Global research landscape and advancements on the links between the gut microbiome and insulin resistance: hot issues, trends, future directions, and bibliometric analysis

The data source for this bibliometric study was Scopus. The bibliographic database Scopus was launched in 2004 and is owned by Elsevier (www.elsevier.com). Scopus is one of the most comprehensive collections of peer-reviewed literature; Scopus includes a wide range of disciplines, including the humanities, social sciences, technology, and health. There are approximately 22,800 current titles in the Scopus database from more than 7,000 different publishers worldwide [26]. Scopus offers broad journal coverage across biomedical and interdisciplinary fields, with strong indexing of publications related to the gut microbiome and IR, and the majority of publications indexed in Web of Science and PubMed are indexed in Scopus [27‐29]. While it is common for systematic reviews and meta-analyses to utilize multiple databases to ensure a comprehensive literature search [30, 31], the majority of bibliometric studies are generally based on one database. This is often due to practical considerations such as data format differences and the complexity of integrating data from various sources.

We used the “Advanced search” tool of the Scopus online database. After we inserted the required keywords, we located the research status pertinent to the gut microbiota and IR (from January 2015 to December 31, 2024). The retrieval and export of publications should be completed within one day to avoid bias from continuously updated databases (as of June 15, 2025). We have updated the citation counts to reflect the latest data, as of August 22, 2025. The following search steps used synonyms for the gut microbiota and IR.

Step 1: To accomplish the objectives of this bibliometric study, terms associated with the microbiota were chosen from the literature [17, 18, 20‐23, 32‐38]. and input into the Scopus database in ‘Article Title or/and Abstract’.

Step 2: To identify relevant studies, we limited our search to publications whose titles contained ‘IR’ or closely related terms. These terms were derived from previous reviews and meta-analyses [39‐41] and included both ‘insulin resistance’ and ‘insulin sensitivity’ in the ‘Article Title.

The overall search query was as follows: TITLE-ABS(“16S rRNA profiling” OR “16S rRNA sequencing” OR “Akkermansia” OR “Bacteroides” OR “Bifidobacterium” OR “Escherichia coli” OR “FMT” OR “colonic flora” OR “colonic microbiome” OR “colonic microbiota” OR “colonic microflora” OR “commensal microbiota” OR “digestive flora” OR “digestive microbiome” OR “digestive microbiota” OR “digestive microflora” OR “dysbiosis” OR “enteric bacteria” OR “enteric flora” OR “enteric microbiome” OR “enteric microbiota” OR “enteric microflora” OR “fecal flora” OR “fecal microbiome” OR “fecal microbiota” OR “fecal microflora” OR “flora” OR “gastric flora” OR “gastric microflora” OR “gastric microbiome” OR “gastric microbiota” OR “gastrointestinal flora” OR “gastrointestinal microflora” OR “gastrointestinal microbial community” OR “gastrointestinal microbiomes” OR “gastrointestinal microbiota” OR “gut bacteria” OR “gut dysbiosis” OR “gut flora” OR “gut microbiome” OR “gut microbiota” OR “gut microorganisms” OR “gut-derived microbiota” OR “intestinal bacteria” OR “intestinal ecosystem” OR “intestinal flora” OR “intestinal microbiome” OR “intestinal microbiota” OR “intestinal symbionts” OR “Lactobacillus” OR “metagenomic sequencing” OR “microbial community” OR “microbial composition” OR “microbial flora” OR “microbial metabolite” OR “microbial metabolite regulation” OR “microbiome” OR “microbiota” OR “microbiota diversity” OR “microflora” OR “probiotic” OR “Saccharomyces” OR “SCFAs” OR “short-chain fatty acids” OR “shotgun metagenomics”) AND ((TITLE(insulin AND resist*) OR TITLE(insulin AND sensitiv*)) AND PUBYEAR > 2014 AND PUBYEAR < 2025) AND (EXCLUDE (DOCTYPE,“er”)).

The raw data were analysed in Microsoft Excel, which included both descriptive statistics and bibliometric mapping. Descriptive statistics, including frequencies and percentages, were calculated for document types, publication counts, journal impact factors, geographical distributions, funding sources, institutional affiliations, and citation metrics of prominent publications. VOSviewer software (version 1.6.20) was subsequently employed for bibliometric mapping, visualizing term cooccurrences within titles and abstracts, generating overlay visualizations, and mapping international collaborative networks [44, 45]. These visualizations represented the prevalent terms in titles or abstracts, emphasizing the main research topics identified in the documents retrieved. Moreover, the mapping included evaluations of international collaboration. In all visual maps, node size indicates frequency, node color denotes relevance, and the distance between nodes signifies the strength of their association. With VOSviewer, you can generate an overlay visualization where recently used author terms are highlighted in yellow. This visualization is created on the basis of the frequency of occurrence and the average number of publications per year.

The search strategy resulted in 584 studies pertaining to the gut microbiota and IR. There were five categories of publications, the majority of which were research articles (n = 480, 82.19%) and reviews (n = 82, 14.04%).

The publication trends, as illustrated in Fig. 1, revealed a clear three-phase growth trajectory. Initially, the field showed limited activity, with fewer than 20 annual publications in 2015. This was followed by a steady increase between 2016 and 2018. A significant acceleration occurred from 2019 to 2024, during which 78.25% of the total publications (n = 457) were produced. Linear regression analysis further confirmed a strong and statistically significant upwards trend in annual publication output over time (R² = 0.7705, p = 0.0002), underscoring the rapidly growing global interest in the relationship between the gut microbiota and IR.

Table 1 presents a ranking of the ten leading countries on the basis of the total number of publications. Specifically, the data indicate that China leads in this area with 254 publications (43.49%), followed by the United States with 96 publications (16.44%), Canada with 44 publications (7.53%), and Germany with 27 publications (4.62%). These figures highlight the prominent contributions of these countries to the growing body of literature on the gut microbiota and IR. Figure 2 shows a study of international collaboration among various countries. Literature from 27 countries was chosen, with a minimum of 5 publications from each country. The United States and China were located at the center of this network map. In the coauthorship map, China and the United States function as network hubs, bridging multiple regional communities (East Asia, North America, Europe). The dense interhub link suggests efficient knowledge diffusion and resource flow across continents.

Table 2 reveals the major institutions leading research on the connection between the gut microbiota and IR. The Ministry of Education of the People’s Republic of China (China) holds the top position, with 18 publications (3.08%), followed by Zhejiang University (China), with 16 publications (2.74%).

A clear hierarchy of funding sources is evident in gut microbiota and IR research. At the forefront is the National Natural Science Foundation of China (China), which supported the largest share of publications (n = 117; 20.03%). This is followed by the National Institutes of Health (United States), with 29 publications (4.97%), and the National Key Research and Development Program of China (China), with 25 publications (4.28%). The National Institute of Diabetes and Digestive and Kidney Diseases (United States) also played a noteworthy role, funding 21 publications (3.60%) (Table 3).

Table 4 provides the rankings of the top journals by the number of publications in the field. Nutrients (Impact Factor 2023: 4.8) contributed 30 publications, followed by Food and Function (n = 21; IF 2023: 5.1) and the Journal of Functional Foods (n = 16; IF 2023: 3.8). Altogether, these three journals account for 67 publications, representing a significant share of the field’s output. Overall, the top ten journals combined published (n = 146) articles, comprising approximately 25.0% of all the documents.

Table 5 ranks the publications by citation count, from highest to lowest. The top ten most-cited articles have a wide range of citations, from 432 to 1752 [46‐55]. Six of these exceed 500 citations.

A total of 13,483 title-and-abstract terms were extracted from the 584 papers; after applying a ≥ 15-occurrence threshold in VOSviewer (v1.6.20), the software retained 251 high-frequency terms, which formed the dataset used to generate the term-cooccurrence map. We selected a ≥ 15-occurrence threshold to balance network interpretability (reducing sparsity/noise) with coverage of salient terminology. In sensitivity checks, a threshold of 10 increased the term count and edge density but merged peripheral topics, whereas 20 pruned informative mid-frequency terms. At 15, the three-cluster structure remained stable with improved label legibility. The software modularity algorithm automatically partitioned these items into three color-coded clusters (Fig. 3), whose thematic identity was inferred by inspecting the most frequent terms in each cluster.

The relative circle sizes in Fig. 3 reflect each term’s raw occurrence count, highlighting the prominence of these keywords and confirming the three clusters as the current research hotspots linking the gut microbiota with IR.

An overlay visualization map highlights the temporal trends of these terms, indicating the evolution of research in this field (Fig. 4). The effects of inflammatory and metabolic signalling should be investigated in the future. To aid synthesis, Table 6 concisely summarizes the three data-driven clusters, listing representative keywords, exemplar papers [46, 49, 51‐54, 56], predominant study designs, recent trends (2019–2024), and actionable gaps/opportunities.

Our literature mapping into three translationally connected clusters—HFD gut–liver axis (C1), patient-centric epidemiology/trials (C2), and inflammatory/metabolic signalling (C3; in Table 6)—justifies a conservative, adjunctive clinical approach and a research agenda. In the clinic, clinicians should proceed with guideline-directed management of insulin resistance and comorbidities but counsel high-fibre, plant-rich eating styles consistent with SCFA- and bile acid–common mechanism signals in C3 and HFD-modulating signals in C1. Currently, strain-specific probiotics, next-generation contenders (e.g., Akkermansia muciniphila), and fecal microbiota transplantation should be restricted until efficacy, dose, durability, and safety are established. In accordance with the transition identified in C2, subsequent research should be weighted in favour of mechanism-anchored randomized trials with biomarker-guided stratification (e.g., baseline HOMA-IR, enterotypes), standard reporting of intervention (diet composition and schedule; strain identity and viability; donor screening), longer follow-up to validate durability, and incorporation of diverse populations to increase generalizability. Pragmatic comparative-effectiveness trials alongside routine care, in combination with multiomic cohorts and planned mediated analyses, will translate evidence in bibliometric into actionable knowledge. Owing to database coverage, keywords chosen, and citation time lags, hotspots identified should be interpreted best as indicative signals to inform trial design and research priorities, not as definitive clinical recommendations.

To our knowledge, using the Scopus database, this bibliometric study is the first to analyse the characteristics of the gut microbiota and IR-related articles published from 2015 to 2024. However, certain limitations should be mentioned. First, the current study used the Scopus database, one of the most widely used resources for researchers in bibliometric analysis. Although Scopus offers broad, interdisciplinary coverage and robust author keyword indexing—assets for term-mapping analyses—it differs from Web of Science (WoS) in terms of journal selection, historical depth, and citation graph construction. Prior comparisons indicate partially overlapping but nonidentical coverage across WoS and Scopus; hence, absolute counts and rankings can vary by database. Given the practical constraints in harmonizing exports across platforms for VOSviewer mapping, we base our analysis on Scopus, acknowledging that multidatabase triangulation could further increase completeness. Second, the publications that have been published for a considerable amount of time are the ones that are cited the most frequently. As a result, some recently released high-quality studies cannot be assessed by us since they do not receive sufficient citations. Third, the dataset is predominantly English-language, reflecting Scopus coverage and the search strategy; non-English outputs may be underrepresented, potentially attenuating signals from some regions and journals. Fourth, false positive and false negative results are inherent challenges, regardless of the precision of the search strategy. However, we believe that the occurrence of such results was minimal and had little to no impact on the overall accuracy of our study’s findings. By focusing the search strategy on article titles, we increased the accuracy and reduced the likelihood of false positives. While one could argue against this approach, we felt it would be unjust to include every article containing the keyword “insulin resistance” in the title, abstract, or keywords. Therefore, restricting the search query to the article title improved the precision of the retrieved articles. Fifth, the current study used a thorough list of keywords related to the gut microbiota and IR based on those found in prior studies; nonetheless, it is probable that some keywords were omitted, which may have led to false-negative results. Sixth, the lists of important active players were taken directly from Scopus when retrieved. The amount of research produced by particular countries, funding agencies, or institutions might occasionally be underestimated in the Scopus database owing to multiple variations in the names or spellings of these entities. Seventh, bibliometric methods map volume, influence, and topic structure but do not assess the risk of bias or methodological rigor. Future systematic reviews/scoping reviews are needed to qualify the evidentiary strength of the clusters we identify. Finally, when searching at different points in time, the citation times of the publications are different; therefore, it is vital to update them for future studies. Despite our efforts to ensure consistency in the data retrieval process, the search period was limited to publications between 2015 and 2024, with data retrieval completed on June 15, 2025. While we used a PUBYEAR < 2025 filter to exclude post-2024 papers, any recent publications that have emerged since the retrieval date were not included in this study. Additionally, 2025 was excluded from the analysis because the year is still open for new issues and is not yet complete. Future studies should consider updating the results to account for new publications beyond the search cutoff.