Clinical efficacy of pharmacological and non-pharmacological interventions for constipation-predominant irritable bowel syndrome

Abstract

BACKGROUND

Irritable bowel syndrome with constipation is the most common gastro-intestinal disorder that substantially affect bowel functionality, the burden of symptoms and the quality of life. Both pharmacological and non-pharmacological treatments are employed in clinical settings; however, there is limited comparative evidence regarding the effectiveness of these interventions.

AIM

To compare and rank the efficacy and safety of pharmacological and non-pharmacological interventions for constipation-predominant irritable bowel syndrome.

METHODS

A systematic review and Bayesian network meta-analysis of randomized controlled trials was conducted according to the preferred reporting items for systematic reviews and meta-analyses-network meta-analyses statement. PubMed, EMBASE, Web of Science, the Cochrane Library, CNKI, Wanfang, and VIP were searched from January 2007 to October 2025. Adults with constipation-predominant irritable bowel syndrome were included. Outcomes were complete spontaneous bowel movements, spontaneous bowel movements, Bristol stool form scale score, irritable bowel syndrome symptom severity scale score, irritable bowel syndrome quality of life score, and adverse events. Risk of bias was assessed using Cochrane RoB 2.0, and treatments were ranked by surface under the cumulative ranking curve.

RESULTS

Thirty-nine randomized controlled trials involving 15865 adults and 14 interventions were included. Acupuncture, tenapanor, and transcutaneous auricular vagus nerve stimulation ranked highest for complete spontaneous bowel movements. Tenapanor, transcutaneous auricular vagus nerve stimulation, and polyethylene glycol ranked highest for spontaneous bowel movements. Transcutaneous auricular vagus nerve stimulation, traditional Chinese medicine, and tenapanor ranked highest for Bristol stool form scale improvement. Acupuncture, probiotics, and transcutaneous auricular vagus nerve stimulation gave the best results for lowest ranked symptom severity denoting best benefit. The three therapies that reported the most benefit for quality of life were transcutaneous auricular vagus nerve stimulation, probiotics, and inulin. The risk of adverse events was lower in patients who received sensory adaptation training and transcutaneous auricular vagus nerve stimulation.

CONCLUSION

The benefits of pharmacological and non-pharmacological treatments for patients with constipation-predominant irritable bowel syndrome vary, but transcutaneous auricular vagus nerve stimulation, tenapanor, acupuncture, and probiotics appear particularly promising.

Key Words: Constipation-predominant irritable bowel syndrome; Pharmacological treatment; Non-pharmacological treatment; Clinical efficacy; Network meta-analysis; Systematic review

Core Tip: A comprehensive review and network meta-analysis has been conducted to determine the effectiveness of pharmacological and non-pharmacological therapies for constipation dominant irritable bowel syndrome. The assessment considered 14 different interventions depending on the important outcomes like frequency of bowel movements, consistency of stool, severity of symptoms, quality of life and adverse effects. According to the research, transcutaneous auricular vagus nerve stimulation, tenapanor, acupuncture, and probiotics are quite effective. These findings suggest the need for more personalized treatment approaches, and the need for further high-quality trials to confirm.

INTRODUCTION

Irritable bowel syndrome (IBS) the most common functional gut condition, is defined as the presence of no identifiable organic disease. The clinical manifestations of it are mainly gastrointestinal dysfunction, which consists of alternate constipation and diarrhea, abdominal pain and bloating. As per Rome IV criteria, IBS affects up to 4.1% of adults across the globe. Under Rome III criteria, it affects up to 10.1% of adults. Notably, there is a high prevalence rate among women and younger individuals[1]. Irritable bowel syndrome, or IBS, has four subtypes. There are the ones with dominantly constipation, those that are diarrhoea predominant, mixed type, and unclassified type. In clinical practice, the first three subtypes are frequently seen[2].

IBS with constipation (IBS-C), which accounts for over one-third of all IBS cases, is characterized chiefly by persistent, recurrent constipation[3]. The precise causes of IBS-C are currently unknown and are thought to involve a network of many factors. Potential causes include irregular gastrointestinal motility, psychological factors, brain-gut communication imbalance, gut microbiota changes and heightened visceral sensitivity[4,5]. Several surveys conducted on patients with IBS-C have revealed critical aspects of their self-image and sexual engagement. IBS-C sufferers reported a larger proportion of problems regarding their self-image, sex life, concentration and being satisfied with life, relative to other types of IBS[6]. These factors can severely disrupt the patient’s daily life and contribute to emotional problems which can worsen the symptoms of IBS-C and complicate disease management. There are pharmacological and non-pharmacological approaches for the treatment of IBS-C. The main types of drug therapies are secretagogues, laxatives, traditional Chinese medicine and sodium-hydrogen exchanger 3 inhibitors[7,8]. Non-pharmacological treatments for IBS-C typically include transcutaneous auricular vagus nerve stimulation (taVNS), dietary and lifestyle modification, probiotics, and behavioral therapy[8]. Although medications can effectively and rapidly improve symptoms such as constipation and abdominal pain, long-term use may impose a financial burden on patients. Non-pharmacological treatments such as acupuncture adverse effects and a low long-term cost treatment burden, although these benefits may vary across modalities and study designs[9,10]. Nevertheless, the wide range of available treatment options and the uncertainty regarding their relative advantages and limitations continue to complicate clinical decision-making. To date, there is minimal evidence that systematically evaluates and contrasts the clinical effectiveness and safety of both pharmacological and non-pharmacological approaches for managing constipation-predominant IBS. Consequently, this study aimed to compare and prioritize these interventions through a Bayesian network meta-analysis to furnish evidence for personalized clinical management.

MATERIALS AND METHODS

Study registration

This research was earlier registered with the International Registry for Prospective Systematic Reviews (PROSPERO) under the registration number: No. CRD420251171965.

Search strategy and study design

The present investigation was carried out in alignment with the standards set forth by the preferred reporting items for systematic reviews and meta-analyses-network meta-analyses (PRISMA-NMA) guidelines. In this process, two researchers, Li C and Yang JW, undertook independent systematic literature searches guided by specific inclusion and exclusion criteria that had been established prior to the search. A wide range of databases was searched, which were the widely used and high-quality databases like PubMed, EMBASE, Web of Science, The Cochrane Library and regional databases CNKI, Wanfang and the VIP database. The searches were conducted carefully in English and Chinese collecting titles, abstracts, and keywords from studies published from January 2007 to October 2025. The review focused on the clinical efficacy of pharmacological as well as non-pharmacological management in adults suffering constipation-predominant IBS (IBS-C). Only randomized controlled trials (RCTs) were eligible for inclusion. To organize the search effectively, the researchers concentrated on four central themes: “Constipation-predominant irritable bowel syndrome”, “clinical efficacy”, “pharmacological vs non-pharmacological interventions”, and “randomized controlled trials”. Additionally, a variety of keywords pertinent to the research topic were utilized, including terms such as IBS, constipation-predominant IBS-C, constipation-predominant IBS, IBS with constipation, and patients with constipation-predominant IBS, among others. Interventions include polyethylene glycols, lactulose, traditional Chinese medicine, linaclotide, plecanatide, lubiprostone, tenapanor, probiotics, dietary interventions [high fiber, inulin, low-fermentable oligosaccharides, disaccharides, monosaccharides, polyols (FODMAP), personalized diet], acupuncture, electroacupuncture, moxibustion, taVNS, biofeedback, and other pharmacological and non-pharmacological treatments; primary outcome measures include: Complete spontaneous bowel movement (CSBM), SBM, Bristol stool form scale (BSFS), IBS severity scoring system (IBS-SSS), IBS-quality of life (QOL), adverse events, etc. In the event of a dispute, the resolution shall be determined by a third-party expert who possesses relevant expertise in the field.

Inclusion and exclusion criteria

Inclusion criteria: (1) Study subjects were patients aged ≥ 18 years diagnosed with IBS-C based on Rome II, Rome III, or Rome IV criteria; (2) Study design: RCTs, regardless of blinding status; (3) Interventions: Pharmacological and non-pharmacological interventions: Polyethylene glycols, lactulose, traditional Chinese medicine, linaclotide, plecanatide, lubiprostone, tenapanor, probiotics, dietary interventions (high fiber, inulin, low-FODMAP, personalized diet), acupuncture, electroacupuncture, moxibustion, TaVNS, biofeedback, etc. Control groups included placebo, conventional treatment, or comparisons between different types of drug/non-drug interventions; (4) Outcome measures: Primarily include CSBM, SBM, BSFS, IBS-SSS, IBS-QOL, adverse events, etc., with at least one measure present in the selected studies; and (5) Publication dates range from January 2007 to October 2025, with language restrictions limited to English or Chinese.

Exclusion criteria: (1) Study subjects were patients under 18 years of age diagnosed with constipation-predominant IBS; (2) Non-RCTs (case-control studies, cohort studies, animal experiments, meta-analyses, reviews, case reports, conference abstracts, etc.); (3) Intervention measures were surgical treatments; (4) Studies lacking relevant primary outcome measures, where primary outcome measures could not be extracted, or where data were incomplete; and (5) Studies with duplicate publications or overlapping data.

Literature screening and data extraction

All records obtained from various databases, including PubMed, EMBASE, Web of Science, the Cochrane Library, CNKI, Wanfang, and VIP, were subsequently uploaded into EndNote reference management software for organization and management. Two researchers, Li C and Yang JW, conducted an independent review of the literature, applying specific inclusion and exclusion criteria that had been established beforehand. If two researchers had any differences in their assessment, the two would discuss the discrepancies together to resolve them. Studies were included in the review if the inter-rater agreement assessed using kappa was > 0.90. If there’s a disagreement, a third-party expert will resolve it who has expertise in the relevant field.

The extracted materials include: (1) The attributes of the study consist of the principal author, year of publication, country involved, and design of the study; (2) The initial patient characteristics include both the intervention and control strategy, sample size, average age with standard deviation, distribution of sex; (3) Participants required to have a diagnosis of constipation-predominant IBS clearly; and (4) The primary outcome measurements were CSBM, SBM, BSFS, IBS-SSS, IBS-QOL and adverse effects. A CSBM referred to a bowel movement that occurred within the past 24 hours, without any use of laxatives and with a self-report of complete evacuation. Therefore, each qualifying bowel was counted as one CSBM. SBM refers to a motion which occurred within a 24 hours times frame, also without the aid of laxatives, but without a sense of complete evacuation. The BSFS scale ranges from 1 to 7, which represent stool consistency from the hardest to the softest, where low-value indicates a more severe case of constipation. The IBS-SSS is a validated measure of the severity of IBS symptoms in five domains measured on a 100-point Likert scale for each domain (maximum total score = 500). Higher scores indicate greater severity of symptoms. IBS-QOL has 34 items organized into 8 domains. The domains are irritability, daily activities, body image, health worry, food avoidance, social interaction, sexual issues, personal relationships. Scores on the IBS-QOL scale range from 0 to 100. A higher score indicates better QOL. The number of participants experiencing at least one adverse effect during the treatment phase was measured. More patients experiencing adverse effects were indicative of high occurrence or prevalence of treatment-related effects.

Risk of bias assessments

All RCTs incorporated into the study had their bias risk assessed independently by two reviewers utilizing the Cochrane Risk of Bias 2.0 tool. This tool evaluates five critical areas that may affect the results of the trial. The first domain on the randomized study article highlights the biases which may arise due randomization process used in the clinical trial. It emphasizes the need of allocation concealment to negate selection bias. The second domain focuses on assessing the risk of bias arising from deviations from the interventions that were allocated to the participants. In particular, it examines whether the analyses departed from the intention-to-treat principle, where relevant, to protect the integrity of randomized assignments. The third domain considers whether missing outcome data can produce a non-robust trial. The fourth domain is related to measurement of outcomes. It highlights the importance of blinding outcome assessors with respect to measurement bias. Finally, the fifth domain assesses the risk of selection bias in the reported results. This can be a critical bias in the study. Each of these domains as well as the overall risk of bias for each trial was assessed and assigned as low risk, some concerns, or high risk. In cases where the reviewers encountered disagreements in their evaluations, they engaged in discussions to reach a consensus; if necessary, a third reviewer was consulted to provide additional insights, thereby enhancing the reliability of the bias assessment process.

Statistical analysis

A meta-analysis using Bayesian networks was conducted with the BUGSnet and gemtc packages in R version 4.4.2. This approach, which builds upon traditional meta-analysis, combines direct evidence with indirect evidence, enabling the comparison of three or more treatments and the establishment of hierarchical rankings regarding the effectiveness of those interventions. Bayesian inference was based on Markov Chain Monte Carlo methods, with four chains run for 50000 iterations each; the first 10000 iterations were discarded as burn-in to ensure model convergence. When closed loops were present in the network, node-splitting analyses were applied to assess the consistency between direct and indirect evidence. A P value exceeding 0.05 suggested that there was no meaningful statistical inconsistency. Subsequently, the surface under the cumulative ranking curve (SUCRA) values was calculated to assess the relative effectiveness of the interventions across various outcome measures. In addition, exploratory conventional pairwise meta-analyses were conducted using STATA version 16.0 for selected outcomes with direct placebo/sham-controlled comparisons to further assess comparability and heterogeneity across studies. Estimates of effect were aggregated through standard meta-analytic models, while the evaluation of heterogeneity was conducted with the I² statistic and Cochran’s Q test. Ultimately, the GRADEpro system was employed to assess the quality of evidence for the primary outcomes.

RESULTS

Literature search results

The initial literature search identified 8088 potentially relevant records. After a preliminary screening, 5977 records were excluded, including duplicate studies, records eliminated by automated screening tools, and studies involving animal experiments, reviews, letters, guidelines, case reports, pathophysiological research, and meta-analyses. The titles and abstracts of the remaining records were thoroughly screened, which led to the exclusion of 1960 studies. These studies failed to meet the predetermined inclusion criteria related to the specific characteristics of the study population, the nature of the intervention, or the design of the study itself. Consequently, 151 studies were included in the present study after applying a thorough full-text reading screening process. After a thorough assessment of their full text, we had to exclude a further 97 studies due to inaccessibility of the full text, ineligibility for outcome measures, poor study quality, and inability to extract or pool outcome data. All in all, there were a total of 39 RCTs that were included in the study based on the preset inclusion criteria for this Bayesian network meta-analysis. The study selection process is depicted in Figure 1. The study selection process represents the studies which are included in this review and how these studies were chosen.

Figure 1  Literature screening process.

Basic characteristics of the included studies

Table 1[11-48] shows that a total of 39 RCTs were included with 15865 adults with IBS-C aimed to investigate 14 interventions: Polyethylene glycol[11,26], plecanatide[12,15,17], linaclotide[13,14,16,18-22,30-32], tenapanor[23,24,29], lubiprostone[33-35], traditional Chinese medicine[25,27,28], probiotics[36,37], inulin[38,43], dietary fiber[39,45], acupuncture[40-42,47,48], moxibustion[41,42], transcutaneous auricular vagus nerve stimulation[44,46], sensory adaptation training[8], and lactulose oral solution[27,47]. Table 1 summarizes the baseline characteristics of the participants, characteristics of the study design and outcome measures.

Table 1 Clinical and demographic characteristics of studies included in network meta-analyses (n = 39 studies)[11-48].

Ref.	Year	Country	Study type	Intervention mode		Number of cases		Age (years)		Gender (male/female)		Inclusion criteria	Outcome indicators
				Experimental group	Control group	Experimental group	Control group	Experimental group	Control group	Experimental group	Control group
Chapman et al[11]	2013	Poland	RCT	PEG	Placebo	68	71	43.6 ± 14.9	39.1 ± 14.5	10/57	13/57	Aged 18-80 years; IBS-C diagnosed according to Rome III criteria	CSBM, SBM, BSFS, adverse events
Brenner et al[12]	2024	United States	RCT	Plecanatide	Placebo	560	544	43.7 ± 13.7	43.2 ± 14.2	142/418	131/413	Adults aged ≤ 85 years with IBS-C (Rome III criteria)	CSBM, SBM, adverse events
Brenner et al[12]	2024	United States	RCT	Plecanatide	Placebo	159	173	43.0 ± 15.7	46.7 ± 14.0	47/112	52/121	Adults aged ≤ 85 years with IBS-C (Rome III criteria)	CSBM, SBM, adverse events
Chey et al[13]	2021	United States	RCT	Linaclotide	Placebo	67	66	44.8 ± 14.9	45.4 ± 14.7	8/59	13/53	Patients aged 18 years or older who met the Rome III criteria for IBS-C	SBM, BSFS, adverse events
Chey et al[13]	2021	United States	RCT	Linaclotide	Placebo	67	66	44.7 ± 13.7	45.4 ± 14.7	8/59	13/53	Patients aged 18 years or older who met the Rome III criteria for IBS-C	SBM, BSFS, adverse events
Chey et al[13]	2021	United States	RCT	Linaclotide	Placebo	67	66	46.5 ± 12.7	45.4 ± 14.7	12/55	13/53	Patients aged 18 years or older who met the Rome III criteria for IBS-C	SBM, BSFS, adverse events
Peng et al[14]	2022	China	RCT	Linaclotide	Placebo	327	332	39.6 ± 12.8	40.5 ± 13.9	63/264	45/287	Patients aged 18 years or older who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Brenner et al[15]	2018	United States	RCT	Plecanatide	Placebo	351	354	43.0 ± 13.8	43.0 ± 13.8	84/267	82/272	Adults aged 18-85 years who met Rome III criteria for IBS-C	CSBM, BSFS, adverse events
Brenner et al[15]	2018	United States	RCT	Plecanatide	Placebo	349	354	43.2 ± 13.3	43.0 ± 13.8	83/266	82/272	Adults aged 18-85 years who met Rome III criteria for IBS-C	CSBM, BSFS, adverse events
Brenner et al[15]	2018	United States	RCT	Plecanatide	Placebo	377	379	44.0 ± 14.6	44.8 ± 14.7	107/270	107/272	Adults aged 18-85 years who met Rome III criteria for IBS-C	CSBM, BSFS, adverse events
Brenner et al[15]	2018	United States	RCT	Plecanatide	Placebo	379	379	43.1 ± 14.2	44.8 ± 14.7	106/273	107/272	Adults aged 18-85 years who met Rome III criteria for IBS-C	CSBM, BSFS, adverse events
Chey et al[16]	2012	United States	RCT	Linaclotide	Placebo	401	403	44.6 (19-82)	44.0 (18-87)	33/368	51/352	Men and women aged 18 years or older who met modified Rome II criteria for IBS-C	Adverse events
Cash et al[17]	2023	United States	RCT	Plecanatide	Placebo	188	168	43.0 (20-80)	42.9 (19-78)	41/147	37/131	Age ≥ 18 years who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Cash et al[17]	2023	United States	RCT	Plecanatide	Placebo	171	168	41.3 (18-76)	42.9 (19-78)	43/128	37/131	Age ≥ 18 years who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Fukudo et al[18]	2018	Japan	RCT	Linaclotide	Placebo	116	112	41.5 ± 11.3	41.6 ± 10.8	7/109	12/100	Male and female outpatients aged 20-64 years with IBS-C based on the Rome III diagnostic criteria	Adverse events
Fukudo et al[18]	2018	Japan	RCT	Linaclotide	Placebo	111	112	41.7 ± 11.9	41.6 ± 10.8	9/102	12/100	Male and female outpatients aged 20-64 years with IBS-C based on the Rome III diagnostic criteria	Adverse events
Fukudo et al[18]	2018	Japan	RCT	Linaclotide	Placebo	112	112	41.8 ± 9.8	41.6 ± 10.8	13/99	12/100	Male and female outpatients aged 20-64 years with IBS-C based on the Rome III diagnostic criteria	Adverse events
Fukudo et al[18]	2018	Japan	RCT	Linaclotide	Placebo	107	112	38.9 ± 11.0	41.6 ± 10.8	8/99	12/100	Male and female outpatients aged 20-64 years with IBS-C based on the Rome III diagnostic criteria	Adverse events
Fukudo et al[19]	2018	Japan	RCT	Linaclotide	Placebo	249	251	41.6 ± 10.7	42.2 ± 11.3	37/212	24/227	Male and female outpatients aged 20-79 years diagnosed as having IBS-C based on the Rome III diagnostic criteria	Adverse events
Rao et al[20]	2020	United States	RCT	Linaclotide	Placebo	26	13	40.1 ± 2.6	46.4 ± 2.1	1/25	0/13	Adults aged 18-85 years who met Rome III criteria for IBS-C	BSFS, IBS-QOL, adverse events
Chang et al[21]	2021	United States	RCT	Linaclotide	Placebo	306	308	46.5 (19-85)	46.8 (18-79)	65/241	53/255	Men or women; ≥ 18 years; met Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Yang et al[22]	2018	China	RCT	Linaclotide	Placebo	417	422	41.0 (18-77)	41.3 (18-80)	84/333	67/355	Men and women ≥ 18 years of age were eligible if they met the Rome III criteria for IBS-C	Adverse events
Chey et al[23]	2017	United States	RCT	Tenapanor	Placebo	88	90	45.8 ± 12.7	46.0 ± 13.8	12/76	13/77	Adults 18-75 years who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Chey et al[23]	2017	United States	RCT	Tenapanor	Placebo	89	90	45.3 ± 14.1	46.0 ± 13.8	12/77	13/77	Adults 18-75 years who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Chey et al[23]	2017	United States	RCT	Tenapanor	Placebo	89	90	45.8 ± 12.2	46.0 ± 13.8	10/79	13/77	Adults 18-75 years who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Chey et al[24]	2020	United States	RCT	Tenapanor	Placebo	307	299	45.0 ± 13.4	44.9 ± 13.0	63/244	50/249	Men and women aged 18-75 years who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Nasab et al[25]	2025	Iran	RCT	Traditional Chinese medicine	Placebo	30	30	39 ± 4.38	38.77 ± 4.22	13/17	14/16	Patients aged 18-50 years who met modified Rome IV criteria for IBS-C	IBS-QOL, IBS-SSS
Yu and Xuan[26]	2017	China	RCT	PEG	Placebo	48	48	56.95 ± 5.43	56.45 ± 5.62	26/22	25/23	Diagnosed with IBS-C according to the Rome III diagnostic criteria for IBS	Adverse reactions
Li et al[27]	2025	China	RCT	Traditional Chinese medicine	Lactulose oral solution	36	36	43.28 ± 14.76	43.03 ± 14.84	20/16	12/24	Age 18-70 years; meeting Rome IV diagnostic criteria for IBS-C	BSFS, IBS-SSS, adverse events
Hao et al[28]	2025	China	RCT	Traditional Chinese medicine	Linaclotide	51	46	42.87 ± 10.18	45.42 ± 9.29	22/29	18/28	Age > 18 years who met the Rome IV criteria for IBS-C	BSFS, IBS-SSS, IBS-QOL, adverse events
Chey et al[29]	2020	United States	RCT	Tenapanor	Placebo	293	300	46.1 ± 13.1	44.8 ± 13.8	53/240	53/247	Men and women aged 18-75 years who met the Rome III criteria for IBS-C	CSBM, SBM, BSFS, adverse events
Rao et al[30]	2012	United States	RCT	Linaclotide	Placebo	405	395	43.3 (19-81)	43.7 (18-84)	38/367	38/357	At least 18 years of age; met modified Rome II criteria for IBS-C	Adverse events
Andresen et al[31]	2007	United States	RCT	Linaclotide	Placebo	12	12	43.4 ± 2.8	38.5 ± 2.8	0/12	0/12	Female, nonpregnant, non-breastfeeding participants aged 18-65 years with IBS-C, based on Rome II criteria	BSFS
Andresen et al[31]	2007	United States	RCT	Linaclotide	Placebo	12	12	35.4 ± 2.3	38.5 ± 2.8	0/12	0/12	Female, nonpregnant, non-breastfeeding participants aged 18-65 years with IBS-C, based on Rome II criteria	BSFS
Rao et al[32]	2014	Not explicitly stated	RCT	Linaclotide	Placebo	805	797	NA	NA	NA	NA	Patients were eligible to participate if they were at least 18 years of age and met modified Rome II criteria for IBS-C	Adverse events
Chang et al[33]	2016	United States	RCT	Lubiprostone	Placebo	325	180	45.4 ± 12.4	46.4 ± 12.2	21/304	6/174	Meeting the Rome II Modular Questionnaire criteria for IBS-C	Adverse events
Johanson et al[34]	2008	United States	RCT	Lubiprostone	Placebo	51	48	46.5 ± 10.1	44.6 ± 11.1	4/47	4/44	Men and women aged 18-80 years or older who met modified Rome II criteria for IBS-C	Adverse events
Johanson et al[34]	2008	United States	RCT	Lubiprostone	Placebo	49	48	48.3 ± 11.9	44.6 ± 11.1	3/46	4/44	Men and women aged 18-80 years or older who met modified Rome II criteria for IBS-C	Adverse events
Johanson et al[34]	2008	United States	RCT	Lubiprostone	Placebo	45	48	43.9 ± 11.6	44.6 ± 11.1	7/38	4/44	Men and women aged 18-80 years or older who met modified Rome II criteria for IBS-C	Adverse events
Drossman et al[35]	2009	United States	RCT	Lubiprostone	Placebo	769	385	46.1 (19.0, 83.0)	47.7 (18.0, 85.0)	71/698	26/359	Eligible patients being at least 18 years of age and meeting the Rome II Modular Questionnaire Criteria for IBS-C	Adverse events
Kwon et al[36]	2024	Republic of Korea	RCT	Probiotics	Placebo	15	15	36.4 ± 13.6	37.1 ± 13.7	NA	NA	Participants qualified for randomization had clinical responses based on the Rome IV criteria	IBS-SSS, IBS-QOL
Mourey et al[37]	2022	France	RCT	Probiotics	Placebo	230	226	41.2 ± 13.96	39.9 ± 14.56	28/202	36/190	Females and males aged 18-75 years were eligible to participate if they met the Rome IV criteria for IBS-C	Adverse events
Akçalı et al[38]	2025	Turkey	RCT	Inulin	Placebo	17	17	36.35 ± 11.6	38.23 ± 12.9	4/13	4/13	Aged between 19 years and 65 years, diagnosed with IBS-C according to the Rome IV criteria	IBS-SSS, IBS-QOL
Camacho-Díaz et al[39]	2023	Mexico	RCT	Fiber	Placebo	24	26	50.0 ± 2.0	50.0 ± 2.0	NA	NA	Screening of patients was based on the Rome III diagnostic criteria for IBS-C	IBS-QOL
Rao et al[8]	2021	United States	RCT	Sensory adaptation training	Placebo	26	23	45.0 ± 3.0	47.0 ± 3.1	0/26	4/19	Age ≥ 18 years who met the Rome III criteria for IBS-C	BSFS, adverse events
Huang et al[40]	2022	China	RCT	Acupuncture	Placebo	26	26	51.4 ± 16.2	49.6 ± 17.5	9/17	10/16	Men and women aged 18-75 years who met the Rome IV diagnostic criteria for IBS-C	CSBM, BSFS, IBS-SSS, IBS-QOL
Lu et al[41]	2015	China	RCT	Acupuncture	Moxibustion	19	21	40 ± 12	42 ± 8	NA	NA	The eligibility of the initial IBS-C patients according to the Rome III diagnostic criteria	BSFS
Zhao et al[42]	2018	China	RCT	Acupuncture	Moxibustion	30	30	40.40 ± 12.67	42.33 ± 8.68	NA	NA	Complied with Rome III diagnostic criteria of C-IBS; were 18-65 years old	BSFS, adverse events
Isakov et al[43]	2023	Russia	RCT	Inulin + vitamins	Placebo	20	20	48.7 ± 17.7	47.7 ± 15.9	0/20	0/20	Age ranging from 18 years to 80 years old; diagnosis of IBS-C based on ROME IV criteria	BSFS
Shi et al[44]	2021	China	RCT	taVNS	Placebo	21	21	41.5 ± 15.4	49.6 ± 15.6	4/17	6/15	Aged 18-75 years; met the Rome IV diagnostic criteria for IBS-C	CSBM, BSFS, IBS-SSS, IBS-QOL
Choi et al[45]	2011	Korea	RCT	Fiber	Placebo	16	18	NA	NA	NA	NA	Volunteers between 18 years and 70 years of age who met the Rome III criteria with IBS-C	BSFS
Liu et al[46]	2024	China	RCT	taVNS	Placebo	20	20	48.10 ± 11.54	48.75 ± 12.23	5/15	5/15	Individuals aged 18-70 years diagnosed with IBS-C according to the Rome IV criteria	CSBM, SBM, BSFS, IBS-SSS, IBS-QOL, adverse events
Pei et al[47]	2015	China	RCT	Acupuncture	Lactulose oral solution	30	30	44 ± 12	44 ± 14	11/19	9/21	Age between 18 years and 65 years who meeting the Rome III diagnostic criteria for IBS-C	IBS-QOL
Guo et al[48]	2021	China	RCT	Acupuncture	PEG	92	45	46.3 ± 14.3	47.0 ± 13.6	33/59	17/28	Age between 18 years and 70 years who meeting the Rome III diagnostic criteria for IBS-C	IBS-SSS, IBS-QOL, adverse events

RCT: Randomized controlled trials; PEG: Polyethylene glycol; taVNS: Transcutaneous auricular vagus nerve stimulation; NA: Not available; IBS-C: Irritable bowel syndrome with constipation; CSBM: Complete spontaneous bowel movement; SBM: Spontaneous bowel movement; BSFS: Bristol stool form scale; IBS-QOL: Irritable bowel syndrome quality of life; IBS-SSS: Irritable bowel syndrome severity scoring system.

The quality assessment of the included studies

This study included 39 RCTs in total. The selected studies’ quality of methodology was systematically evaluated using the Cochrane RoB 2.0 tool with the help of R software version 4.4.2. As depicted in Figure 2, most studies were assessed as having a low risk of bias in several domains, such as randomization process, completeness of outcome data, and selective reporting. This suggests that most of the studies had a high quality and reliable outcomes. Notwithstanding, there were two domains where a high risk of bias was observed, specifically, deviations from intended intervention and measurement of outcome. Numerous studies such as Yu and Xuan[26], Hao et al[28], Lu et al[41], Zhao et al[42], Isakov et al[43], Pei et al[47] show a disconnect between the intervention provided and the study protocol. Variations may induce bias through modified interventions. In addition, the use of a single blind or completely unblinded in the studies of Yu and Xuan[26], Li et al[27], Hao et al[28], Akçalı et al[38], Huang et al[40], Lu et al[41], Zhao et al[42], Isakov et al[43], Liu et al[46], Pei et al[47], Guo et al[48]. might increase the risk of performance and measurement bias. The methodological quality of the included studies was acceptable, but improvements in adherence to intervention protocols and implementation of double-blind designs are required.

Figure 2  Risk of bias in randomized controlled.

Certainty of evidence

Through the Cochrane RoB 2.0 tool, this study evaluated the methodology of the assays of 39 RCTs. The results indicated that most studies showed low risk of bias for several domains so the studies were robust overall. The domains consisted of randomization, completeness of outcome data and selection of reported results. Some domains show residual risk of bias; for instance, some studies did not adhere to intended interventions or missed blinding of participants and personnel. It is essential to assess the potential residual biases in the methodology used in implications. The primary outcome measures in question might be more susceptible to bias, due to them being subjective in nature. This means that these measures can more easily be open for interpretation and variation. The GRADE rating system was used to assess the quality of evidence for the primary outcomes (CSBMs, spontaneous bowel movements, BSFS score, IBS symptom severity scale score, IBS QOL score and adverse event). The findings showed that the majority of outcomes were rated as moderate quality, indicating a reasonable level of confidence. The evidence behind the IBS-QOL was rated high quality which indicates a strong basis for this finding. It was found that the risk of bias was the critical factor restricting the overall certainty of the evidence of the studies. In summary, both pharmacological and non-pharmacological interventions were promising in alleviating constipation and abdominal pain in patients diagnosed with constipation-predominant IBS. Nonetheless, these results need to be validated in high-quality, multicenter RCTs. As shown in Table 2, we can be certain about the evidence.

Table 2 Certainty of evidence.

Quality assessment							Summary of findings
Participants (studies) follow up	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	Overall quality of evidence	Study event rates (%)		Relative effect (95%CI)	Anticipated absolute effects
							With control	With outcomes		Risk with control	Risk difference with outcomes (95%CI)
CSBM (critical outcome; better indicated by lower values)
8318 (17 studies)	Serious	Serious	No serious indirectness	No serious imprecision	Undetected	Moderate due to risk of bias, inconsistency, large effect	4159	4159			The mean CSBM in the intervention groups was 1.1 higher (1.05 to 1.16 higher)
SBM (critical outcome; better indicated by lower values)
5706 (16 studies)	Serious	Serious	No serious indirectness	No serious imprecision	Undetected	Moderate due to risk of bias, inconsistency, large effect	2848	2858			The mean SBM in the intervention groups was 1.65 higher (1.58 to 1.72 higher)
BSFS (critical outcome; better indicated by lower values)
7689 (28 studies)	Serious	Serious	No serious indirectness	No serious imprecision	Undetected	Moderate due to risk of bias, inconsistency, large effect	3834	3855			The mean BSFS in the intervention groups was 0.97 higher (0.94 to 1.01 higher)
IBS-SSS (critical outcome; better indicated by lower values)
544 (9 studies)	Serious	Serious	No serious indirectness	No serious imprecision	Undetected	Moderate due to risk of bias, inconsistency, large effect	245	299			The mean IBS-SSS in the intervention groups was 37.34 lower (46.23 to 28.45 lower)
IBS-QOL (critical outcome; better indicated by lower values)
617 (11 studies)	Serious	No serious inconsistency	No serious indirectness	No serious imprecision	Undetected	High due to risk of bias, large effect	278	339			The mean IBS-QOL in the intervention groups was 7.27 higher (5.32 to 9.22 higher)
Adverse events (critical outcome)
16725 (36 studies)	Serious	Serious	No serious indirectness	No serious imprecision	Undetected	Moderate due to risk of bias, inconsistency, large effect	2720/8076 (33.7%)	3357/8649 (38.8%)	OR = 1.23 (1.15 to 1.32)	Study population
										337 per 1000	48 more per 1000 (from 32 more to 65 more)
										Moderate
										346 per 1000	48 more per 1000 (from 32 more to 65 more)

CI: Confidence interval; CSBM: Complete spontaneous bowel movement; SBM: Spontaneous bowel movement; BSFS: Bristol stool form scale; IBS-QOL: Irritable bowel syndrome quality of life; IBS-SSS: Irritable bowel syndrome severity scoring system.

Network meta-analysis

Change in CSBM: Network evidence graph. This analysis included six interventions and one placebo control group (A: Placebo; B: Polyethylene glycol; D: Plecanatide; E: Linaclotide; F: Tenapanor; L: Acupuncture; N: TaVNS). A Bayesian network meta-analysis performed a pre- and post-intervention change in number of CSBM as outcomes. The evidence network focused on placebo comparator, with nodes assigned to each intervention and placebo. Lines were used to demonstrate direct comparisons between the various interventions, with thicker lines signifying a higher number of included studies (Figures 3, 4, 5, 6 and 7). Figure 3 presents the related evidence network. Comparisons between the fixed-effect and random-effects models (Figure 5) showed that the random-effects model has a lower deviance information criterion (DIC) value (66.1) and smoother posterior distribution, suggesting it fits the data better and more appropriately accounts for between-study heterogeneity. Therefore, the final analysis was carried out using the random-effects model. The evaluations of the bias distributions under the consistency and inconsistency models supported the robustness of the aforementioned analyses. The consistency model displayed smaller bias and dispersion and lower DIC than the inconsistency model (DIC consistency = 66.16 vs DIC inconsistency = 154.45) indicating a better fit (Figure 6) and a robustness of results across grids. As there were no closed-loop structures evident within the network, further consistency testing was not considered necessary.

Figure 3 Network evidence and cumulative ranking surface under the cumulative ranking curve diagram. A: Complete spontaneous bowel movement; B: Spontaneous bowel movement; C: Bristol stool form scale; D: Irritable bowel syndrome severity scoring system; E: Irritable bowel syndrome-quality of life; F: Adverse events. taVNS: Transcutaneous auricular vagus nerve stimulation.

Figure 4 Comparison table of treatment effects League. A: Complete spontaneous bowel movement; B: Spontaneous bowel movement; C: Bristol stool form scale; D: Irritable bowel syndrome severity scoring system; E: Irritable bowel syndrome-quality of life; F: Adverse events. ¹Indicate that statistical significance, which means that 95% of the interval does not cross 0. taVNS: Transcutaneous auricular vagus nerve stimulation.

Figure 5 Bayesian network meta-analysis model fitting diagnostic diagram. A: Complete spontaneous bowel movement; B: Spontaneous bowel movement; C: Bristol stool form scale; D: Irritable bowel syndrome severity scoring system; E: Irritable bowel syndrome-quality of life; F: Adverse events. DIC: Deviance information criterion.

Figure 6 Consistency test results of Bayesian network meta-analysis. A: Complete spontaneous bowel movement; B: Spontaneous bowel movement; C: Bristol stool form scale; D: Irritable bowel syndrome severity scoring system; E: Irritable bowel syndrome-quality of life; F: Adverse events. DIC: Deviance information criterion.

Figure 7 Cumulative sorted curve of pharmaceutical and non-pharmaceutical intervention for outcome indicators chart. A: Complete spontaneous bowel movement; B: Spontaneous bowel movement; C: Bristol stool form scale; D: Irritable bowel syndrome severity scoring system; E: Irritable bowel syndrome-quality of life; F: Adverse events. taVNS: Transcutaneous auricular vagus nerve stimulation.

NMA: The network meta-analysis findings are presented in Figure 4, which shows for each comparison, the estimated differences in efficacy between interventions. The higher the value, the higher the efficacy. The lower the value, the lower the efficacy. The different shades of color shown correspond with the efficacy of the method with a darker shade representing better and a lighter shade worse. The estimate denotes the effect due to the column treatment against that of the row comparator; the 95% confidence/credible intervals are shown in parentheses in each cell. The upper triangle and the lower triangle are said to be reciprocal. Blue shading indicates a positive effect; yellow to orange shading indicates a negative effect, with darker colors indicating a larger absolute effect size. No comparison within treatment in gray diagonal cells. Estimates marked with “1” are statistically significant, which means that 95% of the interval does not cross 0.

Efficacy ranking: According to SUCRA analysis, acupuncture had the highest probability of increase in CSBMs compared with other interventions, followed by tenapanor, taVNS, linaclotide, polyethylene glycol, and placenatide. Higher SUCRA values indicate greater efficacy of the intervention. These results are shown in the Figure 3 and Figure 7.

Change in SBM: Network evidence graph. This analysis included five interventions and one placebo control group (A: Placebo; B: Polyethylene glycol; D: Plecanatide; E: Linaclotide; F: Tenapanor; N: TaVNS). The Bayesian network meta-analysis was performed, in which change in SBM pre-intervention and post-intervention was the primary outcome. Figure 3 shows the associated network structure. Figure 5 indicates that when we apply the random-effects model, which yielded a DIC of 61.08, the model fit diagnostics were far better compared to the fixed-effects model (121.19). This indicates a better fit and more stability of the random-effects model. Moreover, the model consistency assessment in Figure 6 has a DIC value of 60.92. This is much lower than the DIC of the model inconsistency assessment, which is 121.24. That shows strong agreement between direct and indirect evidence. Direct and indirect comparisons revealed no major differences between them.

NMA: According to the network meta-analysis results indicated in Figure 4, each cell has the estimated difference in efficacy of the interventions for the respective comparisons. Also, higher values indicate higher efficacy and lower values indicate lower efficacy. The color gradient represents this difference, with darker colors representing the most effective pesticides and lighter colors the least. See above text for the caption.

Efficacy ranking: The SUCRA analysis findings showed that tenapanor (90.1%) was ranked first followed by taVNS (79.3%), polyethylene glycol (58.5%), linaclotide (48.3%), plecanatide (23.8%) and placebo (0.001%). A higher SUCRA value in an intervention indicates greater efficacy for spontaneous bowel movements. These findings are illustrated in Figure 3 and Figure 7.

Change in BSFS score: Network evidence graph. This analysis included 12 interventions and one placebo control group (A: Placebo; B: Polyethylene glycol; C: Lactulose oral solution; D: Plecanatide; E: Linaclotide; F: Tenapanor; H: Traditional Chinese medicine; J: Dietary fiber; K: Inulin; L: Acupuncture; M: Moxibustion; N: TaVNS; O: Sensory adaptation training). BSFS score change was the primary outcome for the Bayesian network meta-analysis that was done. The structure of the network is shown in Figure 3, and model diagnostics were shown in Figure 5. The DIC value for the random-effects model of 114.45 was substantially lower than the approximately 251.12 DIC value for the fixed-effects model. According to the consistency assessment (Figure 6), the DIC value of the consistency model (114.29) was much lower than the DIC value of the inconsistency model (251.12). This finding indicates that direct and indirect evidence are in strong agreement, and the results are reliable. Since there were no closed-loop structures among the interventions further tests of consistency were not necessary.

NMA: The network meta-analysis results are summarized in Figure 4. The cells contain the estimated differences in efficacy for each comparison. In other words, the value in the cell indicates which treatment is more efficacious. Higher values indicate a more efficacious treatment while lower values indicate a less efficacious treatment. The color gradient corresponds to these differences: A darker shade indicates better performance while a lighter shade indicates worse performance. Refer to the above text for the caption.

Efficacy ranking: The SUCRA analysis ranked the interventions aimed at improving stool consistency as follows: TaVNS at 92.7%, traditional Chinese medicine at 83.4%, tenapanor at 71.0%, acupuncture at 63.7%, inulin at 62.0%, polyethylene glycol at 61.8%, lactulose oral solution at 61.4%, sensory adaptation training at 46.1%, linaclotide at 42.5%, dietary fiber at 28.8%, plecanatide at 25.7%, placebo at 10.5%, and moxibustion at 0.3%. Higher SUCRA values indicate a greater efficacy in softening stool consistency. These results are illustrated in Figure 3 and Figure 7.

Change in IBS-SSS score: Network evidence graph. This analysis included eight intervention groups and one placebo control group (A: Placebo; B: Polyethylene glycol; C: Lactulose oral solution; E: Linaclotide; H: Traditional Chinese medicine; I: Probiotics; K: Inulin; L: Acupuncture; N: TaVNS). Changes in IBS-SSS scores before and after the intervention served as the primary outcome for constructing the Bayesian network meta-analysis model. The corresponding network structure is illustrated in Figure 3. In Figure 5, model fit diagnostics suggest that the DIC for the fixed-effects model (34.36) is better than that of the random-effects model (35.24), with a smoother posterior distribution. It indicates a more superior fit and greater analytical stability. As can be seen from the assessment of consistency shown in Figure 6, the DIC value of consistency model (35.21) was found to be comparable to that of inconsistency model (34.49). According to this, there is a good agreement between direct and indirect evidence. Therefore, it could be confirmed that the network did not suffer from any significant inconsistency. Further testing for consistency was not required as there were no closed loops among the interventions.

NMA: As shown in Figure 4, the network meta-analysis results show the estimated differences in efficacy of interventions for each comparison and are displayed in each cell. When the values get higher, symptoms become less severe and when the value gets lower, symptoms become more severe. The color gradient reflects these differences: Darker areas mean lesser severity while the lighter their shades, the worse it is. Refer to the above text for the caption.

Efficacy ranking: The SUCRA analysis produced the following rankings: TaVNS (10.0%), probiotics (22.2%), acupuncture (31.5%), inulin (36.5%), polyethylene glycol (44.0%), traditional Chinese medicine (49.4%), linaclotide (80.3%), placebo (85.2%), and lactulose oral solution (91.1%). It is noteworthy that lower SUCRA values indicate a greater improvement in the symptom severity of constipation-predominant IBS. These results are depicted in Figure 3 and Figure 7.

Change in IBS-QOL score: Network evidence graph. The meta-analysis included nine interventions and one placebo control (A: Placebo; B: Polyethylene glycol; C: Lactulose oral solution; E: Linaclotide; H: Traditional Chinese medicine; I: Probiotics; J: Dietary fiber; K: Inulin; L: Acupuncture; N: TaVNS). A Bayesian network meta-analysis was conducted, using changes in IBS-QOL scores as the primary outcome. The network structure is illustrated in Figure 3, while model diagnostics are presented in Figure 5. The fixed-effects model had a DIC value of 47.41, while the random-effects model had a DIC value of 44.21. Thus, the random-effects model performed better as it had a lower overall DIC value, reflective also of its greater robustness. As shown in Figure 6, the consistency assessment indicates that direct and indirect evidence are in good agreement, as the DIC value of the consistency model (44.29) is lower than that of the inconsistency model (47.36). Because of the closed loops on the network, we conducted additional node-splitting analysis to check consistency. According to the results, direct and indirect comparisons were not statistically significantly different (P > 0.05), which confirmed the findings (Figure 8).

Figure 8 Node analysis (consistency check diagram). E: Linaclotide; A: Placebo; H: Traditional Chinese medicine; Crl: Credible interval.

NMA: The findings of the network meta-analysis are displayed in Figure 4. Each cell illustrates the estimated differences in efficacy for a certain comparison, with higher values denoting greater efficacy and lower values denoting lesser efficacy. The darker shades denote superior efficacy while the lighter shades refer to less efficacy. Refer to the above text for the caption.

Efficacy ranking: As per SUCRA analysis, interventions to improve QOL were ranked as follows: TaVNS (91.2%), probiotics (83.5%), inulin (65.6%), traditional Chinese medicine (56.5%), acupuncture (55.2%), linaclotide (50.6%), polyethylene glycol (40.7%), lactulose oral solution (28.5%), dietary fiber (25.2%), placebo (3.1%). A higher SUCRA means a better efficacy for improving the QOL of patients with constipation-predominant IBS. These results are shown in Figure 3 and Figure 7.

Change in adverse events: Network evidence graph. This analysis included nine intervention groups and one placebo control group (A: Placebo; B: Polyethylene glycol; D: Plecanatide; E: Linaclotide; F: Tenapanor; G: Lubiprostone; H: Traditional Chinese medicine; I: Probiotics; N: TaVNS; O: Sensory adaptation training). A Bayesian network meta-analysis was performed with the number of participants experiencing adverse events after intervention as the primary outcome. Figure 3 depicts the network structure. Diagnostics for model fit (Figure 5) show that the DIC value of the fixed-effects model (108.1) is lower than the random-effects model (109.41). This indicates a better fit of the data and more stability during analysis of the fixed-effects model. As shown in Figure 6, the assessment of inconsistency based on DIC value demonstrated that the DIC value of the consistency model, which was recorded as 110.64, was lower than that of the inconsistency model that was recorded to be 110.66. Thus, this shows it is good and that there is a direct agreement between the direct and the indirect evidence. Thus, the evidence is of good quality and there is no major inconsistency in the network analysis results.

NMA: The network meta-analysis results are shown in Figure 4. Each cell presents the estimated differences in efficacy among treatments regarding adverse events. Higher numbers correspond to fewer adverse events, while lower numbers denote greater numbers of adverse events. The colors show this difference, as fewer adverse events are shown as darker colors that slowly become lighter with each increase in adverse event. Refer to the above text for the caption.

Efficacy ranking: The SUCRA analysis ranked interventions based on the incidence of adverse events as follows: Sensory adaptation training (2.7%), taVNS (27.3%), placebo (27.8%), lubiprostone (37.8%), linaclotide (56.2%), polyethylene glycol (56.9%), plecanatide (70.2%), tenapanor (72.0%), probiotics (74.6%), and traditional Chinese medicine (74.6%). Lower SUCRA values indicate a reduced incidence of adverse events following interventions in patients with constipation-predominant IBS. These findings are depicted in Figure 3 and Figure 7.

Model fitting diagnostic diagram

Among the six outcome measures, the DIC values of the random-effects models were lower than those of the fixed-effects models for four outcomes: CSBM (66.10 vs 154.56), SBM (61.08 vs 121.19), BSFS (114.45 vs 251.12), and IBS-QOL (44.21 vs 47.41). This indicates a superior model fit and interpretability of the random-effects approach. Additionally, both the effective number of parameters and residual deviance were smaller under the random-effects models, suggesting greater model stability and a more appropriate representation of between-study heterogeneity. See Figure 5 for the model fitting diagnostic diagram.

Consistency test

The rationality of the network structure was assessed from the standpoint of the consistency and inconsistency models’ consistency assumptions. It was shown that the DIC values of the consistency model were lower than those of the inconsistency model in all outcome measures (CSBM: 66.16 vs 154.45; SBM: 60.92 vs 121.24; BSFS: 114.29 vs 251.12; IBS-QOL: 44.29 vs 47.36; adverse events: 110.64 vs 110.66) indicating a better fit of the consistency model. It is worth mentioning that DIC is used for model comparison and not for testing statistical significance. Refer to Figure 6 for consistency test diagram.

SUCRA

The SUCRA plot indicates the behavior of each intervention’s cumulative probability of being rated to obtain different efficacy outcome levels, with those closer to the upper left corner representing higher treatment rankings. For outcomes where higher values indicate better efficacy, including CSBM, SBM, BSFS, IBS-QOL, a better SUCRA score indicates higher efficacy (Figure 7). One the other hand, for outcomes where smaller values indicate better clinical outcomes, such as IBS-SSS and adverse events, a higher SUCRA indicates less efficient outcomes (Figure 7).

Node analysis

As represented in Figure 8, mean differences with their respective 95% credible intervals for direct and indirect methods, as well as network comparisons of pharmaceutical vs non-pharmaceutical interventions. The overlapping 95% credible intervals and P values > 0.05 indicate a strong agreement between the direct evidence and the indirect evidence. This alignment reinforces the overall consistency and robustness of the results obtained from the Bayesian network meta-analysis, indicating that the findings are reliable and credible.

Supplementary pairwise meta-analysis of BSFS

An exploratory conventional pairwise meta-analysis for the BSFS was additionally conducted to further assess the comparability of direct evidence across placebo/sham-controlled trials. The Figure 9 pooled analysis showed that active interventions significantly improved stool consistency compared with controls [weighted mean difference = 0.75, 95% confidence interval (CI): 0.59-0.90]. Importantly, this direction of effect was consistent with the findings of the network meta-analysis, providing supportive evidence for the robustness of the overall conclusion. In subgroup analyses, plecanatide showed low heterogeneity (I² = 0%), linaclotide showed moderate heterogeneity (I² = 51.3%), whereas taVNS showed higher heterogeneity (I² = 75.3%). Overall, these findings suggest that the direct comparative evidence generally supports a beneficial effect of active interventions on stool form, while the observed heterogeneity is likely attributable to differences in intervention characteristics and study design.

Figure 9 Supplementary conventional pairwise meta-analysis of Bristol stool form scale. PEG: Polyethylene glycol; WMD: Weighted mean difference; CI: Confidence interval; taVNS: Transcutaneous auricular vagus nerve stimulation.

Supplementary pairwise meta-analysis of adverse events

An exploratory conventional pairwise meta-analysis for adverse events was additionally conducted to further assess the comparability of direct evidence across placebo/sham-controlled trials. The Figure 10 pooled analysis showed that active interventions were associated with a slightly higher incidence of adverse events than controls [relative risk (RR) = 1.12, 95%CI: 1.08-1.17], with low overall heterogeneity (I² = 14.8%, P = 0.263). Importantly, this result was generally consistent with the safety profile observed in the network meta-analysis, indicating that differences in adverse-event risk across interventions were directionally stable in the direct comparisons. In subgroup analyses, plecanatide (RR = 1.28, 95%CI: 1.11-1.48), linaclotide (RR = 1.11, 95%CI: 1.05-1.17), and tenapanor (RR = 1.35, 95%CI: 1.11-1.65) showed increased risks of adverse events compared with controls, whereas polyethylene glycol and lubiprostone did not show statistically significant differences. Overall, these findings provide supportive direct evidence for the comparative safety results of the network meta-analysis and help explain the observed variation in treatment tolerability.

Figure 10  Supplementary conventional pairwise meta-analysis of adverse events. PEG: Polyethylene glycol; RR: Relative risk; CI: Confidence interval; taVNS: Transcutaneous auricular vagus nerve stimulation.

DISCUSSION

Constipation is the primary clinical manifestation of IBS-C and represents a significant concern for affected patients. This often occurs with recurrent pain in the abdomen, with change in how often the stools are passed and pain with passing them[3]. Persistent symptoms of IBS-C can negatively impact QOL in the patient as it affects the emotional aspect and increases the symptoms. The cause of IBS-C is complex, with multiple mechanisms believed to be responsible. These mechanisms include abnormal motility in the gastrointestinal tract, psychological factors, altered brain–gut interaction, disturbance of the gut microbiota, visceral hypersensitivity, genetically determined gut microbiota, and intestinal inflammation[4,5]. There are many treatments available, which are mainly pharmacological and non-pharmacological therapy[7,8]. The effectiveness of these traditional treatments varies among the studies. However, most of the studies involve a single intervention through RCTs. Moreover, they do not compare the relative effectiveness of different treatments. Therefore, an effective treatment for improving IBS-C symptoms and QOL is unknown. In this research, the authors will make use of a Bayesian network meta-analysis to compare various treatment options for IBS-C.

Findings from a Bayesian network meta-analysis assessing the relative effectiveness of various treatment methods show acupuncture, tenapanor, as well as taVNS are the most effective in increasing CSBMs. Tenapanor, taVNS, and polyethylene glycol were effective for improving spontaneous bowel movement. In terms of stool softening, taVNS, traditional Chinese medicine, and tenapanor were the most effective. The effective treatment for alleviation of symptoms in patients with IBS-C was evaluated in a systematic review. According to this review, the most effective treatment for constipation-predominant IBS was taVNS. This was followed by probiotics and acupuncture. When it comes to improvements in QOL, taVNS, probiotics, and inulin outperformed others while sensory adaptation training, taVNS, and placebo had the least adverse impact. The diagnostics for model fitting and consistency analysis indicate the reliability of the network estimates. However, small-study effects and publication bias cannot be ruled out as some comparisons are based on few trials. According to GRADE assessment, most outcomes were of moderate-quality evidence, while the QOL evidence was rated as high quality. To summarize, the most promising therapies for managing constipation-predominant IBS are taVNS, tenapanor, acupuncture, and probiotics.

Tenapanor is a small-molecule inhibitor of the sodium-hydrogen exchanger type 3 (NHE3). Tenapanor lowers sodium absorption in the intestine by inhibiting this exchanger. Thus, it can reduce levels of sodium. This device causes less sodium absorption but greater secretion of water into the intestinal lumen. Due to actions such modifications to the composition of intestinal contents tenapanor increases transit speed of intestinal contents with acceleration of passage and expulsion stool consistency shape mobility and number indirectly by pharmacological actions quicker intestinal transit[49,50]. In addition, tenapanor reduces intracellular potential of hydrogen, increases actin assembly in the cytoskeleton, and reduces paracellular permeability and enhances intestinal barrier function[51]. Animal studies show that tenapanor may also reduce pain by decreasing visceral sensitivity[51,52]. According to our findings, tenapanor has a positive impact on CSBM, SBM, and BSFS scores (Figure 7). These results indicate that tenapanor is effective for the treatment of constipation and abdominal pain in IBS-C patients. However, limited clinical studies have leveraged inhibitors of NHE3 for treating IBS-C, indicating the necessity for a rise in such clinical trials to ascertain their therapeutic benefits in this condition. The non-invasive treatment of IBS-C by taVNS does not pose the same risks of implanted vagus nerve stimulation devices and is generally considered a safe procedure. Mediating effects can occur through brain-gut-microbiome interaction pathway due to the suggested efficacy of therapeutic effects[53]. TaVNS effectively alleviates symptoms such as constipation and abdominal pain by stimulating the vagus nerve, which enhances intestinal motility and improves bowel habits[54]. Relevant animal studies indicate that taVNS softens stool and promotes defecation by increasing fecal pellet count, wet weight, and moisture content in IBS-C model mice, thereby improving symptoms of constipation[53]. Furthermore, the taVNS-induced enhancement of vagal tone reduces visceral hypersensitivity, alleviating abdominal pain[53]. Studies have also demonstrated that taVNS increases the abundance of bifidobacteria, an effect linked to the brain-gut-microbiota interaction pathway[53]. Reduced levels of bifidobacteria are known to impair intestinal barrier function[55,56], and taVNS appears to maintain intestinal barrier integrity by promoting the growth of bifidobacteria. Our study shows that taVNS yields favorable outcomes across multiple measures, including CSBM, SBM, BSFS, IBS-SSS, IBS-QOL, and adverse events (Figure 7). Therefore, taVNS is considered effective in improving constipation, abdominal pain, and QOL in patients with IBS-C. Acupuncture is a therapeutic modality for IBS-C that exerts its effects through the stimulation of specific acupoints. It enhances gastrointestinal motility by modulating the vagus nerve, the myenteric plexus, enteric ganglion cells, interstitial cells of Cajal, as well as neurotransmitters and their receptors[57]. For instance, electroacupuncture at the Tianshu and Shangjuxu acupoints has been demonstrated to enhance intestinal motility and promote defecation in mice with constipation models[58], while stimulation of the Zusanli acupoint improves gastrointestinal motility through the activation of vagal pathways[59]. Additionally, acupuncture regulates levels of vasoactive intestinal peptide, thereby further enhancing gastrointestinal function[60]. Regarding analgesia, acupuncture primarily promotes the release of endogenous opioid peptides, such as enkephalins, which contribute to pain relief and the alleviation of abdominal discomfort[61]. Overall, acupuncture has shown to be effective in treating IBS-C. In line with these processes, our results showed that acupuncture had notable effects on key outcome measures such as CSBM and IBS-SSS scores. Probiotics are preparations that contain live microorganisms which confer health benefits by modifying the host microbiota[62]. Probiotics typically contain various Bifidobacterium and Lactobacillus species as a treatment for IBS-C[63]. The regulation of intestinal motility is dependent on multiple interactions between the gut environment, the immune system, the enteric nervous system and the central nervous system. When this balance is disturbed, it may cause constipation[63]. Bifidobacterium and Lactobacillus species can aid with constipation as they restore microbial balance in our gut leading to enhanced intestinal motility and secretory function[64]. According to our research, probiotics can significantly reduce symptoms and improve the QOL of IBS-C patients. The evidence gathered suggests that probiotics could help to alleviate constipation-related symptoms in people with IBS-C. It is likely that they act through the modification and improvement of the intestinal environment, leading to a marked relief of symptoms. Analyzing the cost-effectiveness of probiotics and prebiotics, the overall wellbeing and QOL of the patients suffering from the disease can improve a lot.

This study has many strengths that improve its credibility and the depth of analysis. The method used is first, a Bayesian network meta-analysis technique to review the relative efficacy of different interventions for constipation-predominant IBS. Incorporating direct evidence and indirect evidence from 39 RCTs, the study effectively summarizes the treatment of this condition. Further, it offers a sound and firm statement on the treatment. Secondly, the broad range of treatment options incorporated is another notable aspect of the analysis owing to the wide intervention types. The evaluation goes beyond simply assessing stool-related outcomes to include clinically relevant assessments for a range of conditions. These comprise not only CSBMs and spontaneous bowel movements but also the BSFS score, IBS symptom severity scale score and IBS-QOL score. Moreover, we take into account adverse events of the interventions. By incorporating data from various forms of research, the organization achieves an extensive understanding of the safety and efficacy of the discussed vaccine or drug. The analysis was consistent with PRISMA-NMA statement and involved extensive risk-of-bias assessment, model-fitting checks and consistency checks. Supplementary traditional pairwise meta-analyses for BSFS and adverse events were done; they provided supportive direct evidence and further enhanced the interpretation of the network meta-analysis results. That said, there are a few limitations. Initially, some interventions, such as taVNS and acupuncture, were ranked highly. However, these interventions were supported by very few trials with small sample sizes which probably inflated the SUCRA rankings. Subsequently, this very probably overestimated the comparative efficacy of the interventions. In addition, blinding was difficult to implement for the various non-pharmacological trials, especially the ones evaluating acupuncture and taVNS. This raised the risk of performance bias and placebo effects. In addition, the duration of treatment was different in the studies and often was short, limiting inference on long-term efficacy and safety in this chronic disorder. Some of the treatment rankings should be viewed with caution due to the absence of closed loops in a number of networks or the presence of a dominance of indirect comparisons. In the fifth place, the control conditions were not fully equivalent across trials, and the use of sham procedures, placebo tablets or standard care may have resulted in clinical heterogeneity and impacted the transitivity assumption of the network meta-analysis.

CONCLUSION

To conclude, pharmacological and non-pharmacological interventions ease constipation and abdominal discomfort in IBS-C patients and hence improve their QOL significantly. Therapies like taVNS, tenapanor, acupuncture, and probiotics exhibit especially encouraging therapeutic outcomes. TaVNS as a non-pharmacological approach could be beneficial and safe while the pharmacological approach that is tenapanor has the most potent effect but with a weaker safety aspect. Thus, patients can choose one of the four interventions based on their therapeutic needs. Not only with the same mechanism of action as tenapanor, taVNS may soften stool consistency, reduce visceral hypersensitivity, and enhance intestinal barrier function. It is suggested that combined pharmacological and non-pharmacological clinical treatments may achieve better treatment outcomes. This integrated approach has a lot of potential for clinical applications and development. Consequently, more high-quality, multicenter RCTs are needed to confirm these results and to assess the effect of integrated treatment approaches.

ACKNOWLEDGEMENTS

We would like to thank the researchers and study participants for their contributions.