No approval was required from the Institutional Review Board as data were retrieved from the Web of Science (WOS) database (https://www.webofscience.com/wos/woscc/basic-search) and no human subjects were involved.

The WOS is the most commonly used database in scientific and bibliometric research. It contains nearly 9,000 of the world’s most prestigious high-impact journals and more than 12,000 academic conferences, thus providing a comprehensive overview of the global research results in science, technology, medicine, and other fields (13, 14).

Publication information in the DMED area was searched within one day to ensure that no data were updated. The period was January 1, 2001, through September 30, 2022, and the retrieval date was October 1, 2022. The WOS Core Collection was selected, the subject words were “Diabetic erectile dysfunction”, and the article types were “article” and “review”. The retrieved results were exported in “Plain text file” format, and “Full Record and Cited References” was selected for “Record content”.

The search query string was described as follows: Results for “diabetic erectile dysfunction” (Topic) and Article or Review Article (Document Type) and Book Chapters (Exclude - Document Type).

CiteSpace software (Drexel University, Philadelphia, PA, USA), designed by Dr. Chen Chaomei in 2004 (11), is a freely available Java application widely used for the visualization and analysis of trends and patterns in scientific literature (15). CiteSpace (https://citespace.podia.com/download, R6.1.3) to scientometrics, data and information visualization technology as the foundation, through the analysis of the potential knowledge of literature, regularity and distribution, present knowledge structure. In this study, CiteSpace was used for keyword clustering and salient word analysis.

VOSviewer 1.6.18 (https://www.vosviewer.com/) is a software tool for constructing and visualizing bibliometric networks (16). In this study, VOSviewer was used to visualize countries/regions, authors, institutional collaborations, citations, and keyword co-occurrence, as well as to construct density maps.

The aim was to describe all the literature characteristics, including country/institution, journals, highly cited articles, cluster network of co-cited references, and most frequently cited keywords. In particular, in addition to noun phrases extracted from article titles and abstracts, burst detection was applied to the keywords of publications in the article collection assigned to the citation extension.

The number of annual publications is an important parameter for evaluating the development of scientific research and can reflect the growth of knowledge in this field to a certain extent. As of September 30, 2022, a total of 1,403 publications related to DMED were published ( Figure 1 ). The number of articles published per year is shown in Figure 2 . Although the trend was upward overall, it fluctuated slightly in some years. Notably, the number of publications per year can be easily divided into two phases. In the first phase, from 2001 to 2008, the average number of publications per year was around 30; in the second phase, from 2009 to 2022, the average number of publications per year stabilized at more than 50. Overall, the amount of knowledge in the DMED research field showed a linear growth trend (R ^{2 =} 0.6709), reflecting the increasing research interest in this field.

A total of 75 countries/regions have published articles in this field. The 10 countries that made the most significant contributions to DMED-related publications were the USA (294), China (244), Italy (111), South Korea (85), England (79), Turkey (73), Egypt (45), Japan (44), Spain (40), and Germany (38) ( Figure 3 ; Table 1 ). The size of the nodes is determined by the number of publications (the larger the number, the larger the node). The same colors represent the same clusters. The lines between nodes represent the alignment between countries/regions (the stronger the partnership, the wider the boundaries). The total link strength reflects the combined strength between countries/regions. As shown in Table 2 , the USA had the largest number of publications, the highest number of citations (15,620), and the greatest link strength (164). The USA (15,620), Italy (6,917), England (6,645), China (4,495), and Germany (2,381) were the top five countries in terms of citations. The above results indicated that these countries exerted the greatest influence on DMED-related research.

VOSviewer was used to analyze cooperation across countries, with lines between nodes indicating co-authorship between countries (the thicker the line the stronger the cooperation). China, the USA, Italy, South Korea, and England cooperated the most with other countries ( Figure 4 ).

A total of 1,465 institutions participated in the publication of papers related to DMED technology. The 10 institutions that contributed the most to DMED publications were INHA University (39), Huazhong University of Science & Technology (32), Tulane University (29), University of California San Francisco (25), Peking University (23), Seoul National University (20), Cairo University (20), University of Florence (19), Nanjing University (19), and Southern Medical University (18) ( Table 3 ). These results demonstrated that INHA University had the greatest number of publications with the highest number of citations (663) and the greatest link strength (20). Each node represents a different institution. The size of the node is determined by the number of publications (the larger the number, the larger the node). The same colors represent the same clusters. Boundaries between nodes represent a collaboration between organizations (the stronger the partnership, the wider the boundaries). The total link strength reflects the aggregate strength between institutions. As can be seen from the map, 146 institutions actively cooperated both within and between clusters. The three institutions with the highest total link strength were Peking University (34), the University of Florence (24), and the University of California San Francisco (23). The five institutions with the greatest number of citations were Eastern Virginia Medical School (4,067), Harvard University (3,845), The University of Pavia (2,092), the University of Manchester (1,684), and the University of Rome Tor Vergata (1,558) ( Table 4 ). Tulane University, University of Florence, University of California San Francisco, INHA University, and Huazhong University of Science & Technology were at the center of such collaborations. However, most institutions were fragmented and displayed little cooperation ( Figure 5 ).

Author co-occurrence analysis can identify the core authors in a field and the strength of collaboration among authors. Co-citation means that two authors or papers are cited by a third author or paper at the same time. This study included 5,216 authors and 20,297 co-cited authors. Of these, Ji-Kan Ryu (29), Jun-Kyu Suh (29), Wang Tao (26), Yin Guo Nan (25), and Liu Jihong (20) published the most articles ( Table 5 ). At the same time, Jihong Liu and Guiting Lin obtained the most cooperation, forming two stable author cooperation groups ( Figure 6 ). However, there was a lack of cooperation among other authors and teams, and the research was in a relatively scattered state. The results of the co-citation relationship showed that Bivalacqua, TJ (430); Corona, G (340); Feldman, HA (259); Rosen, RC (258); and Burnett, AL (213) were the most frequently cited authors ( Table 6 ), indicating that they play an important role in DMED research.

The selected papers were published in a total of 359 journals. The 10 journals with the most publications were the Journal of Sexual Medicine (156), the International Journal of Impotence Research (77), Andrologia (36), the Asian Journal of Andrology (35), Andrology (35), BJU International (30), the Journal of Urology (29), Diabetes Care (21), Urology (20), and PLoS One (19) ( Table 7 ; Figure 7 ).

A total of 4,634 journals were identified as being co-cited. The top five co-cited journals were the Journal of Sexual Medicine (3,656), the Journal of Urology (2,878), the International Journal of Impotence Research (2,058), Diabetes Care (1,676), and Urology (997) ( Table 8 ).

A total of 1,168 references and 29,269 co-cited references were obtained. Some papers were cited over 300 times, including those by Tesfaye (2010), Boulton (2005), Vinik (2003a), Gallagher (2007), Jurenka (2008), Jones (2011), and Kapoor (2007b) ( Table 9 ). In addition, a total of 19 references were obtained to highlight the analysis results. The three references with the highest intensities were Kouidrat Y, 2017 (doi: 10.1111/dme.13403), citation burst strength: 26.95; Malavige LS, 2009 (doi: 10.1111/j.1743-6109.2008.01168.x), citation burst strength: 21.88; and Rendell MS, 1999 (doi: 10.1001/jama.281.5.421), citation burst strength: 19.24 ( Figure 8 ).

Changing trends in research topics over time can be identified through keyword co-occurrence and salience analysis to better grasp the development of research hotspots. A total of 3,969 keywords were obtained, the top 10 of which were erectile dysfunction (796), men (256), diabetes (254), diabetes mellitus (239), prevalence (180), corpus cavernosum (171), dysfunction (155), mellitus (154), nitric oxide synthase (153), and expression (140) ( Table 10 ; Figure 9 ).

After clustering using CiteSpace software, a total of 18 categories of keywords were obtained. From 2001 to 2011, keyword-based research hotspots in the DMED field mainly included impotence, oral sildenafil, smooth muscle relaxation, nitric oxide synthase, epidemiology, relaxation, gene therapy, in vivo, type 2 diabetes, metabolic syndrome, and sexual function. From 2014 to 2022, meanwhile, the research hotspots (by keyword) included cavernous nerve injury, diabetic rat, pathway, radical prostatectomy, stem cell, apoptosis, model, pathophysiology, and penile prosthesis, among others ( Figure 10 ).

Over the past 20 years, the number of DMED-related articles published in journals showed a linear upward trend (R ^{2 =} 0.6709).

From the perspective of countries/regions and institutions, the number of publications originating from the United States as well as the number of associated citations far exceeded that of other countries. Although the number of articles published in China ranked second, the number of citations was low, ranking only fourth. This indicates that, although the number of papers from Chinese institutions is increasing yearly, there is still a lack of high-quality articles. This may be partly explained by a lack of cooperation with internationally renowned researchers as well as language barriers. Among the 10 institutions with the most published papers, only Huazhong University of Science & Technology, Peking University, and Nanjing University were from China—with the remainder being mainly from Europe and the USA—and these institutions had low levels of research cooperation. These observations highlight the need to strengthen communication and cooperation with global research teams, especially those from countries and institutions in the Asian region.

Regarding authors and references, Ryu, Ji-Kan and Suh, Jun-Kyu had the greatest production efficiency, while Bivalacqua, TJ had the most co-citations. Corona, G; Feldman, HA; Rosen, RC; and Burnett, AL also made important contributions to DMED research. The research undertaken by Ryu, Ji-Kan and Suh, Jun-Kyu is mostly mechanism-related. The former has mainly focused on identifying potential target genes associated with DMED in mouse cavernous pericytes (17). Meanwhile, Suh, Jun-Kyu found that pericellular-derived extracellular vesicles mimicking nanovesicles can promote neurovascular regeneration and erectile function recovery under diabetic conditions through a lipid chain enzyme 2 (LCn2)-dependent mechanism (18). Both authors reported that the preservation of damaged cavernous neurovasculature through the inhibition of the proNGF/p75NTR pathway may represent a novel therapeutic strategy for radical prostatectomy-induced erectile dysfunction (19).

The Journal of Sexual Medicine published the most DMED-related papers and the 10 most cited and co-cited journals were mainly in Q1 and Q2. This indicates that most papers published in this field were high-quality scientific research achievements.

Our analysis further demonstrated that the most influential authors and references were mostly associated with review articles and clinical guidelines from internationally renowned institutions and journals. Combining keyword co-occurrence, clustering, and salience analysis, we identified oral sildenafil, smooth muscle relaxation, acute oxide synthase, gene therapy, metabolic syndrome, cavernous nerve injury, stem cell, and penile prosthesis as the main research topics and hotspots in the field of DMED.

The main treatment for erectile dysfunction is drug therapy, including phosphodiesterase-5 (PDE5) inhibitors, androgen therapy, and vasoactive drugs (20–22). PDE5 inhibitors, such as sildenafil, tadalafil, and vardenafil, are the first-line oral drugs recommended by the World Health Organization (WHO) for the treatment of erectile dysfunction, including that associated with diabetes (23–25). These drugs increase the concentration of cyclic guanosine monophosphate (cGMP) in vascular smooth muscle cells by inhibiting PDE5 expression in the corpus cavernosa, reduce the concentration of intracellular calcium, cause smooth muscle relaxation, increase blood flow to the corpus cavernosa, and improve erection. Treating DMED with a PDE5 inhibitor can improve the International Index of Erectile Function-5 (IIEF-5) score and sexual success in a significant number of patients (26, 27).

Nevertheless, PDE5 inhibitors are ineffective in patients with severe erectile dysfunction (28) and are occasionally associated with side effects such as headaches, flushing, indigestion, nasal congestion, vision abnormalities, and diarrhea. The prerequisite for PDE5 inhibitor to function is that ED patients must have intact molecular and neurological pathways and some degree of sexual stimulation. Therefore, PDE5-Is has poor efficacy in patients lacking upstream nitric oxide (NO) pathway (29) in some disease states, such as diabetes with peripheral neuropathy (30) and prostate cancer (31). In addition to drugs, other alternative strategies are available, such as vacuum contractile devices and penile prosthesis implantation The three-piece inflatable penile prosthesis (IPP) is the most common implant used in penile surgery (32).The satisfaction rate of patients is 90% ~ 100%, which varies with different prosthesis devices (33, 34).However, IPP may only be offered to patients who fail due to the high cost of conservative treatment, its invasive nature, and myriad potential complications. Thus, novel therapeutic approaches are urgently required to overcome these disadvantages in DMED treatment.

Stem cell therapy has attracted increasing attention owing to its ability to promote functional recovery and tissue structural repair in patients with diabetes. Stem cell mainly include bone marrow mesenchymal stem cells (BMSCs) (35), adipose-derived stem cells (ADSCs) (36), neural derived stem cells (37), mesenchymal stem cells (MSCs) (38), umbilical cord derived stem cells (39) and urine-derived stem cells (USCs) (40).Mesenchymal stem cells (MSCs) are adult stem cells derived from mesoderm with a high potential for proliferation, self-renewal, and multidirectional differentiation (41). The culture of bone marrow-derived MSCs is relatively simple; accordingly, they are widely used in animal studies and clinical trials (42). Given their abundant autologous availability, MSCs are one of the most promising candidates for the treatment of DMED (43). Importantly, allogeneic MSCs can also be successfully transplanted owing to the low immunogenicity (44, 45). In animals, bone marrow-derived MSC transplantation can significantly improve the recovery of erectile function by inhibiting apoptosis (46). For DMED, MSC-based treatment has also attained encouraging therapeutic effects (47). The restoration of erectile function is mainly attributed to the increased endothelial cell and smooth muscle content in the corpus cavernosum (48, 49). Various strategies have recently emerged to improve the therapeutic effect of MSCs on DMED (50, 51).In a single-blind study, 6 out of 7 ED patients developed morning erections 3 months after treatment with umbilical cord derived stem cells (39).Chen S et al. found that adipose mesenchymal stem cells were more effective in treating diabetes-related erectile dysfunction than bone marrow mesenchymal stem cells (52).

Erectile dysfunction has been associated with the upregulation of Toll-like receptor 4 (TLR4) expression. It has been reported that TLR4 expression is significantly increased in the corpus cavernosum of diabetic rats compared with that of controls (53). The overexpression of inducible nitric oxide synthase (iNOS) may also be associated with penile microvascular dysfunction in diabetes mellitus, while endotoxemia has been linked with iNOS upregulation (54). Studies on conducting vessels have shown that iNOS overexpression in endotoxemia contributes to endothelial dysfunction by reducing the activity of endothelial nitric oxide synthase (eNOS) (55). iNOS inhibitors can alleviate DMED-related injury by moderating eNOS phosphorylation and chronic iNOS overexpression as well as improving microvascular fibrosis (56).

Trabecular smooth muscle relaxation during sexual stimulation and dilatation of penile resistance arteries in the cavernous body are essential for penile erection (57). Nitric oxide (NO) is a key factor in both processes. NO can be released from nerve endings or endothelial cells to stimulate cGMP production in penile smooth muscle cells, relaxing them and increasing blood flow to the corpus cavernosum (58, 59). Defects in the NO/cGMP pathway at any level will lead to inadequate penile smooth muscle relaxation, which affects erectile function (60).

In vitro low-intensity shockwave therapy (LI-SWT) is a potential treatment option for ED. The micro-damage of Li-SWT to the spongy tissue may stimulate neovascularization and up-regulate some factors related to tissue healing and remodeling (61).A prospective, randomized, sham controlled study reported a clinically significant improvement in erectile function in 40.5% of the treatment group based on the minimum clinically significant difference (MCID) criteria (62).Another randomized clinical trial evaluated changes in penile hemodynamics and IIEF-EF scores in patients with vascular ED. In the IIEF-EF score, 56.7% of the treatment group achieved MCID at 1 month and 75% achieved MCID at 12 months (63).

Through keyword co-occurrence and salience analysis, we also found that there is a regional imbalance in the development of DMED research and influential authors and institutions are mainly concentrated in Europe and Asia, which may be due to greater research interest. Additionally, despite the numerous studies undertaken on DMED, the underlying mechanism remains incompletely understood and represents a key future research direction.

To the best of our knowledge, this is the first bibliometric analysis of DMED-related studies undertaken in the last 20 years. However, our study still had some limitations. First, high-quality articles published in recent years may not have yet reached the ideal citation threshold, which may have introduced research bias. Second, there may be a delay in exploring the frontiers of research. Finally, we only included literature written in English from the WOS, which may have resulted in the omission of important documents in other languages.

Overall, our bibliometric analysis provides comprehensive information on DMED-related publications. Our findings suggested that the field of DMED is thriving and has stimulated great interest in the research community worldwide. Although DMED is a relatively common disease, much remains unknown about both the underlying mechanisms and how to treat it. Likely, the terms oral sildenafil, smooth muscle relaxation, nitric oxide synthase, gene therapy, metabolic syndrome, cavernous nerve injury, stem cell, and penile prosthesis will be at the forefront of DMED-related research.