1 Introduction
Spinal cord injury (SCI) often leads to patients losing motor and sensory functions. About 180,000 new cases occur every year, which causes a heavy economic burden to society and families simultaneously.[
1] The prognosis of SCI is poor, and many people suffer the consequences throughout their lives. SCI is divided into traumatic and nontraumatic etiologies. The former type of SCI is more commonly seen.[
2] With the development of spinal deformity surgery, distraction SCI,[
3] which is the most dangerous iatrogenic injury problem along with the surgery, has become an urgent problem to be solved.[
4]
Since the 1990s, many scholars have begun to pay attention to the possibility of drug treatment for SCI.[
5,
6] The main idea is to control neural inflammation or promote nerve regeneration. Hormonal drugs represented by methylprednisolone (MP) have been widely used in the acute stage of SCI.[
7] However, criticisms suggest that the side effects outweigh the potential neural protection effect.[
8] In recent years, the concept of drug therapy has expanded, and new concepts, such as stem cell therapy and nanomedicine systems, have emerged, bringing hope for the drug therapy of SCI.
The bibliometric analysis method can synthesize existing research in related fields, explain the development of a subfield through the analysis of a large number of documents, reveal the cooperation and relationship between research fields, countries, and authors, and find current research hotspots. In this study, we adopt a broad operational definition of “pharmacotherapy” that encompasses not only traditional small-molecule drugs (e.g., MP, riluzole) but also advanced therapy medicinal products. This inclusion is justified by the evolving regulatory landscape where biologics, cell-based therapies, and exosome-derived agents are increasingly evaluated as “living drugs” or bioactive therapeutics aimed at modifying the pathological microenvironment.[
9] We conducted this bibliometric analysis to synthesize the development and changes in drug therapy for SCI to reveal the research hotspots and expand the current understanding of SCI drug therapy.
2 Materials and methods
2.1 Data acquisition and retrieval strategy
The Web of Science (WoS) database was used to search for publications on drug therapy for SCIs. We used the search terms “SCI”, “drug therapy” and “pharmacological treatment” to find the related studies in WoS core collection. No restriction on language was set. The retrieval time span was set from 1992 to 2025. This period was selected to capture over three decades of intellectual trajectory, beginning from the early 1990s when pharmacological interventions for SCI first gained significant global attention. The exclusion criteria were as follows: (1) meeting abstracts, editorials, letters, and corrections; (2) articles not related to biological or pharmacological interventions (e.g., purely biomechanical studies or epidemiological surveys without a treatment focus).
2.2 Bibliometric analysis and visualization
The analysis of the number of national, institutional, and journal publications and the research areas was realized through the “analyze results” of WoS. Journal impact factors and category quartiles were acquired from the 2024 journal citation reports (Clarivate Analytics, Philadelphia). The trend in publications over the years was visualized by Microsoft Excel 2016.
CiteSpace version 6.4.R1[
10,
11] was utilized to visualize the structural and temporal patterns of the scientific domain. We employed betweenness centrality to identify pivotal nodes (hubs) connecting different clusters and burst detection algorithms to capture rapid increases in scientific interest, indicative of emerging trends. WoS data were converted to txt format for further analysis. We analyze citations of publications, co-authorship of countries/institutions/authors, co-citation of authors/journals/references, and co-occurrence of keywords. In the visual map, various nodes represent countries/institutions/journals/authors/references/keywords. The size of the nodes represents the amount of publications/citations/frequency of occurrence; different colors represent clusters/the average appearing year; and the thickness of the line represents the strength of the link between the nodes. A dual-map overlay for journals was constructed. We conducted cluster analysis for co-cited references and keywords, and detection of references and keywords with intense citation bursts. The parameters of CiteSpace were set as follows: link retaining factor (LR = 3), look back years (LBY = 5),
e for the top
N (
e = 1), time span (1992–2025), years per slice (3), links (strength: cosine, scope: within slices), selection criteria (g-index:
k = 25), and minimum duration for keyword/phrase (
m = 2).
3 Results
3.1 Bibliometric analysis of annual publication output from 1992 to 2025
A total of 3,470 publications about drug therapy for SCI were identified in the WoS database between 1992 and 2025. The distribution depicting years of publication and the number of citations received is presented in Figure 1. The highest number of published studies was in 2025 (244 papers); the number of annual documents showed a linear growth trend. (y = 5.9176x, R2 = 0.966).
3.2 Distribution of countries and institutions
A total of 109 countries worldwide contributed to the research on drug therapy for SCI. As shown in Table 1, the United States (1,221 publications, 35.18% of all articles) is the most productive country, followed by China (807, 23.26%), Canada (252, 7.26%), and Italy (233, 6.71%). As shown in Figure 2, time-series analysis indicates that China has experienced explosive growth in research output in recent years (2022–2025), becoming the most dynamic research entity in this field. The overall network density (0.0959) reflects the formation of a highly mature and closely integrated international collaboration mechanism in this field.
The top ten productivity academic institutions were concluded in Table 1, with the University of California System (128 publications, 3.68%) and the US Department of Veterans Affairs (95 publications) as the leading institutions, eight of the top ten most productive institutions are based in the United States. Betweenness centrality analysis—indicated by purple rings—identifies critical “bridge” nodes that facilitate international knowledge exchange; the institutional hubs such as Ahvaz Jundishapur University of Medical Sciences (0.11) and the University of California San Diego (0.09) demonstrate high centrality despite not having the highest absolute publication counts. Furthermore, temporal evolution analysis using tree rings shows that while Western institutions possess long-standing research foundations, Chinese institutions, notably the Chinese Academy of Sciences and Wenzhou Medical University, have exhibited an explosive growth in activity during the recent 2022 to 2025 period, as evidenced by prominent red outer rings (Fig. 3).
3.3 Authors with research publications analysis
The top ten authors wrote 184 articles on drug therapy for SCI. Michael G. Feling from the University of Toronto published 31 papers with 2,505 citations. His highest cited article revealed that the primary mechanical trauma initiates a destructive secondary injury cascade—including neuroinflammation, ischemia, and apoptosis—which leads to progressive cell death and the formation of an inhibitory microenvironment composed of glial scars and cystic cavities. Adhering to the “time is spine” clinical paradigm, effective management necessitates rapid diagnosis, early surgical decompression, and hemodynamic stabilization to mitigate secondary damage and optimize long-term functional recovery.[
2] The top 10 cited papers and relevant information are listed in Table 2. The immense citation impact of these top articles is driven by two key factors: the enduring clinical reliance on comprehensive reviews regarding neuropathic pain management and SCI pathophysiology, and the explosive academic interest in emerging regenerative paradigms, specifically exosome-mediated therapies and functional biomaterials, which represent the field's cutting-edge direction.
3.4 Analysis of journals and research areas
Since 1992, 3,470 articles have been published in 652 journals. The top ten journals with the most cited counts are listed in Table 3. Journal of Neurotrauma received the top cited counts (1,710), followed by the Experimental Neurology (1,563) and Proceedings of the National Academy of Sciences of the United States of America (PNAS) (1,472). Overall, the identified publications are grouped into 82 research areas. As shown in Table 4, the most representative research area was neurosciences neurology (1,412 records, 35.73% of all articles), followed by pharmacology pharmacy (743, 18.80%), and biochemistry molecular biology (295, 7.46%). In addition, a dual-map overlay of journals was used to analyze the association of subject categories between citing and cited journals. The spline wave from left to right describes the citation path, and this interaction illustrates the linkage of different research areas. Two critical citation paths marked in yellow and pink indicate that articles published in the journals of Molecular/Biology/Immunology and Neurology/Sport/Ophthalmology are usually cited by articles published in Molecular/Biology/Genetics journals (Fig. 4).
3.5 Citation and co-citation analyses
The co-cited references were then clustered based on indexing terms. As shown in Figure 5, 11 main clusters were found: #0 SCI/neuropathic pain, #1 nerve regeneration/stem cells, #2 neuropathy pain/pharmacological treatment, #3 nerve injury/exosomes, #4 small fiber neuropathy/SCI, #5 biomaterial scaffolds/3d bioprinting, #12 cell therapy/neuroregeneration, #13 chronic pain/microglia, #14 clinical trial/central pattern generator, 15# isoform-specific nicotinamide adenine dinucleotide phosphate hydrogen (NADPH) oxidase inhabitor/morphine, and #18 central neuropathic pain/antioxidant. The S value for all the clusters above is over 0.8, indicating that the articles in each cluster have substantial homogeneity. The timeline view of all clusters is presented in Figure 6.
3.6 Co-occurrence analysis of keywords
We obtained a total of 679 keywords, of which the five with the highest frequency were SCI (1,670 times), neuropathic pain (461), spinal cord (408), functional recovery (309), and therapy (281). The chronological order is presented from dark to bright yellow (Fig. 7).
We detected the burst of keywords based on CiteSpace's algorithm of burst detection, where the minimum duration of the burst was set to five years. The discontinuous blue lines represent the timeline. The top 20 keywords with the highest burst strength are shown in Figure 8. The most intense keyword was randomized controlled trial (strength = 19.43), followed by drug therapy (14.07) and controlled trial (12.61). The keyword with the longest burst time was intrathecal baclofen, which lasted 25 years from 1993 to 2018. More meaningfully, the keywords “extracellular vesicles” and “microglia” had outbreak citations most recently (2022), indicating increasing research effort into biological therapies and neuroinflammation for SCI conditions.
4 Discussion
In this bibliometric study, the 3,470 literature on drug therapy for SCI published in the WoS database from 1992 to 2025 was analyzed using CiteSpace 6.4.R1 software. A steady linear increase in the total number of publications was observed, consistent with former studies.[
20,
21] The shifting landscape of research output reflects a transition from a unipolar structure dominated by Western institutions to a more diversified global network. While the United States maintains its historical leadership in citation impact and H-index, China's “explosive growth” from 2022 to 2025 marks a strategic pivot toward intensive investment in regenerative medicine. The explosive growth of publications from Chinese institutions correlates strongly with the rise of clusters related to exosomes and hydrogels, reflecting a strategic pivot toward bioengineering solutions in this region. However, the lower H-index of Chinese entities suggests a potential “quantity–quality gap” that may be bridged through the international collaboration mechanisms identified in our network analysis, where hubs like the University of Toronto and the University of California System act as critical knowledge bridges.
After clustering the selected literature based on the keywords, pain-related clusters occupy 4 of the 12 main clusters. It can be seen from the timeline map that painrelated research has formed the base theme of drug therapy in SCI patients since the 1990s. Ahuja CS et al.[
2] claimed that about 40% of SCI patients will experience pain at or below the injured segment, all those results indicating one goal of drug therapy for SCI is improving the quality of life of patients.
In 2010, Attal et al.[
22] and Finnerup et al.[
23] systematically summarized the evidence of treating neuropathic pain with tricyclic antidepressants, serotonin noradrenaline reuptake inhibitors, gabapentin, pregabalin, and opioids. The author believes that the current drug treatment is mainly to relieve the patient's pain.[
23] High-dose MP therapy is widely used to control inflammation in the acute phase of SCI. However, it may cause significant side effects such as gastrointestinal tract bleeding,[
24] aseptic necrosis of the femoral head,[
25] and the neuroprotective effect is not widely accepted. The effect of MP in treating SCI remains under debate.[
26–
29] This method remains in use because no other drug has successfully replaced MP therapy's effect.
Beyond symptom control, the diagram of drug therapy in SCI has shifted into neuroregeneration. Our citation analysis showed strong links between nerve regeneration, cell therapy, and biomaterial scaffolds clusters.
Current pharmacotherapy has evolved from broad-spectrum antiinflammatory suppression (e.g., MP) to precision microenvironment reprogramming.[
30] As highlighted by Hellenbrand et al.,[
18] the inflammatory response follows a critical timeline, suggesting that therapeutic interventions must align with specific cellular infiltration phases. This shift is evidenced by the citation burst of keywords “microglia” and “macrophage polarization”. Recent high-impact studies, such as Liu et al.,[
17] demonstrate that rather than merely suppressing inflammation, the goal is to orchestrate the M1-to-M2 phenotypic shift. Their work on hypoxic preconditioned mesenchymal stem cell-derived exosomes (HExos) revealed that targeting the TLR4/NF-κB axis can effectively resolve neuroinflammation while preserving reparative immune functions. This marks a transition towards “smart” immunotherapies that leverage specific molecular cargos to modulate rather than ablate the immune response.
A pivotal evolution in the 2022 to 2025 period is the integration of electroconductive biomaterials with bioactive nanocarriers, representing a shift from structural support to functional synergy.[
31] While earlier scaffolds primarily served as physical bridges, recent innovations, such as the electroconductive hydrogels developed by Fan et al.,[
19] mimic the electrical properties of native neural tissue to stimulate axonal regrowth via the PTEN/PI3K/AKT/mTOR pathway. Crucially, these advanced scaffolds are now being utilized as reservoirs for exosomes (Cluster #3), creating a dual-action system that simultaneously modulates the immune microenvironment and provides physical guidance for myelinated axon growth. This convergence of Cluster #5 (biomaterial scaffolds) and Cluster #3 (exosomes) signifies that future pharmacotherapy will likely rely on implantable, multi-functional platforms rather than systemic drug administration alone. Furthermore, the emergence of Cluster #15 (isoformspecific inhibitors) suggests that small-molecule drug discovery is moving toward higher specificity. Unlike broad antioxidants used in the past, targeting specific enzymatic isoforms (e.g., NADPH oxidase subtypes) may offer a way to mitigate oxidative stress without disrupting essential physiological signaling.
5 Limitations
This study has distinct limitations. First, relying solely on the WoS database may introduce database bias, potentially excluding relevant non-English publications, influential clinical trial reports indexed elsewhere. Second, the clustering analysis is inherently interpretative; while it identifies major trends, it may oversimplify the complex, multifactorial nature of certain transitional research areas. Finally, citation metrics inevitably lag behind the most recent discoveries and often favor positive outcomes, potentially underrepresenting studies with negative results or papers published in late 2025.
6 Conclusions
Based on the trajectory of citation bursts and cluster evolution identified in this study, we propose three evidence-based research priorities that will likely define the next decade of SCI pharmacotherapy:
6.1 The rise of “cell-free” and “smart” nanomedicine
Our analysis reveals a decisive shift from direct stem cell transplantation (Cluster #1) to exosome-based therapies (Cluster #3). While stem cells offer regenerative potential, they face hurdles regarding survival rates and tumorigenicity. Future research should prioritize engineered exosomes as superior alternatives. As demonstrated by the high-impact work of Liu et al.,[
17] exosomes derived from hypoxic preconditioned mesenchymal stem cells show enhanced cargo (e.g., miR-216a-5p) capable of precisely modulating microglial polarization. Future strategies should focus on modifying these nanocarriers to cross the blood-spinal cord barrier more efficiently and deliver specific miRNA cargos that target neuroinflammatory checkpoints.
6.2 Integration of electroconductive biomaterials
The emergence of Cluster #5 (
biomaterial scaffolds) highlights a critical gap in traditional pharmacotherapy: the lack of physical guidance and electrical signaling for regenerating axons. Drugs alone cannot bridge the lesion cavity. The future lies in combinatorial bio-hybrid systems. As evidenced by Fan et al.,[
19] the next generation of therapeutics will likely be electroconductive hydrogels loaded with bioactive factors (e.g., exosomes). These “smart” scaffolds perform a dual function: providing a mechanical bridge with electrical properties matching native spinal cord tissue to stimulate axonal regrowth, while simultaneously releasing immunomodulatory agents to resolve the glial scar.
6.3 Addressing the “acute-chronic” translational gap
A persistent disconnect exists between preclinical success and clinical failure. Our keyword analysis shows a heavy focus on “neuroprotection” (acute phase), yet the clinical reality involves patients in the chronic phase with established glial scars. As noted by Hu et al.,[
16] future clinical trials must move beyond broad-spectrum antiinflammatories. Research priority is shifting toward therapies that can remodel the chronic scar tissue and promote neural circuit reconstruction (intermediate neural networks) rather than solely focusing on acute neuroprotection.[
32] Furthermore, given the heterogeneity of human SCI, future pharmacotherapy should incorporate biomarkers to stratify patients for precise therapeutic windows, as the inflammatory timeline varies significantly between individuals.
The Author(s) 2026. Published by Wolters Kluwer Health, Inc. on behalf of Higher Education Press. This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.