Knowledge Mapping of Clinical Evidence and Mechanistic Research on Tongxie Yaofang for Ulcerative Colitis: A Bibliometric Analysis

Summary

This bibliometric study maps English-language Web of Science Core Collection evidence on Tongxie Yaofang (TXYF) for ulcerative colitis and distinguishes clinical, mechanistic, and translational research streams. The analysis identifies a rapidly expanding, China-led field, with network pharmacology, animal models, microbiota, and inflammatory signaling research outpacing standardized clinical endpoint studies.

Abstract

To systematically map the knowledge landscape of clinical evidence and mechanistic research on Tongxie Yaofang (TXYF) for ulcerative colitis (UC), we conducted a bibliometric and knowledge-graph analysis of English-language articles and reviews retrieved from the Web of Science Core Collection through 31 December 2025. After screening, 133 records were analyzed for publication trends, collaboration networks, journal distribution, co-citation clusters, keyword co-occurrence, and clinical-versus-mechanistic thematic structure. Operational criteria were used to classify records as clinical evidence, mechanistic research, or translational/bridge studies. The field expanded markedly after 2020, with Chinese institutions dominating, but international and cross-team collaboration remained limited. Mechanistic research showed strong activity around nuclear factor kappa B (NF-κB)/NOD-like receptor protein 3 (NLRP3), mitogen-activated protein kinase (MAPK)/AKT (protein kinase B), gut microbiota, short-chain fatty-acid metabolism, intestinal barrier repair, macrophage polarization, and computational pharmacology. Clinical evidence in the English-language WoSCC corpus was comparatively sparse and often relied on broad symptom or disease-activity outcomes rather than standardized patient-reported, endoscopic, histologic, and biomarker endpoints. These findings indicate a translational disconnect between mechanistic discovery and clinical evaluation, and they support the development of mechanism-informed, multicenter clinical studies with predefined biological readouts.

Introduction

Ulcerative colitis (UC) is a chronic inflammatory disease of the colon shaped by environmental, genetic, microbial, epithelial-barrier, and immune factors¹. Current management has moved beyond symptom relief toward treat-to-target strategies that emphasize mucosal healing, durable remission, and prevention of complications. Although 5-aminosalicylic acid, corticosteroids, immunomodulators, biologics, small-molecule inhibitors, and interleukin-targeted therapies have improved outcomes for many patients, primary non-response, secondary loss of response, infection risk, cost, and treatment burden remain clinically important challenges^2,3,4. These unmet needs have sustained interest in multi-component adjunctive approaches that may regulate inflammation, barrier injury, microbiota disturbance, and immune imbalance through coordinated mechanisms.

Tongxie Yaofang (TXYF), a classical traditional Chinese medicine (TCM) formula associated with the principle of restraining wood and supporting earth, is commonly discussed for diarrhea and abdominal pain related to liver-stagnation and spleen-deficiency patterns⁵. In contemporary clinical practice, it has become a frequently prescribed treatment for ulcerative colitis and related inflammatory bowel diseases. Recent scientific research on Tongxieyao Formula has progressed from clinical empirical validation to in-depth molecular mechanism analysis^6,7. Modern pharmacological studies reveal that this formula does not act as a single anti-inflammatory agent but exerts effects through multiple pathways⁸. In terms of inflammation regulation, research demonstrates its significant ability to inhibit overactivation of relevant signaling pathways and inflammatory microsomes. Regarding intestinal mucosal barrier repair, the formula upregulates tight junction protein expression and reduces intestinal permeability⁹. Additionally, the restoration of gut microbiota balance and modulation of the neuroendocrine-immune network of the gut-brain axis are emerging as novel focal points for elucidating the therapeutic mechanisms of Tongxieyao Formula. Latest basic experiments indicate that the formula improves immune dysregulation by regulating tryptophan metabolism and receptor activity, providing robust evidence supporting the scientific rationale for its therapeutic efficacy in ulcerative colitis¹⁰.

Despite increasing publication activity, the evidence landscape remains fragmented. Current studies predominantly adopt single-perspective approaches: either focusing on clinical efficacy indicators through small-sample randomized controlled trials or conducting purely basic experiments investigating specific signaling pathway interventions^11,12. There exists a significant disconnect between clinical evidence and mechanistic studies, with insufficient integration of their interactive relationships. Additionally, systematic analysis has yet to be conducted regarding core research team distribution, cross-institutional collaboration intensity, the foundational role of seminal literature, and temporal evolution patterns of research themes. For instance, while keywords such as network pharmacology, gut microbiota, and signaling pathways show high frequency in keyword clustering, quantitative big data analysis remains lacking to elucidate how these research hotspots evolved from traditional symptom evaluation methods and determine their future developmental trajectories¹³. In UC-related research, TXYF has been investigated through reviews, network pharmacology, molecular docking, disease models, microbiota analyses, transcriptomics, macrophage-polarization experiments, and fecal microbiota transplantation validation^{2,14,15,16,17,18,19,20}. Current evidence suggests that TXYF research has shifted from empirical symptom-oriented interpretation toward hypotheses involving nuclear factor kappa B (NF-κB)/NOD-like receptor protein 3 (NLRP3), mitogen-activated protein kinase (MAPK)/AKT (protein kinase B), IL-10RA/NF-kB signaling, epithelial-barrier protection, microbial remodeling, and mitophagy. Clinical studies and protocols tend to emphasize symptom relief, disease activity, recurrence, safety, and TCM syndrome differentiation, whereas mechanistic studies frequently examine computational target prediction, inflammatory signaling, animal or cellular validation, microbiota modulation, and epithelial-barrier repair^{2,14,15,16,17,18,19,20}. These streams are not yet well integrated: mechanistic findings are seldom prospectively embedded into clinical trial endpoints, and clinical studies rarely include standardized biomarker panels capable of testing the biological pathways proposed by preclinical work.

This study therefore, aimed to map the English-language WoSCC research landscape on TXYF for UC, identify key contributors and knowledge structures, and compare the maturity and translational connectivity of clinical evidence and mechanistic research. The goal was not to claim exhaustive coverage of all Chinese and international TXYF studies, but to characterize the internationally indexed English-language corpus and to identify where future work should strengthen reproducibility, collaboration, endpoint standardization, and mechanism-to-clinic study design.

Protocol

Study design

This study used a bibliometric and knowledge-graph design focused on clinical evidence and mechanistic research on TXYF for UC. The workflow is shown in Figure 1 and includes six linked stages: database retrieval, duplicate removal, title/abstract screening, full-text eligibility assessment, structured data extraction, and visualization/synthesis. The final analytical sample consisted of 133 English-language articles and reviews. Table 1 summarizes the search strategy and eligibility criteria, and the Table of Materials lists the bibliographic platform, exported data type, software tools, and reproducibility notes.

The design integrates bibliographic metadata, author and institutional information, countries/regions, source journals, cited references, keywords, abstracts, and manually coded thematic attributes. This approach allows the manuscript to describe publication growth and collaboration patterns while also examining whether clinical evidence and mechanistic research are developing as connected translational streams or as parallel, weakly linked bodies of work.

Data source and search strategy

The data source was the Web of Science Core Collection (WoSCC), selected because it provides standardized citation records, author keywords, institutional affiliations, country/region fields, cited references, and export formats compatible with VOSviewer and CiteSpace. The search covered records available through 31 December 2025. Both retrieval and data export were performed at a single time point to reduce instability caused by database updates.

The exact topic-search query executed in WoSCC was:

TS=(("Tongxie Yaofang" OR "Tong Xie Yao Fang" OR "Tongxie-Yaofang" OR "Tongxie formula" OR "TXYF") AND ("ulcerative colitis" OR "UC"))

Search results were limited to English-language articles and reviews. Conference abstracts, editorials, letters, corrections, book reviews, and records without sufficient bibliographic information were excluded. These restrictions were applied to maintain metadata consistency for network construction; however, the English-only WoSCC design is acknowledged in the Discussion as an important retrieval limitation rather than evidence that Chinese-language clinical work is absent.

For reproducibility, all eligible WoSCC records were exported as full records with cited references. Restrictive descriptors such as clinical, mechanism, network pharmacology, signaling pathway, microbiota, or trial were not added to the initial query. After retrieval, records were manually screened and coded by title, abstract, and, when necessary, full text. Geographic fields were standardized before visualization: affiliation records from Taiwan, Hong Kong, and Macao were treated as China regional records for country-level counting and were not interpreted as separate sovereign countries.

Eligibility criteria and literature screening

To ensure high relevance between the included literature and the research topic, and to enhance the accuracy and interpretability of subsequent knowledge graph analysis results, this study established clear inclusion and exclusion criteria prior to formal analysis and conducted staged screening of retrieved literature using standardized procedures. The overall screening process comprised four steps: document deduplication, initial screening of titles and abstracts, full-text re-screening, and final inclusion determination (see Figure 1 for detailed workflow). Following systematic retrieval and progressive screening, a total of 133 articles were ultimately included in the bibliometric analysis of this study.

Regarding inclusion criteria, this study primarily selected literature meeting the following requirements: First, the research topic must explicitly involve Tongxie Yaofang and ulcerative colitis, meaning both Tongxie Yaofang and ulcerative colitis serve as core research subjects rather than being merely mentioned in background or discussion sections. Second, the literature type must be strictly limited to formally published articles or reviews to ensure completeness of bibliographic information and citation data. Third, the language of the literature must be restricted to English to enhance data source consistency and meet the fundamental requirements of international bibliometric analysis. Fourth, the literature records should contain relatively complete metadata information, including title, author, source journal, abstract, keywords, and citation details, enabling subsequent collaborative network analysis, keyword co-occurrence analysis, and co-citation analysis.

The exclusion criteria primarily include the following categories: First, duplicate literature records; Second, studies not directly related to the themes of Tongxie Yaofang (Pain-Relieving Formula) or ulcerative colitis, such as those solely discussing other traditional Chinese medicine formulations or focusing on general inflammatory bowel diseases without explicit reference to Tongxie Yaofang; Third, conference abstracts, editorials, reader letters, errata, book reviews, and other informal academic documents lacking complete research structures; Fourth, records with incomplete bibliographic information or unavailable valid abstracts and basic citation details; Fifth, literature where Tongxie Yaofang is not the primary research subject despite search term matches, or ulcerative colitis is not the main disease focus. Through these criteria, the study aims to minimize interference from non-target literature and low-information-content materials on research outcomes.

The literature screening process was conducted using a hierarchical progressive approach. First, all original records imported from databases were transferred to literature management software and spreadsheets, where title, author, year, and journal information were verified, and duplicate records were removed. Subsequently, the initial screening phase for titles and abstracts was initiated, with each record's title and abstract independently reviewed against predefined inclusion/exclusion criteria to preliminarily exclude documents that clearly did not meet the thematic requirements. Following this, retained documents underwent full-text re-examination to assess whether their research subjects, content, and core themes genuinely centered on "Pain Relief Formula for Treating Ulcerative Colitis," while also verifying compliance with literature type classification and metadata completeness requirements.

Operational definitions were applied before coding. Clinical evidence studies were defined as records whose primary aim was to evaluate patient-level efficacy, safety, recurrence, symptom change, disease activity, quality of life, TCM syndrome outcomes, inflammatory markers, endoscopic outcomes, histologic outcomes, or other clinical endpoints. Mechanistic studies were defined as records whose primary aim was to explain TXYF action through network pharmacology, molecular docking, molecular dynamics, target prediction, animal models, cellular validation, microbiota analysis, multi-omics, inflammatory signaling, immune regulation, epithelial-barrier repair, oxidative stress, apoptosis, mitophagy, or related pathways. Translational or bridge studies were defined as records that connected clinical evaluation with mechanistic measurement, evidence synthesis, biomarker mapping, or experimental validation in a way that could inform clinical trial design.

Data extraction and thematic classification

After completing literature screening and ultimately selecting 133 target publications, this study conducted structured extraction and organization of relevant bibliographic information to establish the data foundation for subsequent knowledge graph analysis. The data extraction primarily encompassed three components: basic literature information, citation details, and thematic attribute data. Basic literature information included title, author, publication year, source journal, country/region, institution, document type, and abstract. Citation data comprised citation frequency, references, keywords, and author keywords. Thematic attribute information was utilized for research content stratification and comparative analysis, covering research subjects, methodologies, models, and core thematic directions.

Thematic attributes were extracted in addition to standard bibliometric fields. For each record, the authors coded the dominant study type, principal model or method, main clinical or mechanistic focus, and whether the study directly linked mechanism-level findings to clinical endpoints. Ambiguous records were classified according to the primary objective stated in the title, abstract, and methods rather than by brief speculative discussion statements.

Clinical endpoint coding was refined into patient-reported outcomes (PROs) and objective endpoints. PROs included diarrhea, abdominal pain, stool frequency, global symptom improvement, quality of life, and TCM syndrome scores. Objective endpoints included disease activity indices, inflammatory markers, endoscopic or mucosal-healing measures, histology, recurrence, safety events, microbiota profiles, epithelial-barrier indicators, cytokines, immune-cell phenotypes, and pathway biomarkers.

Studies combining network pharmacology prediction with animal, cellular, microbiota, or molecular validation were coded as mechanistic studies if mechanism elucidation was the primary aim. Studies focused on efficacy observation but mentioning possible mechanisms only in the Discussion were coded as clinical evidence studies. Protocols or evidence-synthesis records that explicitly attempted to integrate clinical-effect evaluation with mechanism exploration were coded as translational/bridge studies.

Bibliometric analysis and visualization

Bibliometric analysis was conducted with spreadsheet software, VOSviewer, and CiteSpace; the Table of Materials reports the version/identifier fields and flags items requiring author confirmation from the original analysis logs. Spreadsheet software was used for duplicate checks, descriptive statistics, annual publication counts, research-type classification, and preparation of source tables. VOSviewer was used to build co-authorship, institutional, country/region, journal, and keyword co-occurrence networks. CiteSpace was used for co-citation analysis, cluster detection, timeline interpretation, and burst-keyword detection.

Co-citation analysis was used to identify knowledge-base modules repeatedly cited by TXYF-UC studies. Clusters were manually interpreted by reading representative titles, abstracts, and cited-reference contexts, with attention to UC inflammatory mechanisms, TCM compound intervention evidence, gut microbiota and mucosal barrier studies, and computational pharmacology methods.

Keyword preprocessing included synonym consolidation, spelling standardization, case normalization, removal of non-informative terms, and harmonization of equivalent biological concepts. Keyword co-occurrence and temporal analyses were then used to identify major research hotspots and emerging frontiers, including inflammatory signaling, TCM clinical evidence, microbiota-metabolism interactions, computational pharmacology, and barrier-immunity regulation.

Clinical-versus-mechanistic mapping combined the thematic classification with keyword clusters and coded endpoint domains. The comparison examined six domains: symptom control, inflammatory response, intestinal barrier repair, gut microbiota, immune regulation, and disease progression.

Results

Literature screening and overall publication characteristics

Based on established search strategies and screening criteria, this study ultimately included 133 English-language articles related to the use of Tongxieyao Formula in the treatment of ulcerative colitis, forming the foundational sample for subsequent knowledge graph analysis. Overall, this field has evolved from early sporadic explorations into a research direction with considerable scale and continuity, exhibiting ...

Discussion

This study provides a structured map of the English-language WoSCC literature on TXYF for UC and shows that the field has expanded rapidly since 2020. The mapped corpus is China-led, institutionally concentrated, and increasingly mechanism-oriented. Clinical evidence remains visible but comparatively less standardized, whereas mechanistic work has become denser around NF-kB/NLRP3, MAPK/AKT, IL-10RA/NF-kB signaling, microbiota remodeling, epithelial-barrier protection, macrophage polarization, mitophagy, and computational...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
CiteSpace	Chaomei Chen, Drexel University	https://citespace.podia.com/	Co-citation analysis, cluster analysis, timeline visualization, and burst keyword detection.
Microsoft Excel	Microsoft Corporation	https://www.microsoft.com/en-in/microsoft-365/Excel	Data cleaning, duplicate checking, descriptive statistics, annual publication counts, research-type classification, and table preparation.
VOSviewer	Centre for Science and Technology Studies, Leiden University	https://www.vosviewer.com/	Construction and visualization of author, institution, country, journal, and keyword co-occurrence networks.
Web of Science Core Collection	Clarivate Analytics	https://clarivate.com/academia-government/scientific-and-academic-research/research-discovery-and-referencing/web-of-science/web-of-science-core-collection/	Primary bibliographic database used to retrieve English-language articles and reviews on Tongxie Yaofang and ulcerative colitis.