Knowledge structure and thematic organization of PFAS-associated cardiac toxicity: A bibliometric analysis

Abstract

Per- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants increasingly associated with cardiovascular outcomes. While mechanistic studies have reported oxidative stress, developmental toxicity, and gene expression alterations in cardiac models, it remains unclear how these findings integrate with structural cardiomyopathy frameworks.

Methods

A bibliometric analysis was conducted using 469 publications retrieved from Scopus, PubMed, and Web of Science (1992–2025) using PFAS and cardiac-related search terms. Knowledge structure was evaluated using co-occurrence network analysis (VOSviewer), thematic mapping (Biblioshiny), overlay visualization, and citation timeline analysis (CiteSpace).

Results

Keyword co-occurrence analysis identified two dominant thematic domains: (1) mechanistic and developmental cardiotoxicity (e.g., oxidative stress, gene expression, zebrafish), and (2) clinical cerebrovascular outcomes centered on patent foramen ovale and stroke. Thematic mapping positioned PFAS exposure descriptors as central but broadly defined themes, whereas structural cardiomyopathy constructs did not form cohesive or central clusters. Overlay analysis demonstrated temporal expansion toward molecular toxicology after 2018, without parallel growth in myocardial remodeling descriptors. Citation burst analysis revealed stable thematic progression rather than rapid cardiomyopathy-centered acceleration.

Conclusions

PFAS-associated cardiac research demonstrates progressive growth but remains structurally segmented. Mechanistic toxicology and clinical cardiovascular outcome studies operate largely in parallel, with limited integration into adult structural cardiomyopathy frameworks. These findings suggest the need for greater incorporation of myocardial phenotyping and remodeling endpoints in PFAS cardiac toxicology research.

Graphical Abstract

Highlights

• •PFAS cardiac research shows a dual-domain knowledge structure.
• •Mechanistic and clinical studies evolve with limited integration.
• •Structural cardiomyopathy themes are underrepresented.
• •Recent growth emphasizes molecular, not structural, endpoints.
• •Fragmentation limits translation to chronic cardiac disease models.

1. Introduction

Per- and polyfluoroalkyl substances (PFAS) are persistent environmental contaminants widely used in industrial and commercial applications due to their chemical stability and resistance to degradation, PFAS accumulate in biological systems and are detectable in human serum across global populations [1], [2]. Toxicological investigations have linked PFAS exposure to metabolic disruption, endocrine perturbation, immune effects, and hepatotoxicity [1].

Emerging evidence also implicates PFAS in cardiovascular dysfunction. Epidemiological studies report associations between PFAS exposure and dyslipidemia, hypertension, and elevated cardiovascular risk [2], [3], [4], [5]. Experimental models further demonstrate PFAS-induced oxidative stress, mitochondrial impairment, altered lipid metabolism, and developmental cardiotoxicity [6], [7], [8]. These findings suggest potential cardiac vulnerability to PFAS exposure across both developmental and adult life stages.

However, it remains unclear how mechanistic PFAS cardiotoxicity research aligns with structural myocardial disease frameworks. Contemporary cardiac pathology emphasizes myocardial remodeling, fibrosis, and ventricular dysfunction as central mechanisms underlying cardiomyopathy and heart failure [9], [10]. Whether PFAS-associated cardiac research has integrated these structural phenotypes or remains concentrated on developmental and cerebrovascular outcomes has not been systematically examined.

Bibliometric knowledge mapping provides an objective approach to evaluate thematic organization, research centrality, and temporal evolution within a scientific domain [11]. By analyzing co-occurrence networks, thematic positioning, and citation dynamics, it is possible to determine whether mechanistic toxicology and structural myocardial constructs are conceptually integrated or compartmentalized [12], [13].

The present study aimed to characterize the knowledge structure of PFAS-associated cardiac research from 1992 to 2025. Specifically, we evaluated thematic centrality, network architecture, and temporal progression to assess the extent of integration between PFAS exposure toxicology and structural myocardial remodeling constructs.

2. Methodology

2.1. Data sources and search strategy

A bibliometric analysis was conducted using Scopus, PubMed, and Web of Science (Core Collection). The search strategy was designed to capture publications addressing PFAS exposure and cardiac-related outcomes. The search string used was: TITLE-ABS-KEY(("perfluoroalkyl substance*" OR PFAS OR PFOA OR PFOS OR PFHxS OR PFNA OR "perfluorinated compound*" OR "perfluoroalkyl acid*") AND (heart OR cardiac OR myocardial OR cardiotoxicity)).

The search was limited to articles and reviews published between 1992 and 2025. Non-English records and document types other than original articles and reviews were excluded. Duplicate records between databases were identified and removed prior to analysis. A total of 469 publications met inclusion criteria and were included in the final dataset.

Analyses were conducted using Bibliometrix/Biblioshiny (R package), VOSviewer [11], CiteSpace, and Microsoft Excel for preprocessing.

3. Results

3.1. Publication characteristics and temporal distribution

A total of 469 publications met the inclusion criteria after database screening and deduplication. Scientific production was sparse before 2010 but demonstrated steady growth thereafter, with a noticeable increase after 2018. The overall trajectory indicates expanding research interest in PFAS-associated cardiac effects over the past decade, consistent with broader acceleration of PFAS toxicology research reported in recent reviews [2], [4], [5].

Fig. 1.

Annual scientific production of PFAS-associated cardiac research (1995–2025). The number of publications per year was extracted from Scopus and Web of Science using predefined PFAS- and cardiac-related search terms. Scientific output remained minimal prior to 2010, followed by progressive growth after 2014 and a marked increase from 2018 onward. The post-2018 expansion corresponds temporally with increasing emphasis on mechanistic cardiotoxicity studies, as reflected in subsequent thematic and network analyses. This trajectory indicates sustained growth of the field rather than episodic publication bursts.

Citation burst analysis did not reveal strong transient accelerations in cardiomyopathy-related keywords. Instead, dominant terms such as "patent foramen ovale" and "perfluorooctane sulfonate" showed sustained presence across multiple time intervals, reflecting the longitudinal stability of these clinical and exposure-related research streams [14], [15].

3.2. Thematic mapping

Thematic mapping based on co-word analysis demonstrated distinct quadrant positioning of research themes.

Fig. 2.

Thematic map of PFAS-associated cardiac research based on co-word analysis. Themes are plotted according to Callon’s centrality (x-axis) and density (y-axis), representing relevance and degree of development, respectively. Developmental toxicity, zebrafish, gene expression, and PFAS–metabolomics clusters occupy the motor theme quadrant (high centrality and density), indicating well-developed and influential research streams. PFOS and PFOA appear within the basic theme quadrant (high centrality, low density), reflecting foundational but broadly defined exposure constructs. Patent foramen ovale and stroke are positioned in the emerging or declining quadrant, while general PFAS descriptors and cardiotoxicity-related terms are located near the central region with moderate development. Structural myocardial remodeling constructs are not represented as dominant themes.

Thematic mapping demonstrated that developmental toxicity and molecular descriptors (e.g. gene expression, zebrafish models, and metabolomics) function as motor themes, indicating strong internal cohesion and high conceptual influence [6], [7], [8]. Zebrafish-based models have been central to PFAS cardiotoxicity research given their genetic tractability and cardiac transparency during early development [6], [8], [16]. In contrast, specific exposure compounds (PFOS, PFOA) occupy the basic theme quadrant, suggesting central but broadly defined roles within cardiac toxicology [2], [4]. Notably, stroke- and PFO-related constructs are positioned within the emerging or declining quadrant, indicating reduced developmental density despite historical prominence [14], [15]. Structural cardiomyopathy descriptors do not form a central thematic cluster, further supporting limited integration of PFAS exposure research with myocardial remodeling frameworks [9], [10].

Clinical cerebrovascular constructs, including "patent foramen ovale," "stroke," and "cryptogenic stroke," were located in the motor theme quadrant (high centrality, high density), indicating well-developed and highly interconnected research streams [14], [15]. Large randomized controlled trials such as the REDUCE trial have substantially shaped this clinical cluster [14].

PFAS-related exposure descriptors ("perfluoroalkyl substances," "PFOS," "PFOA") occupied the basic theme quadrant (high centrality, low density), reflecting foundational but broadly defined roles across cardiac toxicology studies [2], [3], [4], [5].

Developmental toxicity, oxidative stress, and gene expression appeared in emerging thematic regions (low centrality, low density), consistent with recent growth in mechanistic investigations [6], [7], [8], [17].

Structural cardiomyopathy descriptors, including "ventricular remodeling," "myocardial fibrosis," and "dilated cardiomyopathy," did not form cohesive thematic clusters and exhibited low centrality within the network [9], [10].

3.3. Keyword co-occurrence network

Keyword co-occurrence analysis identified two dominant thematic domains across the Scopus, PubMed, and Web of Science datasets.

Fig. 3.

Keyword co-occurrence network of PFAS-associated cardiac research. Network visualization generated using VOSviewer based on keyword co-occurrence analysis (minimum occurrence ≥5). Node size reflects keyword frequency, and link strength represents co-occurrence intensity. Two dominant thematic domains are observed: a clinical cardiovascular cluster (green) centered on patient outcomes and stroke-related constructs, and a mechanistic toxicology cluster (red) characterized by developmental toxicity, gene expression, zebrafish models, and myocardial tissue descriptors. Limited bridging nodes connect the domains, indicating restricted cross-thematic integration.

Keyword co-occurrence network analysis revealed two dominant and relatively discrete thematic clusters. The first cluster comprised clinical and procedural descriptors, including patient, stroke, cryptogenic stroke, and PFO-related constructs [14], [15]. The second cluster consisted of mechanistic toxicology terms, including heart, tissue, development, gene expression, zebrafish, and developmental cardiotoxicity [6], [7], [8]. Inter-cluster connectivity was present but comparatively sparse, and structural cardiomyopathy descriptors did not form an independent or central subnetwork.

The first domain comprised mechanistic toxicology constructs such as "developmental toxicity," "zebrafish," "oxidative stress," and "gene expression," reflecting experimental and molecular approaches to PFAS-associated cardiotoxicity [6], [7], [8].

The second domain centered on clinical cardiovascular outcomes, dominated by "patent foramen ovale," "stroke," "echocardiography," and related procedural descriptors [14], [15]. This clinical cluster reflects a well-established line of investigation into the mechanistic and interventional relationships between PFO and cryptogenic stroke, largely independent of PFAS exposure science.

Inter-cluster connectivity between these domains was limited. Structural cardiomyopathy-related keywords were sparse and did not constitute an independent subnetwork [9], [10]. Comparable dual-domain organization was observed inn both databases.

3.4. Temporal overlay and timeline analysis

Overlay visualization demonstrated early emphasis (2014–2017) on stroke- and PFO-related constructs [14], [15]. From 2018 onward, mechanistic toxicology terms, including oxidative stress, gene expression, and developmental cardiotoxicity, became more prominent, reflecting a field-wide shift toward experimental and molecular approaches [6], [7], [8], [17].

Earlier publications were dominated by clinical and procedural terms related to patient cohorts, stroke, and PFO-related constructs [14], [15]. In contrast, more recent publications increasingly emphasize mechanistic toxicology descriptors, including development, gene expression, heart tissue, and zebrafish models [7], [8], [18]. Despite this shift toward molecular cardiotoxicity, structural myocardial remodeling descriptors did not demonstrate marked temporal emergence [9], [10].

Fig. 4.

Overlay visualization of keyword co-occurrence network showing temporal evolution of PFAS-associated cardiac research. Colors represent the average publication year of keywords (blue/purple = earlier; yellow = more recent). Early research (approximately 2014–2017) was predominantly centered on clinical and procedural constructs, including patient, PFO closure, stroke, and device-related terms. From 2018 onward, increased prominence of mechanistic toxicology descriptors such as development, heart, tissue, gene expression, zebrafish, and developmental cardiotoxicity is observed. Structural myocardial remodeling constructs do not demonstrate comparable temporal intensification.

Timeline analysis indicated persistent separation between PFAS toxicology clusters and cerebrovascular outcome clusters across the study period (1992–2025). No dominant integrative cluster linking PFAS exposure with structural myocardial remodeling constructs was observed [9], [10].

Fig. 5.

CiteSpace timeline and cluster visualization of PFAS-associated cardiac research (1992–2025). (A) Timeline view showing major keyword clusters across two-year slices. Dominant clusters include #0 cryptogenic stroke, #1 perfluorooctane sulfonate, #2 perfluoroalkyl substances, and #4 atrial septal defect. Horizontal lines indicate cluster persistence over time, with limited cross-cluster convergence. (B) Cluster view illustrating modular separation of thematic domains. Clinical cerebrovascular clusters (e.g., cryptogenic stroke, atrial fibrillation, myocardial infarction) are spatially distinct from PFAS exposure clusters (e.g., perfluoroalkyl substances, perfluorooctane sulfonate). Node size reflects citation frequency, and color indicates time slice.

CiteSpace timeline analysis demonstrated sustained thematic separation across the study period (1992–2025). Major clusters included cryptogenic stroke (#0), perfluorooctane sulfonate (#1), perfluoroalkyl substances (#2), and atrial septal defect (#4). These clusters persisted longitudinally without evidence of structural convergence. Cluster visualization further revealed spatial separation between PFAS exposure clusters and clinical cerebrovascular clusters. Cardiomyopathy-related constructs did not form dominant or persistent clusters, consistent with the broader pattern of limited myocardial remodeling integration

4. Discussion

The present bibliometric analysis demonstrates progressive growth of PFAS-associated cardiac research but reveals a structurally segmented knowledge architecture. Across thematic mapping, network topology, and temporal analyses, two dominant domains consistently emerged: mechanistic/developmental cardiotoxicity and clinical cerebrovascular outcomes.

Thematic mapping positioned PFAS exposure descriptors as central but broadly defined constructs. This suggests that exposure science forms the conceptual foundation of the field but is not consistently differentiated into specific myocardial phenotypes [2], [3], [4], [5]. In contrast, stroke- and PFO-related constructs function as motor themes, reflecting well-developed and cohesive clinical research streams [14], [15].

Mechanistic toxicology, characterized by oxidative stress, gene expression alterations, and zebrafish-based developmental models, has expanded in recent years [6], [7], [8]. However, these investigations remain weakly integrated with structural cardiomyopathy frameworks. Canonical descriptors of ventricular remodeling, myocardial fibrosis, and dilated cardiomyopathy were not central within the thematic architecture [9], [10].

A critical finding of this study is the peripheral positioning of structural cardiomyopathy frameworks. Although mechanistic depth has increased since 2018, this expansion is largely confined to developmental and molecular endpoints, such as oxidative stress and gene expression in zebrafish models [6], [8], [16]. Notably, canonical descriptors of ventricular remodeling, myocardial fibrosis, and dilated cardiomyopathy remain absent from the central thematic clusters [9], [10]. Emerging studies using human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) have begun to bridge this gap, demonstrating that combined PFAS exposure induces extracellular matrix remodeling pathways, TGF-β signaling activation, and pro-fibrotic protein enrichment consistent with early cardiac fibrosis, hallmarks of structural cardiomyopathy [9], [10]. This suggests that chronic exposure paradigms have not yet been systematically aligned with the adult structural phenotypes that define clinical heart failure.

The co-occurrence network further supports a dual-domain organization. Mechanistic toxicology and clinical cardiovascular research appear to operate largely in parallel, with limited bridging nodes [12], [13]. This pattern was reproducible across Scopus, PubMed, and Web of Science datasets, indicating that the observed fragmentation is structural rather than database-specific.

Clinical investigations are primarily focused on cerebrovascular outcomes and PFO-related constructs [14], [15], whereas mechanistic studies emphasize developmental toxicity and molecular stress responses [6], [7], [8]. Although limited cross-linkage exists, the relative separation of these domains indicates constrained conceptual integration. Structural myocardial remodeling constructs do not occupy a central position within the network, suggesting that chronic exposure paradigms have not been systematically aligned with adult cardiomyopathy phenotypes [9], [10].

Temporal overlay analysis suggests increasing mechanistic depth since 2018, yet this expansion has not been accompanied by proportional integration into adult myocardial remodeling constructs [9], [10]. Citation burst analysis did not demonstrate rapid acceleration of cardiomyopathy-centered themes, reinforcing the notion of gradual, domain-specific evolution rather than convergence.

The overlay analysis indicates increasing mechanistic depth in PFAS-associated cardiac research over the past decade. However, this expansion appears confined to developmental and molecular endpoints rather than extending toward adult myocardial remodeling constructs. Thus, while the field demonstrates thematic progression, it does not show clear convergence between exposure toxicology and structural cardiomyopathy phenotyping [9], [10]. The temporal pattern reinforces the dual-domain architecture observed in network analysis and suggests parallel evolution rather than integrative maturation.

From a toxicological perspective, PFAS-associated cardiac research has primarily emphasized developmental endpoints and molecular stress responses [6], [7], [8]. A growing body of zebrafish studies has confirmed dose-dependent impairment of cardiac morphogenesis, hatching, and heart rate following PFOS and related compound exposures [6], [16], [18]. Similarly, population-based cohort studies and meta-analyses have reported associations between PFAS serum levels and subclinical cardiovascular markers including dyslipidemia and arterial hypertension [1], [3], [5]. Adult structural myocardial phenotypes central to cardiomyopathy classification remain comparatively underrepresented. Integration of chronic exposure paradigms with myocardial remodeling endpoints may enhance mechanistic understanding of PFAS-related cardiac dysfunction [9], [10].

The CiteSpace timeline analysis confirms that thematic fragmentation has persisted for more than two decades. PFAS exposure clusters and cerebrovascular outcome clusters have evolved in parallel trajectories, with limited structural integration [14], [15]. The absence of a dominant cardiomyopathy-centered cluster across time slices suggests that myocardial remodeling has not become a unifying framework within PFAS cardiac toxicology [9], [10]. These findings indicate that the field has matured through domain-specific expansion rather than cross-disciplinary consolidation.

Overall, the field demonstrates thematic expansion but limited cross-domain consolidation, suggesting opportunities for future studies incorporating structural phenotyping, metabolic remodeling markers, and functional cardiac endpoints within PFAS exposure models [9], [10].

5. Conclusions

PFAS-associated cardiac research has expanded substantially over the past decade [2], [5]; however, bibliometric mapping reveals a segmented thematic structure. Mechanistic developmental toxicology [6], [7], [8] and clinical cerebrovascular outcome research [14], [15] constitute the dominant domains, while structural myocardial remodeling constructs remain peripheral [9], [10]. Greater integration of myocardial phenotyping and remodeling endpoints into PFAS exposure studies may strengthen mechanistic insight into PFAS-related cardiac toxicity.

CRediT authorship contribution statement

Muna ‘Izzah Azman: Data curation, Investigation, Validation, Writing & Editing. Normala Abd Latip: Conceptualization, Data curation, Investigation, Methodology, Supervision, Project administration, Funding, Writing & Editing.

Funding

This research is partly funded by MOHE HICoE CARE-I (001/2025-4) and UiTM DUCS A 2025 (600-UiTMSEL (PI. 5/4/8) (011/2025)).

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Handling Editor: Prof. L.H. Lash

Data Availability

Data will be made available on request.