensitivity by increasing the pressure threshold of abdominal inwards reflex affected by colorectal distension so that to decrease the visceral sensitivity [59]. Besides, a clinical trial showed that Shugan therapy can also regulate the IBS-D patients' immune system by decreasing the number of IgM in the serum while enhancing the transformation of T-lymphocyte and increasing the number of T8 + lymphocyte [60]. Jianpi-Chushi therapy, which mostly contains herbs such as Atractylodes macrocephala Koidz. (Bai zhu), Citrus Reticulata (Chen pi), Poria Cocos (Schw.) Wolf (Fu ling), can regulate the intestinal flora by reducing the number of aerobes as well as increasing the probiotics, which can significantly relieve the clinical symptoms and achieve ideal effect [61].

PLOS ONE
Pharmacological treatments of Chinese herbal medicine for irritable bowel syndrome in adults Wenshen-Jianpi therapy, largely consists of Myristicae Semen (Rou dou kou), Zingiberis Rhizoma (Gan jiang), can regulate the expression of gastrointestinal hormones and their receptors such as melatonin, cholecystokinin, and CGRP [62]. Another clinical study also found that Wenshen-Jianpi therapy can regulate the expression of neurotransmitters such as 5-HT, neuropeptide Y, and immune factors such as TNF-γ [63]. In conclusion, CHM formulae can act on the IBS patients through multi-targets and multi-pathway, so that to improve their clinical symptoms There were several limitations to this meta-analysis. Firstly, the quality of included trials was not optimal due to methodological shortcomings. When evaluating these studies, we found that many lacked details on allocation concealment or blinding. The former will cause selection bias and the latter will result in detection bias. Besides, although many studies reported the dropouts, only 3 studies [34,41,44] performed intention-to-treat (ITT), which, to

PLOS ONE
Pharmacological treatments of Chinese herbal medicine for irritable bowel syndrome in adults some extent, may lead to incomplete outcome data and increase the attrition bias. Secondly, most of the included trials were single center with only 5 multicenter studies [27,33,34,35,40] and small sample sizes. The study contains 20 arms of SJ with 1361 patients, but there were only 4 arms of JC with 130 patients and 4 arms of WJ involving 163 patients. Due to the limited number of trials, the results of JC and WJ may cause bias. Therefore, more multi-center and large-scale trials should be conducted to offer more proofs in the future. Thirdly, the diversity of different CHM formulae may generate heterogeneity. Although we classified CHM formulae into 3 categories based on their function, the constitution of herbs was different from one formula to another and the dosage of each formula was personalized. Therefore, the differentiation of herbs and ingredients may affect the final effects. Moreover, the variation in the herbs themselves such as source, preparation, complication proportion, and decoction time might all be the source of heterogeneous. Besides, the differentiation of Chinese medicine formulations such as decoction, capsules, and powder, may influence the chemical composition and may result in heterogeneous. Fourthly, nearly all of the included trials were conducted in China and the populations were Chinese, which will generate publication and cultural bias. In addition, the positive-controlled in this study were not strictly in accord with the guideline. Hence, it does limit the value of the evidence, and more clinical trials using standard treatments as a comparison should be conducted in the future. Further, most of our included studies involved patients with IBS-D, which makes it hard to evaluate the efficacy of TCM in other subtypes of IBS. Finally, the treatment course of the included studies varied from 4 to 8 weeks, most of which lack long-term follow-up. Consequently, the recurrent rate remained unclear after treatment and thus was unable to evaluate the long-term efficacy of CHM formulae. In conclusion, it is still hard to find out whether patients with IBS in large-scale trials and other races can still get similar benefits from CHM formulae in the long-term use.

Conclusion
Evidence from this NMA confirmed that Shugan-Jianpi therapy, Jianpi-Chushi therapy, and Wenshen-Jianpi therapy could be beneficial for patients with IBS in relieving their different dimensions of clinical symptoms and improving their quality of life. These findings could provide physicians and patients with appropriate treatments based on the specific characteristics of IBS. However, additional high-quality RCTs should be performed to provide more powerful evidence in a wider population of IBS patients.