Gut microbiota during pregnancy: a bibliometric analysis of global research trends and collaborative networks

To ensure comprehensive coverage and data accessibility, a systematic literature search was conducted on September 10, 2025, in the WoSCC and Scopus databases. The specific search strategy is illustrated in Figure 1. The search query for WoSCC was constructed using the "Topic" (TS) field, which includes title, abstract, author keywords, and Keywords Plus: TS = (("pregnan*" OR "gestation*" OR "maternal") and ("gut microflora" OR "gut microorganism" OR "gut microbi*" OR "gut flora" OR "gut dysbiosis" OR "intestinal microflora" OR "intestinal microorganism" OR "intestinal microbi*" OR"intestinal flora" OR "intestinal dysbiosis" OR "gastrointestinal microb*")).

For Scopus, the search was built using the "TITLE-ABS-KEY" (title, abstract, keywords) field: TITLE-ABS-KEY = ("pregnan*" OR "gestation*" OR "maternal") AND ("gut microflora" OR "gut microorganism" OR "gut microbi*" OR "gut flora" OR "gut dysbiosis" OR "intestinal microflora" OR "intestinal microorganism" OR "intestinal microbi*" OR"intestinal flora" OR "intestinal dysbiosis" OR "gastrointestinal microb*").

The literature types were limited to "Articles" and "Review Articles" published between 1991 and 2025, with no language restrictions. During the data screening process, 5,432 and 5,542 topic-relevant records were identified from WoSCC and Scopus, respectively. The following data were extracted and exported separately from each database for independent analysis: publication year, authors, subject categories, affiliations, journal names, countries/regions, keywords, and references. Additionally, 3,837 overlapping publications were identified between the two databases.

Bibliometric analysis was performed using the bibliometrix package (version 5.1.1) within the R programming environment (version 4.5.1), supplemented by VOSviewer (v1.6.19) and CiteSpace (v6.2. R3). Initial analyses utilized the built-in database statistics and the bibliometrix package to examine publication distributions, countries/regions collaborations, authors, and the top 10 highly cited and co-cited references. VOSviewer was further employed to extract and visualize institutional and keyword clustering networks. CiteSpace was used to perform country co-occurrence analysis, reference co-citation clustering analysis, and to generate a keyword timeline map. Additionally, a burst detection analysis was conducted for the top 30 keywords with the strongest citation bursts.

Analysis of the WoSCC and Scopus databases identified 5,432 and 5,542 publications, respectively (Figures 2A,C). There were 3,837 overlapping publications between the two databases. Figures 2B,D illustrate the annual publication trends, which show a high degree of consistency between the two databases, with annual growth rates of 15.97 and 15.52%, respectively. During the early stage (1991–2006), the annual publication count was generally below 20, and annual citations were fewer than 300. A phase of steady development occurred between 2007 and 2013, with a gradual increase in annual publications. Exponential growth has been observed since 2014, with both databases recording over 700 publications in 2024, reflecting the sustained high impact and growing academic interest in maternal and infant health research.

From 1991 to 2025, research on "gut microbiota during pregnancy" has garnered widespread global attention. In the WoSCC database (Table 1), China took the lead, with 1,561 publications (accounting for 21.67%), a total of 33,826 citations (average of 21.67 citations per article), an H-index of 80, a total link strength of 443 in the VOSviewer collaboration strength analysis, and a betweenness centrality of 0.06 in the CiteSpace collaboration network analysis (Figure 3A). The United States ranked second, with 1,511 publications (accounting for 27.82%), a total of 96,109 citations (average of 63.61 per article), an H-index of 144, a total link strength of 1,018, and a betweenness centrality of 0.3 in the CiteSpace collaboration network analysis, highlighting its advantages in the quality of basic research. Canada ranked third with 342 publications (6.3%), a total of 22,126 citations (average of 64.7 per article), an H-index of 75, a total link strength of 352, and a betweenness centrality of 0.14, indicating its role as a bridge in high-quality research and regional collaboration. England (total link strength 511, betweenness centrality 0.18) and France (betweenness centrality 0.2) emerged as key nodes in global scientific collaboration due to their high collaboration strength.

In the Scopus database (Table 1), China surpassed the United States with 1,729 publications (31.20%) compared to 1,436 (25.91%), a total of 36,944 citations (average of 21.37 per article), an H-index of 90, a total link strength of 428, and a betweenness centrality of 0.12 (Figure 3B), reflecting its leading output volume and high activity in open-access journals. The United States maintained a total of 101,868 citations (average of 70.94 per article), an H-index of 149, a total link strength of 996, and a betweenness centrality of 0.5, underscoring its leading position in high-quality research and collaborative networks. The United Kingdom (total link strength 469, betweenness centrality 0.17) and Germany (betweenness centrality 0.16) served as core nodal countries in Scopus due to their strong interdisciplinary collaboration.

Both WoSCC and Scopus identified the United States, China, and Canada as the top three most productive countries. The United States consistently ranked in the top two across both databases, demonstrating stable research quality and collaboration. China led in output volume in Scopus but still requires improvement in average citations and H-index. Canada ranked in the top three in both databases, reflecting its sustained research investment.

Keyword analysis to elucidate the knowledge structure and research hotspots in this field, we employed VOSviewer to perform a co-occurrence network analysis of literature keywords. To ensure clustering accuracy and representativeness, we excluded core search terms and their synonyms, while merging synonymous terms in the database. This step effectively mitigated clustering distortion and centralization bias caused by terminological diversity. In the co-occurrence network, line thickness visually represents the frequency of keyword co-occurrence in the same literature, with thicker lines indicating higher co-occurrence rates.

Results from the WoSCC database (Figure 7A) reveal that "Obesity" (607 occurrences) is the most frequent keyword, underscoring its central role in current research. "Risk" (572 occurrences) and "Probiotic" (555 occurrences) rank second and third. The co-occurrence network is clearly divided into four main clusters, each representing a distinct and concentrated research direction.

The green cluster centers on "Obesity" and focuses on the pathophysiological mechanisms of obesity and metabolic syndrome. It traces a research trajectory from phenotypic markers (e.g., "Overweight," "Body-Mass Index," "Gestational Diabetes Mellitus," "Preeclampsia," "Hypertension") to molecular mechanisms, emphasizing inflammation ("Inflammation," "Oxidative Stress," "Activation") and metabolic regulation ("Metabolism," "Insulin-Resistance," "Glucose"). Studies investigate how dietary factors ("Diet," "High-Fat Diet," "Dietary Fiber") influence gut microbiota metabolites ("Short-Chain Fatty Acids," "Butyrate," "Metabolites"), triggering host epigenetic changes ("Epigenetics," "DNA Methylation") and gene expression alterations ("Gene-Expression," "Expression"), ultimately impacting maternal-fetal health ("Maternal Obesity," "Maternal Diet," "Offspring," "Placenta"). This cluster integrates multi-level evidence from animal models ("Mice," "Cells") to human studies, unraveling the complex network of obesity-related metabolic diseases.

The blue cluster, led by "Risk" and "Probiotic," addresses early-life microbiota interventions and their long-term health outcomes, particularly allergic and immune diseases. This direction is characterized by clinical intervention and epidemiological features, as evidenced by keywords like "Double-Blind," "Supplementation," "Meta-analysis," and "Prevalence." Research focuses on early-life interventions ("Early-Life," "1st Year," "Children") using probiotics ("Probiotic," "Lactobacillus"), prebiotics ("Prebiotics"), or avoiding disruptors like antibiotics ("Antibiotics") and cesarean section (CS) ("Cesarean-Section") to modulate the immune system ("Regulatory T-Cells") and prevent diseases such as "Asthma," "Allergy," "Atopic-Dermatitis," "Eczema," "Food Allergy," and "Inflammatory-Bowel-Disease."

The red cluster revolves around "Health," "Infant," and "Breast-Milk," emphasizing the establishment and transmission of maternal–infant microbiota and its impact on infant health. Research threads begin with delivery and feeding modes ("Delivery," "Mode," "Breastfeeding," "Lactation"), exploring their effects on neonatal gut microbiota colonization ("Colonization") and diversity ("Diversity"). Studies link early microbiota establishment to infant growth ("Growth," "Birth-Weight"), immune responses ("Immune," "Responses"), and disease risks ("Necrotizing Enterocolitis," "Infection"). A highlight is the in-depth investigation of breast milk components ("Breast-Milk," "Milk," "Oligosaccharides," "Human-Milk Oligosaccharides"), revealing mechanisms by which breast milk shapes neonatal gut microbiota ("Bifidobacteria") to promote health. Keywords like "Sequences" and "Identification" reflect heavy reliance on high-throughput sequencing technologies.

The smaller yellow cluster, centered on "Dysbiosis" and "Brain," represents the rapidly growing "gut-brain axis" field. It explores links between gut microbiota and central nervous system function, as well as psychiatric and behavioral disorders. Research spans clinical conditions ("Autism," "Depression," "Anxiety," "Irritable-Bowel-Syndrome") and animal models ("Model," "Maternal Separation," "Maternal Immune Activation," "Prenatal Stress"). Keywords like "Stress," "Behavior," and "Neurodevelopment" highlight the focus: early-life (including prenatal) stress and immune activation induce dysbiosis, impacting brain development and behavior, potentially leading to neuropsychiatric disorders.

To validate and comprehensively examine the knowledge structure of the research field, this study further conducted a keyword co-occurrence analysis on the literature in the Scopus database. As shown in Figure 7B, its network structure shares commonalities with that of the WoSCC database while also exhibiting unique emphases.

In the Scopus database, "Controlled Study" (2,116 occurrences) emerged as the most frequently mentioned keyword. It was followed by "Animal" (1,944 occurrences) and "Infant" (1,778 occurrences). Through cluster analysis, the keywords were divided into three clusters.

The red cluster, the largest and most content-rich, was dominated by "Controlled Study" and "Animal," clearly pointing to mechanism-based explorations through animal experiments. This direction delved into the role of "Dysbiosis" under specific pathophysiological conditions, with core models including "Obesity," "Gestational Diabetes Mellitus," and "Inflammation." Research utilized interventions such as "Probiotic Agent," "Prebiotic Agent," "Diet Supplementation," "Fecal Microbiota Transplantation," and "High-Fat Diet" to observe their effects on the "Metabolism" and "Immunity" of "Progeny" or adult individuals. Mechanistic studies extended to the molecular level, employing techniques like "Real Time Polymerase Chain Reaction," "Enzyme Linked Immunosorbent Assay," and "Metabolomics" to investigate changes in "Gene Expression," "Protein Expression," "Signal Transduction," and "Metabolites" such as "Short Chain Fatty Acid," as well as the roles of key inflammatory factors like "Interleukin 6" and "Tumor Necrosis Factor." This cluster systematically revealed how interventions from "Prenatal Exposure" to the "Perinatal Period" influence host pathophysiology through molecular pathways.

The green cluster, centered around "Infant," "Newborn," and "Adult," primarily represented population-based clinical and translational medical research. Its research focus centered on the early life stages, encompassing the growth and development process from "Prematurity" and "Birth Weight" to "Child" ("Growth, Development and Aging"). Classic clinical research methods such as "Cohort Analysis," "Prospective Study," and "Randomized Controlled Trial" were employed to explore key exposure factors affecting infant health, including "Cesarean Section," "Vaginal Delivery," "Breast Feeding," and the use of "Antibiotic Agent." Health outcomes were closely tied to specific clinical issues like "Necrotizing Enterocolitis" and "Body Mass," with microbiome analysis conducted through "Feces" samples. This cluster overall embodied a longitudinal observational research paradigm from "Mother" to "Progeny."

The blue cluster featured distinct methodological characteristics, with "RNA 16S" (16S rRNA gene sequencing) and "Metagenomics" at its core. This cluster encompassed the key technical processes of gut microbiome research, from "Dna Extraction" to "Polymerase Chain Reaction," then to "High Throughput Sequencing" and "RNA Sequencing," culminating in "Population Abundance" analysis and "Classification." Correspondingly, the keywords in this cluster also listed in detail the core gut microbial phyla and genera identified through these methods, such as "Bacteroides," "Bifidobacterium," "Lactobacillus," "Firmicutes," and "Proteobacteria." This cluster essentially served as the technical and analytical foundation supporting the entire research field, providing data sources and research tools for the other two clusters.

To identify research hotspots with high burst strength and distinct temporal characteristics in the "gut microbiota during pregnancy" field, this study detected the top 30 keywords with the strongest citation bursts in the WoSCC and Scopus databases using CiteSpace (Figures 7C,D) and generated a timeline view of the field over the past 35 years (Figures 8A,B) to reveal thematic evolution and development dynamics. CiteSpace parameters were set as follows: time slicing (1991–2025), years per slice (1), node type (keyword), selection criteria (g-index, k = 25), no pruning.

The two databases revealed differentiated characteristics of the research frontiers. The WoSCC database timeline viewer showed the top five clusters as insulin resistance, gut-brain axis, allergy, vaginal microbiome, and necrotizing enterocolitis (Figure 8A). Keywords clearly displayed a three-stage evolutionary path from descriptive research to mechanistic and systemic studies (Figure 7C): the early stage (before 2010) was dominated by terms such as "intestinal microflora," "flora," and "atopic disease," reflecting the initial focus on microbial composition and its association with allergic diseases. Between 2010 and 2018, keywords such as "irritable bowel syndrome," and "anxiety like behavior" indicated a shift toward mechanistic exploration, particularly focusing on the gut-brain axis and immune regulation mechanisms. Recently (2023–2025), the emergence of "cesarean born infants" and "preeclampsia" showed an extension of research to delivery modes and associations with pregnancy complications.

In contrast, the Scopus database exhibited a more detailed cluster structure (Figure 8B), with the top seven clusters being feces; infant; animal experiment; blood glucose; adaptive immunity; brain; pregnane x receptor. The burst word list in Scopus included more general medical or methodological keywords (Figure 7D), such as "article," "priority journal" and "drug effects," which had high burst strength but weaker direct research relevance. Nevertheless, Scopus also captured biologically significant keywords with strong bursts, such as "bacterial colonization," "probiotics," and "gastrointestinal tract." Notably, "gestational diabetes" and "Shannon index" showed high burst strength between 2023 and 2025, indicating current research frontiers focused on the association between gestational diabetes and microbiota, as well as quantitative analysis of microbial diversity.

In summary, both databases indicate that the field has moved beyond mere description of microbiota composition toward exploring underlying mechanisms and multi-omics integration. Mechanistic research on pregnancy complications, especially GDM and preeclampsia, constitutes the current clinical frontier. The microbiota-gut-brain axis and its impact on maternal and infant neuropsychiatric health (e.g., anxiety, depression, autism spectrum disorder) form another key direction of shared interest.

A synthesis of the bibliometric data, including keyword co-occurrence, citation analysis, and the review of seminal publications, allows for the identification of several dominant and interconnected research hotspots that define the current intellectual structure of the field.

Analysis of the highly cited and co-cited literature provides insight into the intellectual foundation of the field. Seminal works, such as Koren et al.'s (2012) study demonstrating trimester-specific microbial changes (Koren et al., 2012), and research by Dominguez-Bello et al. (2010) and Backhed et al. (2015) on how delivery mode and feeding practices shape neonatal colonization, form cornerstone references. The frequent co-citation of methodological papers introducing tools like QIIME (Caporaso et al., 2010) and DADA2 (Callahan et al., 2016) underscores the indispensable role of bioinformatic advancements in propelling the field forward. The temporal evolution of research hotspots, as revealed by keyword burst analysis, indicates a clear trajectory. Early research (pre-2010) primarily focused on exploring associations between microbial composition and allergic diseases. As the field matured, themes expanded toward mechanistic investigations of the gut-brain axis, metabolic syndromes, and immune regulation. In recent years, the research frontier has sharpened its focus on specific pregnancy complications (e.g., GDM, preeclampsia), the detailed impact of delivery modes, and the rigorous evaluation of probiotic, prebiotic, and dietary intervention strategies. This evolution signifies a collective shift from initial correlational observations toward a deeper exploration of causal mechanisms and clinical translation.

Research in this field is characterized by robust international collaboration. The China, United States, and Canada lead in terms of publication volume. The United States' leadership in metrics such as total citations, H-index, and betweenness centrality demonstrates its substantial research accumulation, high-quality output, and central role in coordinating international collaborative networks. While China leads in publication count, its relatively lower average citation rates suggests a future focus on enhancing the international impact of its research output. European nations and Canada also play key hub roles in international collaboration, as evidenced by their advantages in betweenness centrality. Institutional collaboration clustering reveals that the field has formed several efficient collaborative networks centered around regions or research themes, such as North American, European, and Asia-Pacific clusters. This structured collaborative model facilitates resource integration and accelerates knowledge innovation.

A relatively stable and highly productive community of authors has emerged, with several research teams establishing distinctive and influential lines of inquiry. The work of these teams often aligns with and has helped define the major research clusters identified above.

For instance, the team associated with Collado, Maria Carmen; Isolauri, Erika; and Salminen, Seppo has systematically investigated the multifaceted impacts of gut microbiota during pregnancy on maternal and offspring health. Their research has elucidated mechanisms spanning microbial vertical transmission, metabolic-immune regulation, and nutritional interventions. They have confirmed that microbial communities from the maternal gut and breast milk are directly transferred to the newborn, a process influenced by delivery mode and feeding type (Flores Ventura et al., 2025; Saez-Fuertes et al., 2024a; Saez-Fuertes et al., 2024c). Their work has also linked maternal metabolic states, such as overweight/obesity, to gut dysbiosis and systemic inflammation (Zacarias et al., 2018), and explored the potential of probiotic (e.g., Bifidobacterium breve M-16V) and synbiotic supplementation (e.g., scGOS/lcFOS) to enhance maternal gut barrier integrity, immune status, and offspring health outcomes (Saez-Fuertes et al., 2024b; Saez-Fuertes et al., 2024c).

The collaborative work of Cryan, John F. and Dinan, Timothy G. has been pivotal in advancing the microbiota-gut-brain axis within the perinatal context. Their research investigates how early-life environmental factors (e.g., stress, antibiotic exposure, diet) affect neurodevelopment via the gut microbiota (Nuzum et al., 2024). Using animal models, they have demonstrated that maternal separation stress induces sex-specific changes in microbiota and leads to long-term visceral hypersensitivity and affective disorders (Wilmes et al., 2024). They have also shown that maternal high-fat diet impacts offspring brain development through microbial-metabolite pathways, such as the kynurenine pathway (Ratsika et al., 2024). Their interventional studies provide experimental support for nutritional and microbial strategies to ameliorate early-life stress-induced behavioral abnormalities (Donoso et al., 2020; McVey Neufeld et al., 2019; O'Mahony et al., 2020; Pusceddu et al., 2015).

Another illustrative example is the team led by Tain, You-Lin; Hsu, Chien-Ning; and Hou, Chih-Yao, which has expounded on how maternal factors (e.g., sweetener intake, chronic kidney disease, high-fat diets) can program offspring hypertension through mechanisms involving oxidative stress, aberrant renin-angiotensin system activity, and gut dysbiosis (Tain et al., 2024a,b; Tain et al., 2024d; Tain and Hsu, 2025). Their research demonstrates the potential of interventions like resveratrol butyrate esters (Tain et al., 2024d), chondroitin sulfate (Tain et al., 2024c), and specific probiotics to reshape microbial structure and prevent hypertensive programming in offspring through pathways involving hydrogen sulfide and nitric oxide signaling (Tain et al., 2023a, 2023b; Tain et al., 2024a,b,c,d).

The main methodological contribution of this study is its comparative analysis using both the WoSCC and Scopus databases, which provides valuable insights for bibliometric research practices. The high consistency in annual publication trends between the two databases validates the objective trajectory of the field's rapid development. However, discrepancies in rankings at the countries/regions, institution, and author levels reflect the different inclusion policies and biases of each database, as previously discussed. Therefore, utilizing both databases allows for a more comprehensive and equitable academic landscape.

Despite these insights, this study has several limitations. Firstly, it primarily focused on journal articles and reviews, excluding conference proceedings, book chapters, or preprints, which might omit some emerging perspectives or data. Secondly, although no formal language restriction was applied, the inherent predominance of English literature in both WoSCC and Scopus means that significant contributions from non-English-speaking regions may not be fully captured, potentially introducing a linguistic and geographical bias. Furthermore, bibliometrics is inherently a quantitative science; it can reveal macro-trends and structural patterns but cannot replace in-depth qualitative assessment of specific research findings or judgments on scientific rigor and validity.

Based on the synthesized research landscape, future directions are likely to focus on several key areas. First, there will be a continued push toward utilizing integrated multi-omics analyses (metatranscriptomics, metabolomics, and proteomics) to move beyond correlation and deeply elucidate the specific molecular pathways through which the gut microbiota influences pregnancy outcomes and offspring health. Second, promoting large-scale, well-designed randomized controlled trials based on existing mechanistic evidence will be crucial to verifying the efficacy and safety of probiotics, prebiotics, or dietary interventions in preventing pregnancy complications and improving long-term offspring health. Third, investigating microbiota characteristics across diverse geographical, ethnic, and socioeconomic populations is essential to ensure that findings are globally applicable and to promote health equity. Finally, exploring the use of microbiota-derived signatures as biomarkers for the early diagnosis and risk prediction of pregnancy complications represents a promising avenue for developing personalized microecological modulation strategies.

The future prospects of maternal and infant microbiome research are profoundly significant. This field holds the potential to fundamentally reshape preventive medicine in the perinatal period. By elucidating the critical role of microbial ecosystems in programming health and disease, it paves the way for novel, non-invasive strategies to improve pregnancy outcomes, optimize infant development, and reduce the long-term burden of non-communicable diseases across the lifespan. As research transitions from observation to intervention, the focus will increasingly be on translating this knowledge into actionable clinical guidelines and public health initiatives that support the microbial health of mothers and their children from pregnancy through the first critical years of life.