1. Introduction

Diabetic macular edema (DME) is one of the ocular complications of diabetes mellitus (DM). It is also a major cause of vision loss in patients with diabetic retinopathy (DR). Given the lifelong disease of DM, macular edema can be persistent and have an extremely elevated recurrence rate after treatment.[1] According to statistics, the global age-standardized prevalence has been on the rise since 2019, and about 25% of patients could experience more severe vision loss in about 3 years without standardized treatment.[2] The macular area is like the “heart” of the eye, a sensitive area of the retina where 90% of the visual cells are concentrated, which is responsible for sharp central vision to show the colorful world. When DM occurs, the small blood vessels and microvessels will be damaged, the blood-retinal barrier (BRB) would be broken, resulting in increased leakage and fluid accumulation in the inner retinal layer of the macular area, finally causes the swelling of macular area.[3] Macular edema can lead to visual cell death and degeneration, and the visual acuity will be severely reduced and even blindness. In most studies, DME is defined as hard exudate in the presence of microaneurysms and punctate hemorrhages within the range of disc diameter of the fovea.[4] The clinically significant macular edema is the more severe type of DME which is defined as edema within 500 μm of the centeral foveal, or a focal photocoagulation scar in the macular area.[5]

The mechanism of etiology and pathogenesis in DME is really complex. Hyperglycemia is the main risk factor for DME, and the primary structural change is the disruption of the BRB. High glucose environment leads to the activation of protein kinase C, increased intraretinal vascular endothelial growth factor (VEGF) concentration, impairment of tight junction proteins, alteration of vascular permeability, and BRB damage, leading to hyperpermeability and vascular leakage.[6,7] Inflammation is a major pathological feature of DME. During the inflammatory process, the increased concentrations of pro-inflammatory cytokines and inflammatory mediators, combined with elevated VEGF levels, leukocyte activation, and stasis, could lead to increased BRB catabolism and vascular permeability.[8,9] For the primary management of DME in general, it was most beneficial to start strict glycemic control early in the onset of DM and before progression to DME.[10] The trial found that there was no significant effect in prognosis improvement by controlling the blood glucose, hyperlipidemia and nonce the patients progressed to DME.[11] A large number of publications about DME have been published in recent years. DME has been discussed from various aspects in these literatures, but there is a lack of systematic summary and overall analysis. Therefore, the research status of DME can be analyzed from the collected publications, which was conducive to researchers to grasp the current research hotspots and research directions, so as to further study DME.

Bibliometrics was an independent discipline formed in the late 1970s and widely used in document analysis.[12] Its analysis, with the help of modern computer technology, can generate a clear visual knowledge map, which provided a quantitative analysis method for the review and investigation of the existing literature in specific fields. Bibliometrics focuses on literature systems and characteristics, investigates hot spots and trends in the scientific literature, describes the relationship between cited and co-cited references, and describes the contributions of different authors, countries, institutions, and journals.[13,14] The Web of Science Core Collection (WoSCC) database has been recognized as one of the most suitable high-quality digital literature resources for bibliometric analysis by numerous researchers, and it has been widely used in various fields such as medicine, geology, ecology, etc. The CiteSpace and VOSviewer are popular software tools for visualizing literature information with their own advantages when performing mapping, and researchers often use them together to play a complementary role. Based on the papers collected by WoSCC as a data sample, this study performed a bibliometric analysis of the DME-related studies published in the last decade. A quantitative and visual analysis of the literature using bibliometric methods was conducted to demonstrate the current state of DME research, to reveal research hotspots and overall development trends, and to provide references for future related research.

2. Methods 2.1. Date collection

To improve the accuracy of the search, we obtained the subject headings from MeSH. We have set the search term as “diabetic macular edema” or “DME,” and searched the relevant literature in WoSCC, with the search period between January 1, 2012 and November 28, 2022. The retrieval method was advanced retrieval. The documents which defined as “Article and Review” in WoSCC were collected for analysis. There were no restrictions on language or document types. The documents unrelated to DME should be excluded. Other documents, such as meeting abstract, guidelines, letters or editorials, were excluded too.

2.2. Date analysis

After extracting the data, all documents were analyzed using Microsoft Office Excel 2019, VOSviewer (v.1.6.18), and CiteSpace (v.6.1.R3). Microsoft Office Excel 2019 was used to analyze the temporal distributions and identify highly cited DME-related studies over the past decade. Based on each set parameter, the included documents were processed by VOSviewer for large-scale data processing to construct a relational networks and visualize the knowledge maps of country/region, authors, research institutions, and discipline distributions. The CiteSpace software was used to visualize the basic knowledge and research hotspots in the field of DME research, to predict the development path and research frontiers of DME research. The CiteSpace parameters of time slicing was set as 2012 to 2022, and years per slice were set as (1), node type as cited reference, selection criteria as “K = 25,” and pruning was set as pathfinder. The flowchart of literature screening and processing is shown in Figure 1.

Figure 1.:

The flowchart of literature screening and processing.

3. Results 3.1. Temporal distribution map of the literature in publications and citation

A total of 6708 records comprising 11 types of documents were obtained. After excluding diseases unrelated to DME, 5165 documents which defined as “Article (n = 4347) and Review (n = 818)”in WoSCC were collected for analysis. The trend of publications of the retrieved literature was analyzed and shown in Figure 2. The number of publications on DME has generally increased in the recent decade, with the highest number published in 2020 and 2021. The number of publications was the highest in 2020 (684 of papers), but suddenly declined in 2022 (475 of papers). Herein, the number of citations is an important indicator of academic quantitative evaluation.[15] From 2012 to 2022, the frequency of literature cited increased year by year, with the highest number of cited articles published in 2021 (24,448 of papers). But the number suddenly declined in 2022. The overall trend was consistent with the number of publications. The temporal distribution of the 2 illustrates that the current research on DME is gradually becoming a research hotspot (Fig. 2). However, it will show a decreasing trend in 2022 due to the impact of the Corona Virus Disease 2019.

Figure 2.:

Trends in the growth of publications and citations worldwide from 2012 to 2022.

3.2. Country/region distribution

As shown in Table 1, the country with the highest number of published documents was the United States (US) with 28.75% of the total reports, followed by China and Italy, accounting for 14.54% and 7.09%, respectively. The total number of studies in these 3 countries accounted for half of the total number of reports, indicating that DME research has received wide attention in these 3 countries. England (54.97) and Italy (52.38) had the highest average reference times of single item, indicating that these 2 countries may be relatively mature and authoritative for DME researches.

Table 1 - Top 10 productive countries regarding the researches on DME. Rank Country Region Quantity Proportion/% ACI H-index Total link strength 1 USA North America 1485 28.75 49.25 475 1141 2 China East Asia 751 14.54 44.14 505 376 3 Italy South Europe 366 7.09 52.38 469 526 4 India South Asia 354 6.85 42.24 355 392 5 Japan East Asia 349 6.76 42.54 400 160 6 England West Europe 343 6.64 54.97 400 658 7 Germany Central Europe 319 6.18 48.63 400 560 8 Turkey East Europe 239 4.63 32.96 168 145 9 South Korea East Asia 234 4.53 35.32 166 109 10 France West Europe 209 4.05 40.16 532 438

ACI = average citations per item, DME = diabetic macular edema.

The VOSviewer parameter “Minimum number of documents for a country” was set to 10. The results were obtained from 100 countries, of which 44 reached the threshold value. Based on the co-occurrence analysis, VOSviewer divided the countries into 4 clusters. Figure 3 shows the cooperation among countries. The US worked mainly with Germany, Brazil, South Korea, and Sweden; China worked closely with Russia, Iran, Poland, Saudi Arabia, and the Czech Republic; and Italy tends to work more closely with Portugal, Switzerland, and Spain.

Figure 3.:

Cooperation map of countries of DME. The coloring is done according to the time course of their appearance, superimposing the time on the network of co-occurring countries. The different colors corresponded to different years, with the more purple color corresponding to an earlier appearance of the country, and the yellow color corresponded to a later. The closer 2 countries were located to each other, the stronger their correlation. DME = diabetic macular edema.

3.3. Distribution of authors and research institutions

As shown in Table 2, Wong Tien Yin from Tsinghua University was the author with the most published articles, followed by Jennifer K Sun from Harvard University and Ernesto Rodriguez from the JAEB Center for Health Research. Of the remaining 7 authors, 4 were from US universities and research institutions, 1 was from Austria, 1 was from Italy and 1 was from Israel. Although the overall authors were closely related to each other, no clear core group of authors was seen.

Table 2 - Top 10 authors in the studies of DME. Rank Author Country Institution TP ACI H-index 1 Tien Yin Wong China Tsinghua University 62 57.08 216 2 Jennifer K. Sun USA Harvard University 56 41.48 136 3 Ernesto Rodriguez USA JAEB Center For Health Research 55 17.25 48 4 Charles Wykoff USA The Methodist Hospital System 55 32.15 91 5 Neil M. Bressler USA Johns Hopkins Medicine 51 19.94 147 6 Ursula Schmidt-Erfurth Austria Medical University of Vienna 50 47.74 304 7 Srinivas Sadda USA Doheny Eye Institute 49 52.49 375 8 Francesco Bandello Italy IRCCS Ospedale San Raffaele 48 41.27 207 9 Quan Dong Nguyen USA Stanford University 44 41.11 141 10 Anat Loewenstein Israel Tel Aviv University 41 50.80 171

ACI = average citations per item, DME = diabetic macular edema, TP = total publications.

The VOSviewer parameters “Minimum number of documents of an author” with the method (Linlog/modularity), were set as 15. The results obtained were for 18,465 authors, and 98 of which met the thresholds. As shown in Figure 4, the different clusters represented the collaboration among authors. The results were shown in close collaboration between Wong Tien Yin, Mitchell Paul, Lamoureux Ecosse L, Sabanayagam Charumathi, Meriaudeau Fabrice, Klein Ronald. Glassman Adam R worked closely with Silva Paolo S, Gardner Thomas W, Melia Michele, and Aiello Lloyd Paul.

Figure 4.:

Cooperation map of authors.

As seen in Table 3, the University of California System (176 of papers) was the institution that published the most research papers. It was closely followed by the University of London (175 of papers) and Johns Hopkins University (173 of papers). University of London was the research institution with the highest ACI (57.91), followed by University College London (55.78) and Moorfields Eye Hospital NHS Foundation Trust (55.68). In terms of institutional affiliation, most of them originated from universities and research institutions, with fewer links to enterprises, hospitals, or other administrative institutions.

Table 3 - Top 10 institutions in the studies of DME. Rank Institution Country Quantity SOTC ACI H-index 1 University of California System USA 176 9091 51.65 475 2 University of London England 175 110,134 57.91 400 3 Johns Hopkins University USA 173 6554 37.88 292 4 Harvard University USA 157 8394 53.46 469 5 Johns Hopkins Medicine USA 148 5337 36.06 292 6 University College London England 144 8033 55.78 400 7 Harvard Medical School USA 139 7447 53.58 469 8 Moorfields Eye Hospital NHS Foundation Trust England 117 6514 55.68 395 9 Udice French Research Universities France 113 5000 44.25 532 10 Egyptian Knowledge Bank EKB Egypt 111 4905 44.19 343

ACI = average citations per item, DME = diabetic macular edema, SOTC = sum of times cited.

The VOSviewer parameters “the minimum number of documents of an institution” with method (Linlog/modularity), were set as 20. Results were obtained from 5184 institutions, 108 of which met the threshold. As shown in Figure 5, it could be seen that the institutions were relatively closely linked to each other, forming a cooperative relationship with Johns Hopkins University, Singapore National Eye Centre, University Sydney, and Universita Vita-Salute San Raffaele MILANO as the main clusters, and the DME was studied earlier in these organizations.

Figure 5.:

Cooperation map of research institutions.

3.4. Distribution of disciplines and journals

Regarding the number of published papers, Ophthalmology (58.10%), Computer Science Information Systems (7.18%) and Biochemical Research Methods (6.93%) were the top 3 disciplines as shown in Table 4. Since DME was the leading cause of irreversible vision loss in DM, it was more studied in specialties. However, cross-discipline collaboration was also likely to promote progress and win-win between disciplines. The study of Computer Science Information Systems and Biochemical Research Methods was also on the rise.

Table 4 - Disciplines within the top 20 in the studies of DME. Rank Quantity WoSCC categories Percentage/% 1 3001 Ophthalmology 58.10 2 371 Computer Science Information Systems 7.18 3 358 Biochemical Research Methods 6.93 4 328 Pharmacology Pharmacy 6.35 5 241 Multidisciplinary Sciences 4.67 6 215 Medicine Research Experimental 4.16 7 141 Biochemistry Molecular Biology 2.73 8 141 Surgery 2.73 9 132 Engineering Biomedical 2.56 10 105 Engineering Electrical Electronic 2.03 11 97 Radiology Nuclear Medicine Medical Imaging 1.88 12 88 Cell Biology 1.70 13 82 Optics 1.59 14 72 Health Care Sciences Services 1.39 15 69 Computer Science Interdisciplinary Applications 1.39 16 66 Computer Science Artificial Intelligence 1.28 17 63 Chemistry Multidisciplinary 1.22 18 62 Computer Science Information Systems 1.20 19 60 Biochemical Research Methods 1.16 20 51 Chemistry Medicinal 0.99 21 47 Computer Science Theory Methods 0.91

DME = diabetic macular edema.

As shown in Table 5, Retina the Journal of Retinal and Vitreous (305 of articles) was the most published journal in the field of DME, followed by Investigative Ophthalmology Visual Science (179 of articles) and Ophthalmology (137 of articles). The journals with the highest ACI were Clinical Medicine, Ophthalmologica, and Investigative Ophthalmology Visual Science in order. From the impact factors of the journals, most of them were dominated by Q1 and Q2, indicating that the relevant papers in DME were of high academic value.

Table 5 - Journals within top 20 the studies of DME. Rank Quantity Journal ACI IF (2022) JCR 1 305 Retina the Journal of Retinal and Vitreous 32.25 3.975 Q2 2 179 Investigative Ophthalmology Visual Science 41.23 4.925 Q1 3 137 Ophthalmology 32.44 14.277 Q1 4 128 Graefes Archive for Clinical and Experimental Ophthalmology 35.33 3.535 Q2 5 120 European Journal of Ophthalmology 30.52 1.922 Q4 6 118 Plos One 39.86 3.752 Q2 7 110 Clinical Ophthalmology 39.53 4.383 Q1 8 103 Scientific Reports 40.88 4.996 Q1/3 9 100 American Journal of Ophthalmology 33.01 5.488 Q1 10 98 Ophthalmic Surgery Lasers Imaging Retina 24.77 1.296 Q4 11 96 JAMA Ophthalmology 26.57 8.253 Q1 12 93 Journal of Ophthalmology 40.20 1.974 Q4 13 91 International Journal of Ophthalmology 31.75 1.645 Q4 14 89 Indian Journal of Ophthalmology 36.57 2.969 Q3 15 84 Ophthalmologica 43.11 3.757 Q2 16 82 BMC Ophthalmology 32.54 2.086 Q3 17 82 International Ophthalmology 31.46 2.029 Q3 18 69 Ophthalmology Retina 29.38 0.000 N/A 19 67 Eye 41.00 4.456 Q1 20 58 Journal of Clinical Medicine 58.79 4.964 Q2 21 58 Ophthalmic Research 32.09 3.031 Q3