A bibliometric analysis of research trends in mesenchymal stem cell therapy for neonatal bronchopulmonary dysplasia: 2004–2024

Bibliometric analysis offers several advantages over traditional literature reviews for mapping research landscapes. Unlike narrative reviews that may be influenced by subjective selection biases, bibliometric approaches employ quantitative methods to analyze publication metadata, identifying patterns that might not be apparent through conventional literature synthesis. This approach is particularly valuable for examining the structural characteristics of a research field, including collaborative networks, citation patterns, and thematic evolution over time.

The Web of Science Core Collection (WoSCC) served as the primary database for this bibliometric analysis. WoSCC was selected for its comprehensive coverage of peer-reviewed literature, standardized metadata, and robust citation tracking capabilities, as highlighted by Zhu and Liu (17). While databases such as PubMed offer broader coverage of biomedical literature, WoSCC provides more complete citation data necessary for co-citation analyses. Indexing was restricted to the Social Sciences Citation Index (SSCI) and Science Citation Index Expanded (SCI-EXPANDED) to ensure the inclusion of high-quality academic literature (17).

The search period was defined as January 2004 to November 2024, with the starting date corresponding to the emergence of the first publications exploring MSC applications in BPD-related models. This 20-year timeframe allows for comprehensive tracking of the field's evolution from early conceptual studies through preclinical investigations to clinical translation. All data were retrieved on 5 November 2024, at 11:01:52 GMT+0800 (CST), ensuring the most up-to-date results from the database.

The search strategy employed Medical Subject Headings (MeSH) terms and related keywords to capture comprehensive data on bronchopulmonary dysplasia (BPD) and mesenchymal stem cells (MSCs). The final search string was formulated as follows: •Query: TS = (Bronchopulmonary Dysplasia OR Dysplasia, Bronchopulmonary OR Ventilator-Induced Lung Injuries OR Ventilator Induced Lung Injury) AND TS = (Mesenchymal Stem Cells OR Mesenchymal Stromal Cells OR Mesenchymal Progenitor Cells OR Wharton's Jelly Cells OR Adipose-Derived Mesenchymal Stromal Cells OR Multipotent Bone Marrow Stromal Cells).

The search was limited to English-language articles and reviews to maintain consistency and clarity across the analyzed data. Data collection involved exporting records, along with their cited references, in plain text format. Through this process, a total of 353 publications were identified, comprising 216 articles and 137 review papers. Table 1 and Figure 1 illustrate the retrieval process for this study.

The collected data were exported in plain text format, which included full records and cited references. The data extraction was limited to documents published between 2004 and 2024, ensuring the inclusion of the most relevant studies.

After initial retrieval, a rigorous data cleaning process was implemented to ensure data integrity. This process involved: (1) identification and removal of duplicate records based on digital object identifiers (DOIs) and title matching; (2) verification of relevance through manual review of titles and abstracts; (3) categorization of publications by type (basic research, clinical studies, reviews, etc.); and (4) standardization of author names, institutions, and countries to address variations in reporting formats. Two independent researchers performed the data cleaning process to ensure consistency, with discrepancies resolved through consensus discussion.

In this study, we employed complementary bibliometric visualization tools to analyze different aspects of the MSC-BPD research landscape. Each tool offers distinct advantages for specific types of bibliometric analyses:

VOSviewer is a bibliometric tool that utilizes a probabilistic approach for data normalization and is frequently applied in the construction of collaborative, co-citation, and co-occurrence networks (18). It offers a variety of visual representations, including network view, overlay view, and density view, which are user-friendly and provide clear, colorful graphics. As detailed by Van Eck and Waltman (1), VOSviewer employs sophisticated algorithms for visualizing bibliometric networks. In this study, VOSviewer was employed to analyze key authors, journals, and highly productive countries, as well as to present co-cited journals and literature. In the generated visual representations, each node represents an entity, such as a country, institution, journal, or author. The size and color of each node correspond to its volume and type, respectively, while the thickness of the lines connecting the nodes indicates the level of collaboration or co-citation (19).

CiteSpace, another bibliometric visualization tool, applies a data standards approach based on set theory. As described by Chen et al. (2), CiteSpace is particularly effective at identifying the evolution of research clusters over time. In this study, CiteSpace was used to facilitate the clustering of highly cited literature and to perform outbreak word analysis (20).

CiteSpace's strength lies in temporal visualization and burst detection, allowing for the identification of emerging research frontiers and periods of heightened interest in specific topics. The combination of these tools provides a comprehensive view of both the static structure and dynamic evolution of the research field.

Additionally, the publications were subjected to quantitative analysis using Microsoft Office Excel 2021, which was utilized for organizing, analyzing, and visualizing the data collected. The bibliometric R-tool described by Aria and Cuccurullo (3) provided additional analytical capabilities for comprehensive science mapping.

Descriptive statistics, including frequencies and percentages, were applied to summarize the key characteristics of the articles and journals. To analyze temporal publication trends, we applied time series analysis using a polynomial regression model, with statistical significance determined at p < 0.05. The fit of our data to established bibliometric distributions was assessed using logarithmic transformation and linear regression for Bradford's law, with R² values greater than 0.9 considered indicative of good fit.

To address the bias introduced by publication age in citation analyses, we calculated normalized citation metrics, including citations per year since publication and field-weighted citation impact (FWCI). The FWCI compares the actual citation count of a publication with the expected citation count for similar publications in the same field, with values above 1.0 indicating above-average impact. Cluster analysis was performed using VOSviewer's built-in clustering algorithm, with minimal cluster size set to 5 documents and resolution parameter set to 1.0.

Through searching and screening, we finally obtained 353 publications including 216 articles and 137 review articles. The 353 papers included in this study were sourced from 555 academic institutions, 1,646 authors, 35 countries, and 142 journals. Additionally, they were cited a total of 15,089 times from 2,267 journals. As demonstrated in Table 2.

Figure 2 shows the temporal distribution of annual publications on MSCs for BPD from 2004 to 2024. The data reveal a significant upward trend in publication volume (p < 0.01), with particularly accelerated growth observed after 2015. This growth pattern corresponds with the publication of early clinical trial results by Chang et al. (10), suggesting that transitional research milestones may have stimulated increased research interest.

To analyze this trend further, a binomial function was applied to model the cumulative publication data:y=1.2185×2−60.0652x+70.4298y=1.2185x2−60.0652x+70.4298y=1.2185×2−60.0652x+7.4298R2=0.9957

The high R² value (0.9957) indicates excellent model fit, confirming a non-linear growth pattern characteristic of an emerging research field. The compound annual growth rate was calculated at 18.2%, significantly exceeding the average growth rate of 3.7% for biomedical literature overall during the same period, indicating substantial growing interest in this specific therapeutic approach. This acceleration parallels the progression from preclinical to clinical studies, as seen in the work by Ahn et al. (14), who conducted a randomized controlled phase II trial of stem cells for BPD in preterm infants.

Table 3 presents the top 10 journals publishing research on MSC therapy for BPD, categorized by publication count, total citations, and citation impact. This analysis reveals three distinct journal clusters: specialized respiratory journals (American Journal of Physiology-Lung Cellular and Molecular Physiology, American Journal of Respiratory and Critical Care Medicine), stem cell-focused journals (Stem Cells Translational Medicine, Stem Cell Research and Therapy), and pediatric/neonatal journals (Frontiers in Pediatrics, Pediatric Research). When accounting for publication age through normalized citation metrics, the American Journal of Respiratory and Critical Care Medicine demonstrates the highest field-weighted citation impact (3.8), indicating substantial influence despite publishing fewer articles than other journals. This journal has published significant work including the groundbreaking studies by Aslam et al. (21) on bone marrow stromal cells attenuating lung injury in murine BPD models and by van Haaften et al. (22) on airway delivery of MSCs preventing arrested alveolar growth.

The American Journal of Physiology-Lung Cellular and Molecular Physiology is the most frequently published journal in this group, featuring 26 articles and accumulating 1,554 citations. Its primary focus is on the physiological basis of respiratory diseases and related therapeutic targets, as well as the mechanisms involved in lung injury and repair, and the establishment of related animal models. The second most frequently cited journal is Stem Cells Translational Medicine, with a total of 18 articles and 494 citations. Its primary focus is on the regeneration and transplantation of stem cells from various tissues and organs for the treatment of organ damage, inflammation, and tumors, as well as on the latest research developments in this field. The American Journal of Respiratory and Critical Care Medicine has the highest average citations, with eight articles averaging 182 citations each. This journal has published landmark research by Willis et al. (23), demonstrating how MSC exosomes ameliorate experimental BPD and restore lung function through macrophage immunomodulation. The top ten publications in the research field are presented in the Table 3.

Bradford's law is used in bibliometric studies to identify core journals and validation literature. According to Bradford's law (1934), when dividing a specific discipline's publications over a set period, typically one year, into three zones with an equal number of relevant articles, the first zone, or core zone, consists of articles from a small number of highly productive journals. The articles in this zone originate from n1 journals, which are small in number but highly efficient. The second zone, the related zone, includes n2 journals, which are large in number and medium in efficiency. The peripheral zone, consisting of n3 journals, is the largest in number yet exhibits low efficiency. The distribution of journals across the three zones can be represented as n1, n2, and n3, following the ratio 1: a: a² (a > 1) (24, 25). Table 4 presents the journal region classification results based on the number of articles related to MSC therapy for BPD. The distribution of papers across the three regions is roughly equal, with the journal ratio of 1:3:9 reflecting Bradford's law.

To validate Bradford's law mathematically, we plotted the cumulative number of journals against the logarithm of journal rank, resulting in a linear relationship with R² = 0.947, confirming good adherence to Bradford's distribution. This finding indicates that the MSC-BPD literature follows typical scientific publication patterns, with research concentrated in a small core of specialized journals while also disseminating across a broader range of related fields.

Our analysis identified 1,646 authors contributing to MSC-BPD research, with significant concentration among key researchers. Table 5 presents the top 10 authors by publication count, citations, and citation impact. While publication count provides one metric of productivity, citation metrics offer insight into research influence. Of particular note, Mitsialis, S. Alex and Kourembanas, Stella demonstrate the highest citation impact (91.07 and 89.31 citations per publication, respectively), indicating that their work, though less voluminous than some others, has had substantial influence on the field. Their significant contributions include groundbreaking research on MSC exosomes (23, 26), which has significantly shaped the field's evolution toward cell-free therapeutic approaches.

1,646 scholars published articles about MSC therapy for BPD. Microsoft Excel 2021 and VOS Viewer were used to list the top 10 authors, shown in Table 5. Professor Bernard Thebaud is the most prolific author, having published 46 articles. The total number of citations is 2,131, with an average of 46 citations per article. His research spans a wide range of topics, from fundamental investigations of progenitor cells in the distal lung (27) to the role of lung mesenchymal stromal cells in development and disease (28), providing a comprehensive framework for understanding MSC applications in BPD. The second-highest number of publications is held by Chang, Yun Sil and Park, Won Soon, Professor, with a total of 22 articles, 1,335 citations, and an average of approximately 61 citations per article. Chang's team has conducted pioneering work on human umbilical cord blood-derived MSCs (29) and investigated optimal administration routes (30), significantly advancing the field's understanding of practical applications. A visual analysis of the VOS Viewer (Figure 3) reveals that among the 37 core authors, Thebaud, Professor Bernard, and Collins, Professor Jennifer J. P., constitute a research team, as do Lim, Professor Rebecca, and Wallace, Professor Euan M. These two teams engage in academic exchanges and collaborations. Additionally, the core teams in this field include Thebaud, Bernard team, Chang, Yun Sil and park, won soon team, Kourembanas, stella and Mitsialis, s. Alex team, Bellusci, Saverio team, goldsmith, Adam m. team, and so on. As illustrated in the VOS viewer visualisation analysis (Fig. Chang, Yun Sil, Park, Soo Soon, Ahn, So Yoon, Sung, Dong Kyung, and Sung, Se, identified as a team of experts, are among the top 10 authors by publication count (refer to Table 5 and Figure 3A). AM, Kourembanas, Stella and Mitsialis, S. Alex constitute a team, as do Bellusci, Saverio and Morty, Rory. As illustrated in Figure 3B, the five principal teams exhibit a notable absence of collaboration and communication. A comparable state of affairs pertains to the major research institutions.

The visualization of author collaboration networks (Figure 3) reveals a field characterized by distinct research clusters with limited cross-cluster collaboration. Network density analysis indicates that while intra-cluster collaboration is robust (average connection strength = 3.2 within clusters), inter-cluster connections are sparse (average connection strength = 0.4 between clusters). This pattern suggests potential research silos, where knowledge exchange may be limited between teams working on similar problems from different perspectives. Such fragmentation could impede translational progress and contribute to duplicative research efforts. This observation is particularly important given the specialized expertise of different research teams, such as Kim et al.'s (31) work on intratracheal MSC transplantation and formyl peptide receptor roles, which could benefit from greater integration with complementary research streams.

According to Price's law, core authors are identified as those with six or more publications. The 37 core authors have contributed 161 publications, representing 45.6% of the total output. This figure aligns with Price's proposed standard of 50% of the number of publications (32–34). Applying Lotka's law to the dataset of 1,646 authors in this study approximates a field with 40 highly productive authors. Furthermore, the number of core authors aligns with the law, suggesting the formation of a more stable author collaboration group.

Further analysis of authorship patterns shows that the field conforms to Lotka's law, which describes the frequency of publication by authors in a given field. A logarithmic plot of author productivity (number of authors vs. number of publications per author) yielded a linear relationship with a slope of approximately −2 (R² = 0.91), consistent with Lotka's inverse square law. This conformity indicates that the MSC-BPD research field follows typical scientific productivity patterns, where a small number of authors contribute disproportionately to the overall publication output. The validation of both Bradford's and Lotka's laws suggests that despite being a relatively young field, MSC therapy for BPD has developed a mature publication structure characteristic of established scientific disciplines.

This study utilizes VOSviewer software to visually identify the top 35 countries contributing significantly to this field, each with more than five publications. The resulting visualisation is presented in Figure 4. Figure 4 shows that most articles were published from 2016 to 2022. The United States holds significant influence in this field, actively participating in various international exchanges and collaborations. China has particularly strong links with the US, Germany, and Canada.

Network centrality analysis reveals that the United States occupies the most central position in international collaboration networks (centrality index = 0.83), functioning as a primary hub for global research exchanges. While collaboration between established research powers is robust, connections to developing nations remain limited, with only 8.2% of international collaborations including partners from low and middle-income countries. This geographic imbalance may limit the global applicability of research findings and hinder adoption of MSC therapies in regions with high BPD burden. This pattern is evident in publication trends, with Park et al. (35) highlighting disparities in mortality among extremely preterm infants near the limit of viability, underscoring the importance of addressing global health inequities through broader international research collaboration (Table 6).

Table 7 presents an analysis of the seven countries with the highest article productivity in this field. Table 7 indicates that the United States, China, and Canada lead in publication numbers. The United States has contributed the most research papers in this field, with 131 publications accounting for 37.1% of the total and garnering 6,432 citations. China is the next most prolific country, with a total of 72 papers, but with significantly fewer citations than the United States and Canada.

Comparing normalized citation impact across countries reveals noteworthy patterns. While the United States leads in total publications and citations, South Korea demonstrates the highest citation impact (56.04 citations per publication), followed by Canada (46.17) and Italy (40.30). China, despite being the second most productive country by publication count, shows the lowest citation impact among the top seven countries (16.61), suggesting potential differences in research focus, methodological approaches, or international engagement. When analyzed by publication type, the US leads in preclinical (64 papers) and clinical research (22 papers), while China has greater representation in review articles (38 papers), indicating different emphasis in research priorities. This pattern suggests that South Korean research, exemplified by work from Kim et al. (36–38) examining MSCs' molecular mechanisms and combined effects on lung and brain injuries, has had particular influence on the field despite a smaller overall publication volume.

Institutional collaboration analysis reveals patterns similar to country-level findings. Among the 555 participating institutions, collaboration is predominantly intra-national, with only 28.4% of institutional partnerships crossing national boundaries. The ten most productive institutions account for 41.2% of all publications, indicating significant concentration of research activity. Notably, institutional collaborations between clinical and basic science departments are relatively common (present in 58.7% of multi-institutional publications), suggesting effective integration of clinical and laboratory perspectives despite the geographic concentration of research.

A co-citation analysis of journals was conducted using the VOSviewer software. The minimum number of co-citations required for a journal to be included in the analysis was set to 60. Consequently, 89 journals were screened for co-citation analysis. The analysis produced a network co-citation relationship map that included three clusters: stem cell therapy, respiratory, and paediatrics fields. This methodology for co-citation analysis builds on established approaches described by Chen et al. (2) for mapping research dynamics.

Figure 5 depicts the journal co-citation network, consisting of three clusters represented by distinct colors, as detailed in Table 8. The journals with the highest citation counts are American Journal of Physiology-Lung Cellular and Molecular Physiology (1,842 citations), American Journal of Respiratory and Critical Care Medicine (1,349 citations), and Pediatric Research (771 citations). All three journals are JCR 1 journals of excellence.

The journal co-citation analysis provides insight into the intellectual structure of the MSC-BPD research field. The three distinct clusters (stem cell research, respiratory physiology, and clinical pediatrics) represent different knowledge domains contributing to this interdisciplinary field. Visualization of connections between clusters reveals that while respiratory and pediatric journals show substantial interconnection (inter-cluster linkage strength =.68), the stem cell literature demonstrates fewer connections to these clinical domains (inter-cluster linkage strength =.41). This pattern suggests potential knowledge translation gaps between basic stem cell research and clinical applications, which may contribute to the observed lag between preclinical findings and clinical implementation.

A Vosviewer co-citation analysis of literature from 2004 to 2024 identified the top eight cited works in the field (Table 9), with three authored by the eminent scholar Chang YS, highlighting his significant contribution and influence. Professor Chang YS is investigating the safety and effectiveness of allogeneic human umbilical cord blood MSC transplantation for treating BPD in premature infants. Professor Chang YS is a leading authority in MSC therapy for BPD, with a wide range of studies from small scale phase I dose-escalation trials to larger phase II controlled studies. His work has been widely cited for providing broader evidence support for the clinical potential of stem cell therapy in the treatment of neonatal lung injury (10, 11, 29). Chang et al. (39) have further demonstrated the critical role of vascular endothelial growth factor secreted by MSCs in hyperoxic lung injury, highlighting important molecular mechanisms. The most cited article is Van Haaften et al.'s study (22), “Airway delivery of mesenchymal stem cells prevents arrested alveolar growth in neonatal lung injury in rats”, with 149 citations. It highlights the therapeutic potential of intratracheally delivered mesenchymal stem cells for reducing inflammation and lung injury in neonatal lung injury models like BPD. Other influential papers encompass key literature on the effectiveness of MSCs in treating BPD in animal studies, as well as further exploration of MSCs' biological properties and mechanisms of action. The seminal work of Aslam et al. (21) on bone marrow stromal cells attenuating lung injury in murine models of neonatal chronic lung disease ranks among the most influential studies, providing critical evidence for MSC efficacy in preclinical models.

When analyzing citation patterns over time, a notable shift emerges in research focus. Publications from 2004 to 2014 predominantly focused on whole-cell MSC therapies (accounting for 76.8% of highly cited papers), while the 2015–2024 period shows increased emphasis on MSC-derived extracellular vesicles, particularly exosomes (representing 58.2% of highly cited papers in this period). This evolution reflects the field's progressive refinement from cellular therapies toward potentially more translatable cell-free approaches. Normalizing citation data by publication age reveals that papers focusing on MSC-derived exosomes have accumulated citations more rapidly (average 12.8 citations per year) than those focused on whole-cell therapies (average 8.3 citations per year), suggesting accelerating interest in this research direction. This trend is exemplified by Willis et al. (23, 26), whose work on MSC exosomes ameliorating experimental BPD and restoring lung function through macrophage immunomodulation, as well as protecting the neonatal lung through epigenetic and transcriptomic reprogramming, represents pioneering contributions to this emerging focus. Similarly, Lithopoulos et al. (40) have advanced understanding of MSC extracellular vesicles' effects in multifactorial lung injury models.

Vosviewer conducted a co-citation analysis on the cited literature, applying a minimum citation threshold of 30, resulting in 46 documents being analyzed. Figure 6 illustrates the mapping of co-citation relationships. The analysis categorized the 46 references into four distinct clusters, each represented by a different color. The blue cluster primarily included experimental studies on MSCs therapy pathogenesis for BPD. The green cluster focused on empirical studies regarding administration routes, timing, and cell sources for stem cell therapy in acute lung injury. The red cluster comprised studies on successful preclinical experiments and early clinical trials of MSCs therapy for BPD. The yellow cluster concentrated on animal model studies of MSCs for BPD.

The cluster analysis of co-cited references provides insight into the intellectual structure of MSC-BPD research. Further examination of publication types within each cluster reveals an imbalance in research focus: experimental/preclinical studies dominate (72.4% of highly cited papers), while clinical studies represent only 18.6%, and methodological papers account for 9.0%. This distribution highlights a potential translational gap, where abundant preclinical evidence has not yet led to proportional clinical investigation. When analyzed chronologically, the average year of publication for preclinical papers (2,013.4) precedes that of clinical papers (2,017.8) by over four years, suggesting a substantial lag in clinical translation. This translational challenge is reflected in the work of Baker et al. (41), who discuss the current status and future therapeutic potential of stem cell therapy for lung disease.

The cited literature was clustered into the above groups using Citespace software (Figure 6B). The clusters identified include “mesenchymal stem cell”, “human mesenchymal stem cell”, “pulmonary disease”, “novel treatment approach”, “bone marrow”, “stem cell secretome”, “resident lung stem cell”, and “systematic review evidence map”, highlighting the primary focus areas in MSC therapy for BPD over the last twenty years. The emergence of “stem cell secretome” as a cluster theme aligns with the work of Alphonse and Thébaud (42), who examined growth factors and stem cells in BPD pathogenesis and treatment.

Figure 7 shows the keyword co-occurrence network view for the last five years, generated using Citespace. Node size reflects frequency of occurrence, with larger nodes indicating research hotspots. The most frequently occurring keywords are as follows: “bronchopulmonary dysplasia”, “mesenchymal stem cells”, “therapy”, “brain injury”, “preterm infants”, “respiratory distress syndrome”, “pulmonary hypertension”, “animal models”, “bone marrow”, “extracellular vesicles” and so on. The emergence of “brain injury” as a significant keyword reflects the growing recognition of MSCs' potential for addressing multiple complications of prematurity simultaneously, as demonstrated by Kim et al. (29), who showed that intratracheal MSC transplantation can attenuate both lung and brain injuries in hyperoxic newborn rats.

Temporal analysis of keyword occurrence reveals distinct research phases in the MSC-BPD field. From 2004 to 2010, frequently occurring terms primarily focused on “bone marrow”, “animal models”, and “lung development”, reflecting early foundational work. The 2011–2017 period saw increased emphasis on “inflammation”, “hyperoxia”, and “umbilical cord”, corresponding with expanded preclinical investigations into different MSC sources and mechanisms. Since 2018, “extracellular vesicles”, “exosomes”, and “microvesicles” have rapidly increased in frequency, indicating a shift toward cell-free therapeutic approaches. This evolution suggests the field is progressing from proof-of-concept studies toward more refined and potentially more translatable therapeutic strategies. The investigation of specific administration routes, as examined by Sung et al. (30), has contributed to understanding optimal delivery methods for MSC therapy.

In order to elucidate the research topics that have recently emerged as prominent in this field over the past five years (2019–2024), 12 key words were identified through the Bursts (Burst Words) analysis function of CiteSpace. Figure 7C illustrates that the distribution of these burst words in time is relatively even, with no instances of a sudden emergence of a large number of burst words in a specific year. Of the identified burst words, “microvesicles' exhibits the highest intensity. The emergence of “umbilical cord blood” and “NEC” from 2022 onwards suggests that this area has become a prominent research frontier in recent years. This interest in umbilical cord blood as an MSC source is reflected in Chang et al.'s (29) work demonstrating that human umbilical cord blood-derived MSCs can attenuate hyperoxia-induced lung injury in neonatal rats.

The burst detection analysis identifies terms that have experienced rapid increases in usage frequency, indicating emerging research frontiers. The term “microvesicles” shows the highest burst strength (8.45), followed by “exosomes” (6.21) and “umbilical cord blood” (5.87). The timing of these bursts (all occurring after 2020) suggests recent acceleration in interest surrounding cell-free therapies and specific MSC sources. The emergence of “NEC” (necrotizing enterocolitis) in 2022 indicates expanding interest in MSC applications beyond lung pathology to other neonatal conditions, potentially reflecting recognition of MSCs' broad immunomodulatory properties that could benefit multiple preterm complications simultaneously.