The Web of Science (WoS) Core Collection was selected for its rigorous journal curation (based on Bradford’s law) and widespread adoption in bibliometric studies (15).

A systematic literature search was conducted in the Web of Science Core Collection on 16 January 2025 to identify relevant publications on the association between oral health and frailty. The search strategy was constructed using comprehensive sets of terms related to “oral health” (e.g., Mouth Diseases, oral hygiene, tooth loss) and “frailty” (e.g., frailty syndrome, debilitation). The search was limited to English-language articles and reviews published between 2000 and 2024. The full Boolean query and detailed search strategy are provided in Supplementary Material 1.

Inclusion criteria: original articles or reviews published in peer-reviewed journals. Through a combination of manual screening and the “delete duplicates” function provided by CiteSpace 6.3.1, we excluded ineligible documents based on the following exclusion criteria: (1) errata; (2) duplicate publications; (3) irrelevant documents.

We exported the retrieved articles in plain text format with complete records and references named “download_XXX.txt” and then imported them into CiteSpace 6.3.1 for further analysis. In mapping the visual knowledge graph window, we followed the main procedural steps of CiteSpace, including time slicing, thresholding, modeling, pruning, merging and mapping (13). Nodes in different maps represent authors, institutions, countries, or keywords. The node size indicates the frequency of occurrence or citation and the color of the node indicates the year of occurrence or citation. In addition, nodes with purple pruning indicate higher mediator centrality. Centrality reflects the role of a node in a knowledge network, and nodes with high intermediate centrality (> 0.1) are usually considered as turning points or key points in a domain (14, 16). Links between nodes indicate co-occurrence relationships, and the thickness of the links indicates the strength of co-occurrence; the more links there are, the thicker and tighter the relationship between nodes. The specific parameters configured in CiteSpace for the construction and pruning of each network type are detailed in Supplementary Material 2.

Analysis of the 270 included publications revealed a clear upward trajectory in annual output (Figure 1). Although no relevant studies were published before 2006, there was notable growth between 2019 and 2022, culminating in a peak of 76 publications in 2024. The observed temporary declines in 2019 and 2023 did not alter the overall upward trend. This surge aligns with global demographic shifts and the heightened focus on aging-related health, particularly the role of oral health in frailty. The expansion is likely driven by increased research investment and interdisciplinary collaboration.

A total of 42 countries contributed to research on oral health and frailty, forming 68 collaborative networks (Figure 2). The top 10 contributing countries alone accumulated 5,316 citations (see Supplementary Table 1), corresponding to an average of approximately 21 citations per publication within this group. Japan was the most productive nation, contributing 83 publications (approximately one-third of the total) and accumulating the highest citation count of 2,295. Its average of approximately 28 citations per publication reflects a substantial output and influence, underscoring the country’s strong research focus and significant impact in this domain. This leadership position likely stems from Japan’s rapidly aging population (17) and its advanced research infrastructure at the intersection of dental science and geriatric medicine (18). China (n = 36) and the United States (n = 30) ranked second and third in publication output. Notably, Japan, the United States, the United Kingdom, and Switzerland demonstrated the highest betweenness centrality scores (> 0.1), indicating their pivotal roles as knowledge hubs in this research field (see Supplementary Table 1). The resulting network density of 0.079 indicates limited and fragmented international cooperation. Japan not only emerged as the most productive nation but also served as a central hub in the collaboration network, with direct links to 12 other countries. The United States demonstrated the highest betweenness centrality (0.46), functioning as the most critical bridge for international knowledge exchange. European nations, particularly England and Switzerland, formed a distinct collaborative cluster, while Asian countries showed relatively sparse cross-regional partnerships.

At the institutional level, organizations from the UK and Japan dominated the research landscape (Figure 3). The University of London led with 17 publications, followed by the Tokyo Metropolitan Institute of Gerontology (n = 14) and the University of Tokyo (n = 13). Interestingly, while the University of Newcastle, University of Adelaide, and University College London published fewer articles (not ranking in the top 10 by volume), they exhibited high centrality scores, suggesting their disproportionate influence as bridging nodes in the research network (see Supplementary Table 2). The institutional collaboration network exhibited a much lower density of 0.0141 (Figure 3), quantitatively confirming that research partnerships are highly concentrated and have not yet formed a widely interconnected network. Together, these network metrics (country density: 0.078; institutional density: 0.0141) quantify a clear trend: research collaboration in this field, while growing, remains concentrated around a few highly productive nations and institutions, rather than being a pervasive, worldwide network.

A total of 297 distinct authors were identified as contributors to the 270 publications analyzed (Figure 4). The top 10 authors, ranked by number of publications, were collectively cited 5,816 times, with an average of approximately 582 citations per author. Professor Hirohiko Hirano emerged as the most influential scholar, with 10 publications that accumulated 1,288 citations. Notably, his work achieved an average of approximately 129 citations per publication, a rate that substantially exceeds the group average and underscores his pivotal influence in the field. His seminal work has been particularly impactful in three key areas: (1) conceptualizing oral frailty, (2) developing standardized assessment tools, and (3) exploring innovative intervention strategies. Among the remaining top 10 authors, publication counts showed minimal variation (see Supplementary Table 3), suggesting the absence of a dominant research group. A key trend observed in authorship is its structural fragmentation. This observation is quantitatively supported by the author collaboration network density of 0.0149 and the co-cited author network density of 0.0154, both of which are very low. These values confirm the fragmentation of the research community and the lack of a tightly-knit collaborative group or a dominant intellectual school. While this fragmentation may reflect the interdisciplinary nature of oral frailty research, it also suggests the field would benefit from more coordinated collaborative networks to advance consensus and standardization.

The 270 publications were disseminated across more than 85 academic journals (Figure 5). The top 10 journals published 143 articles, which received 3,008 citations, resulting in an average of approximately 21 citations per article. The Journal of the American Geriatrics Society demonstrated the highest impact per publication among these, with an average of 71 citations per article. Gerodontology emerged as the most prolific journal, publishing 20 articles (7.4% of total publications) on this topic (see Supplementary Table 4), underscoring its role as a key platform for oral health and frailty research. The journal co-citation network density was 0.0434, meaning that researchers in this field tend to cite a common, albeit not monolithic, set of core journals. Notably, while the Journal of the American Geriatrics Society published only 8 articles, it achieved a remarkable citation count of 568 (see Supplementary Table 4), demonstrating its substantial impact in this field. The journal’s publications addressed critical aspects of the oral health-frailty relationship, including mechanistic investigations and intervention efficacy assessments, providing valuable theoretical and practical guidance for subsequent research.

Co-citation relationships occur when two or more publications are cited together by subsequent works, indicating conceptual connections in the academic literature. Our analysis generated a co-citation network comprising 295 reference nodes (Figure 6), where node size corresponds to citation frequency and connecting lines represent co-citation relationships. The co-citation frequency served as our primary analytical metric, as it objectively measures a publication’s academic influence within a research domain. The five most frequently co-cited studies (see Supplementary Table 5) predominantly examined: (1) risk factors associated with oral health decline and frailty, and (2) mechanistic relationships between oral dysfunction and physical frailty in aging populations. The five most co-cited references were cited 206 times in total, averaging approximately 41 co-citations per study. This high frequency highlights their foundational role and widespread recognition as key literature in the field.

A total of 157 keywords were analyzed to identify research hotspots and trends (19) (Figure 7). The network’s low density (0.0182) indicates a research landscape composed of several distinct, internally cohesive thematic clusters that have not yet coalesced into a tightly integrated framework. The most prominent keywords, based on frequency of occurrence and betweenness centrality, are summarized as follows. The top 10 keywords by frequency, along with their corresponding frequency and centrality values, were: “oral health” (Count = 123, Centrality = 0.42), “older adults” (Count = 60, Centrality = 0.05), “health” (Count = 53, Centrality = 0.41), “association” (Count = 47, Centrality = 0.36), “adults” (Count = 47, Centrality = 0.12), “tooth loss” (Count = 44, Centrality = 0.60), “people” (Count = 35, Centrality = 0.22), “quality of life” (Count = 33, Centrality = 0.20), “prevalence” (Count = 27, Centrality = 0.05), and “frailty” (Count = 26, Centrality = 0.04). Among these, “tooth loss” demonstrated the highest centrality value (0.60), highlighting its critical function as a conceptual bridge within the research network, despite not being the most frequently occurring term. Other keywords with notably high centrality included “oral health” (Centrality = 0.42) and “health” (Centrality = 0.41). The keywords related to oral health and frailty are: elderly population, quality of life, tooth loss, oral frailty and care. The hot keywords for the research direction were association and risk. Notably, “tooth loss” and “oral health-frailty association” demonstrated particularly high betweenness centrality scores (see Supplementary Table 6), suggesting their pivotal role as conceptual bridges in this research network.

The clustering analysis yielded eight distinct thematic groups using log-likelihood ratio (LLR) statistics (Figure 8). The robustness of our clustering results was confirmed by silhouette coefficients exceeding 0.7 for all clusters (20, 21), indicating both methodological soundness and substantive meaningfulness of the identified research themes.

The temporal evolution of research trends in oral health and frailty was examined through a citation burst analysis of the 16 most prominent keywords from 2000 to 2024 (Figure 9). In the generated visualization, the blue timeline represents the entire study period, while red segments highlight intervals during which specific keywords experienced statistically significant surges in citation frequency, defined as a threefold or greater increase relative to baseline levels (21). This analytical approach provides valuable insights into shifting research paradigms, as emerging keywords with strong citation bursts effectively map the migration of scientific frontiers. Recent burst keywords are particularly indicative of current research hotspots, revealing a clear trajectory of conceptual advancement in the field. The analysis demonstrates a progressive refinement of research focus, evolving from the initial epidemiological characterization of frailty in elderly populations, through investigations of specific oral function parameters, and ultimately to the development and validation of standardized assessment tools. This evolution reflects both the maturation of the field and an increasing recognition of oral health as a critical determinant of frailty syndromes, with contemporary research emphasizing quantifiable clinical indicators and their predictive value for adverse health outcomes in aging populations.

The marked growth in research output from 2000 to 2024, as evidenced by our bibliometric analysis, corroborates a robust association between oral health status and the development of frailty. This expanding body of literature reflects the growing scientific recognition of oral health as a critical determinant of geriatric health outcomes. Three nations—Japan, China, and the United States—have emerged as predominant contributors, collectively accounting for the majority of published works and significantly advancing the field’s theoretical and methodological foundations. However, our co-authorship network analysis revealed a notable deficiency in international research collaboration, characterized by limited cross-border partnerships among investigators. This finding is quantitatively underscored by the low network densities at the author (0.0149), institutional (0.0141), and even country (0.079) levels. These values confirm that the field is characterized by fragmented and decentralized research efforts, rather than strong, interconnected collaborative networks. Collectively, these findings underscore the need for enhanced multinational cooperation to facilitate knowledge exchange and comparative studies across diverse populations. Current research efforts primarily focus on three interconnected domains: (1) elucidating the pathophysiological mechanisms linking specific oral health conditions to frailty progression, (2) examining the clinical consequences of oral health-related frailty across functional domains, and (3) developing and validating standardized assessment tools capable of capturing the oral health-frailty continuum with greater precision.

The keyword burst analysis (Figure 9) provides direct bibliometric evidence of a methodological shift in the field, marked by strong citation bursts for quantifiable functional parameters such as “masticatory performance” (2018–2020) and “occlusal force” (2018–2020). This trend indicates that the research frontier is actively moving from broad conceptual associations toward the refinement of specific, objective assessment tools to operationalize the diagnosis of “oral frailty.”

This data-driven evolution aligns with recent advances in frailty assessment, which have begun incorporating oral health parameters, reflecting a growing recognition of their prognostic value. Approximately 40% of contemporary frailty indices now include at least one oral health component, typically evaluating the very functions highlighted by the burst keywords (3, 50–52). Parallel developments in oral health assessment tools, such as the Geriatric Oral Health Assessment Index (GOHAI), now incorporate frailty-relevant domains including nutritional intake and social functioning (53).

However, the bibliometric landscape, characterized by the sequential emergence of distinct measurement concepts, also underscores persistent standardization challenges. The heterogeneity in measurement approaches limits cross-study comparability, an issue compounded by the fact that only a minority of existing instruments have undergone rigorous validation in frail populations, indicating a clear need for future methodological refinement (53).