A History of Anti-VEGF Inhibitors in the Ophthalmic Literature: A Bibliographic Review

Abstract

Purpose:

This work aimed to analyze the 100 most-cited articles on antivascular endothelial growth factor (anti-VEGF) inhibitors.

Methods:

A literature search for anti-VEGF inhibitors using the Web of Science was completed using the bibliographic databases for peer-reviewed literature published in Ophthalmology, the New England Journal of Medicine, Journal of the American Medical Association, and Lancet. Primary outcomes were the most frequently cited articles and journals with the most citations as well as the specific drug and disease process studied.

Results:

There were 42 696 cumulative citations among the top 100 articles. The articles were published between 2004 to 2016, with most articles published in 2006. Ophthalmology published the greatest number of articles among the top 100 at 48, whereas the New England Journal of Medicine has the most citations per publication at 1714. Ranibizumab was the medication researched in most articles at 56, followed by bevacizumab at 48, aflibercept at 10, and pegaptanib at 9. Forty-two articles investigated treatment of age-related macular degeneration, followed by 24 investigating diabetic macular edema, 10 for retinal vein occlusion, 8 for proliferative diabetic retinopathy, 2 for retinopathy of prematurity and polypoidal choroidal vasculopathy, and 1 for corneal neovascularization.

Conclusions:

As evidenced by the considerable number of citations accumulated over the past 20 years, anti-VEGF inhibitors have led to significant research in ophthalmology.

Introduction

Antivascular endothelial growth factor (anti-VEGF) medications have revolutionized the field of vitreoretinal surgery and medical retina with their efficacy and numerous indications. Anti-VEGF medications began to emerge within the literature in 2002, with the evaluation of bevacizumab’s efficacy in breast and colorectal cancer treatment. ^{1 -4} Its use in the eye started with pegaptanib in 2002, ⁵ ranibizumab in 2003, ⁶ and aflibercept in 2007. ⁷ Initially, indications for ophthalmologic use were exclusively for age-related macular degeneration (AMD). ⁸ Eventually, studies were released that showed their efficacy in diabetic macular edema (DME) and retinal vein occlusion (RVO) in 2005 and retinopathy of prematurity (ROP) in 2007. ^9,10

The purpose of this study was to identify trends in the 100 most-cited papers on anti-VEGF medications since their advent nearly 20 years ago to provide a historic perspective on these medications and illustrate their importance within ophthalmic research.

Methods

The Web of Science, maintained by Clarivate Analytics and previously known as the Web of Knowledge, was used for this bibliographic analysis because of its sophisticated ontology and analysis capabilities both in science and medicine. ¹¹ The Web of Science contains more than 90 million records and covers more than 12 000 high-impact journals. ¹² The Web of Science, the first web-based citation indexing resource of scholarly and professional journals, is one of the most trusted and authoritative sources for bibliometric data for peer-reviewed global research knowledge across disciplines. ¹³

Our search included the generic and trade names of the 5 most-used anti-VEGF medications in ophthalmology. The search terms used include pegaptanib (Macugen, Bausch + Lomb), bevacizumab (Avastin, Genentech, Inc), ranibizumab (Lucentis, Genentech, Inc), aflibercept (Eylea, Regeneron Pharmaceuticals Inc), and brolucizumab (Beovu, Novartis Pharmaceuticals Corporation). Trade names were searched in addition to generic names; however, no results exclusively contained the trade name without the generic name.

The bibliographic search was completed in January 2020. No specific date parameter was set for the search. Four searches were performed in total, all of which were refined to include research articles exclusively. The first was within the Web of Science category “ophthalmology,” which searches the peer-reviewed journals in the field of ophthalmology (59 journals), and then 3 separate searches were completed within the New England Journal of Medicine (NEJM), Lancet, and the Journal of the American Medical Association, including JAMA Ophthalmology. The goal of these additional searches was to expand the initial search of just ophthalmology journals to collect articles from high-impact journals related to anti-VEGF medications and ophthalmology.

Articles were compiled from each search and sorted by the number of times cited in all databases. Further details were investigated for each article, including year published, average citations per year since publication, first and senior (or last) author, institutions and countries affiliated with both first and senior author, number of authors, which medications were referenced, disease category, number of patients studied, type of article, and source title.

Our primary outcomes were to obtain the most frequently cited articles and which journals had articles with the most citations, as well as the type of drug and disease process studied. Our secondary outcomes included the number of authors, the number of patients, institutions of the first and last authors, and countries of the authors.

Certain articles were eliminated from the search despite their number of citations being among the top 100, because they either did not pertain to ophthalmology or an anti-VEGF medication was not the subject of the article. ^{14 -17} For example, articles about target oxygen levels in ROP ¹⁸ or epidemiologic studies on diabetic retinopathy were eliminated. ¹⁹

For statistical analysis, Pearson correlation coefficients were derived between continuous variables. Calculations were performed using Sheets, a Google-operated spreadsheet.

Results

There were 42 696 cumulative citations among the top 100 articles (Supplemental Tables 1 and 2). The range of citations between least-cited and most-cited articles involving anti-VEGF medications was 169 to 3579. The median and mean were 266 and 427, respectively, with a skew in article distribution due to 25% of citations coming from the top 7 of the 100 most-cited articles. Figure 1A displays the disproportionate amount of citations among the most-cited articles. Figure 1B is a graph of cumulative citations per year for the previously identified top 100 articles. It shows an initial sharp increase in citations after the first articles were published, followed by a second consistent increase until the year 2015, after which the number of times cited per year decreases.

Figure 1.

(A) Number of citations vs rank of corresponding article. (B) Cumulative citations of all top 100 articles per year.

The articles on our list span the years 2004 to 2016 and have a cumulative average of 37 citations per year. There was a strong relationship between total citations and average citations per year at a correlation coefficient of 0.94. However, the relationship between total citations and number of years since publication has a correlation coefficient of 0.15. This discrepancy suggests that although the time since publication was critical in the accumulation of citations, it is not the only factor that contributed to an article’s end-total number of citations. Graphing the number of articles published per year resulted in a bimodal distribution as shown in Figure 2A. This was further broken down by the article’s medication(s) of topic in Figure 2B. Ranibizumab was the medication used in most articles (56 articles), followed by bevacizumab (48), aflibercept (10), and pegaptanib (9). Fifteen articles focused on more than 1 medication. Brolucizumab was also included in the search but was not among the list of the top 100 most-cited articles, likely because of its recent approval by the US Food and Drug Administration (FDA).

Figure 2.

(A) Number of articles from each publication year. (B) The number of times medications were referenced.

The 100 articles were published among 12 journals. Table 1 provides a list of the journals sorted by the number of articles they published on the top 100 list along with their cumulative number of citations. Although the most articles were published in Ophthalmology, the NEJM produced approximately 3 times as many citations for each of its publications. The top 4 most-cited articles were from the NEJM, followed by Lancet with the fifth most-cited article, and finally Ophthalmology rounded off the remaining 5 spots of the top 10 most-cited articles. A bibliographic review performed in 2019 of all anti-VEGF—related articles in all fields reported that Investigative Ophthalmology & Visual Science published the most articles in ophthalmology research. ²⁰ Among the top 100 from our search, however, Investigative Ophthalmology & Visual Science produced just 4 of the articles.

Table 1.

Journals’ Publications and Citations of the Top 100 Articles on Anti-VEGF Inhibitors for Ophthalmic Use.

No. of publications		Total citations per journal
Ophthalmology	48	Ophthalmology	17 616
American Journal of Ophthalmology	14	New England Journal of Medicine	10 286
Retina—The Journal of Retinal and Vitreous Diseases	12	Archives of Ophthalmology	4221
New England Journal of Medicine	6	Retina—The Journal of Retinal and Vitreous Diseases	3695
British Journal of Ophthalmology	5	Lancet	2094
Investigative Ophthalmology & Visual Science	4	British Journal of Ophthalmology	1321
Archives of Ophthalmology	3	Ophthalmic Surgery Lasers & Imaging	1100
Lancet	3	Investigative Ophthalmology & Visual Science	982
Ophthalmic Surgery Lasers & Imaging	2	Archives of Ophthalmology	639
Current Opinion in Ophthalmology	1	Graefe’s Archive for Clinical and Experimental Ophthalmology	335
Eye	1	Eye	224
Graefe’s Archive for Clinical and Experimental Ophthalmology	1	Current Opinion in Ophthalmology	180

The breakdown of the diseases addressed in the top 100 articles can be seen in Table 2. We categorized the articles into 4 research types. These types, in order of most frequently used among the top 100, were clinical trials, clinical experience, basic sciences, and other. Their distribution is represented in Figure 3. The clinical experience research category encompassed case series (15 articles), retrospective studies/analyses (8), and cohorts (3). The basic science category comprised animal and physiologic studies (11 articles) that encompassed toxicity and pharmacokinetics. The “other” category consisted of methods not distinguished by the previously mentioned categories such as reviews (4 articles), a simulation (1), a meta-analysis (1), and a survey (1).

Table 2.

Number of Articles Per Disease.

Disease	No. of publications
Age-related macular degeneration	42
Diabetic macular edema	24
Vein occlusion	10
Proliferative diabetic retinopathy	8
Retinopathy of prematurity	2
Polypoidal choroidal vasculopathy	2
Corneal neovascularization	1

Figure 3.

Distribution of types of articles.

The number of patients in each study ranged from 0 to 5228. Seventeen studies did not have patients and represented animal studies (11 articles), reviews (4), a meta-analysis (1), and a simulation (1). More than 50% of the 40 859 cumulative patients came from the 10 most patient-populated studies, resulting in a mean of 492 patients compared with the median of 189 patients.

The number of authors varied from 1 to 23. The median and mean number of authors for the articles had little variability at 8 and 8.2, respectively. There were 66 different first authors compared with 75 different senior authors, despite 5 of the articles having only 1 author. Table 3 provides the number of articles published, authorship, and the institution affiliated with those authors who published at least 3 articles among the top 100 list. No correlation existed between the number of citations an article received and the number of patients in a study (r = 0.13) or the number of authors (r = –0.01). Forty of the 100 articles were written by 23 study groups. Seven of the 10 most-cited articles were affiliated with a study group. Table 4 shows the 11 study groups that wrote at least 2 articles.

Table 3.

Authors With At Least 3 Articles, Their Institution, and Their Order of Authorship.

Author	No. of articles	Authorship	Institution
Brown, David M.	7	First and senior	The Methodist Hospital System
Heier, Jeffrey S.	6	First and senior	Ophthalmic Consultants of Boston
Rosenfeld, Phillip J.	6	First and senior	University of Miami
Campochiaro, Peter A.	5	First and senior	Johns Hopkins University
Nguyen, Quan Dong	5	First and senior	University of Nebraska and Johns Hopkins University
Avery, Robert L.	4	First and senior	California Retina Consultants
Schmidt-Erfurth, Ursula	4	First and senior	University of Vienna
Elman, Michael J.	4	First	Elman Retina Group
Chakravarthy, Usha	3	First	Queens University Belfast
Spaide, Richard F.	3	First	Manhattan Eye, Ear & Throat Hospital and Vitreous Retina Macula Consultants of New York
Martin, Daniel F.	3	First and senior	Cleveland Clinic
Rubio, Roman G.	3	Senior	Genentech, Inc
Shams, Naveed	3	Senior	Genentech, Inc

Table 4.

Study Groups With At Least 2 Articles.

Study group	No. of articles
Diabetic Retinopathy Clinical Research Network	6
VISION Clinical Trial Group	3
MARINA Study Group	3
CATT Research Group	3
Tubingen Bevacizumab Study Group	2
RISE and RIDE Research Group	2
READ-2 Study Group	2
Pan-American Collaborative Retina Study Group	2
Macugen Diabetic Retinopathy Study Group	2
IVAN Study Investigators	2
ANCHOR Study Group	2

There were 69 different institutions represented between first and senior authors. The institutions that were responsible for at least 4 of the articles among the top 100 list are listed in Table 5. Genentech, Inc was affiliated with the first author in just 1 paper but had a senior author in 21. The countries of origin of the first and senior authors were combined and analyzed. Authors from the United States wrote 70% of the articles. A list of the countries affiliated with either the first or senior author that produced at least 3 articles can be seen in Table 6.

Table 5.

Number of Articles per Institution.

Institution	No. of articles
Genentech, Inc	22
University of Miami	15
Johns Hopkins University	10
Ophthalmic Consultants of Boston	8
The Methodist Hospital System	7
California Retina Consultants	5
Mayo Clinic	5
Vitreous Retina Macula Consultants of New York	5
University of Vienna	4
Elman Retina Group	4
University of London	4
National University of Singapore	4
Cleveland Clinic	4
Eyetech Pharmaceuticals Inc	4
Harvard University	4
Osaka University	4
Tulane University	4

Table 6.

Countries of Authors With At Least 3 Articles.

Country	Total no. of articles
United States	137
Germany	10
England	6
Japan	6
Singapore	6
Austria	4
Switzerland	4
Ireland	3
Venezuela	3

Conclusions

VEGF is implicated in neovascularization that occurs in many diseases of the eye. The binding of VEGF to its tyrosine kinase receptor and the subsequent signaling cascade that occurs have proven to be rate-limiting in the complex process of angiogenesis throughout the body. ²¹ When the allele for VEGF is absent, angiogenesis does not occur, emphasizing the necessity for VEGF. ²² Retinal pigment epithelium and the retina produce and release VEGF while under metabolic distress. Metabolic changes such as oxidative stress, changes in the Bruch membrane, and reduction in blood flow have been shown to promote the release of VEGF within the eye. ²³ VEGF promotes vascularization as a chemoattractant for endothelial cell precursors by preventing endothelial cell apoptosis, and by promoting metalloproteinases that cause tissue degradation and allow invasion of new vessels. ^24,25 Unfortunately, there can be adverse downstream effects of these mechanisms, such as increased vascular permeability, aberrant vessel growth, and vessel occlusion. ^24,26,27

Anti-VEGF medications have seen rapid development and increase in use since the early 2000s. Bevacizumab (Avastin) is a humanized antibody designed to block all isoforms of VEGF. Genentech of San Francisco, California, started phase 1 trials for bevacizumab in 1997, which led to its FDA approval for colon cancer in 2004. ²⁸ Bevacizumab has never been granted FDA approval for ophthalmologic use, but it has been used off label for AMD in response to the results of a small study that showed improvement in visual acuity (VA), optical coherence tomography findings, and angiographic findings. ²⁹

Pegaptanib (Macugen) was the first anti-VEGF medication approved for ophthalmic use by the FDA for AMD in 2004 after 2 large studies within the VISION (VEGF Inhibition Study in Ocular Neovascularization) Clinical Trial Group displayed its efficacy. It is an aptamer (single strand of nucleic acid) that specifically binds to the 165 isoform of VEGF and was licensed by OSI Pharmaceuticals of New York, New York, in 2000. ^30,31

FDA approval for use of ranibizumab (Lucentis) in the treatment of AMD was granted in 2006 based on the MARINA (Minimally Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular AMD) and ANCHOR (Anti-VEGF Antibody for the Treatment of Predominantly Classic CNV in AMD) clinical trials, both of which were published by the NEJM and are listed among the top 100. ^32,33 Ranibizumab is an antigen-binding fragment with affinity for the receptor-binding site of VEGF-A. Genentech developed this medication in response to the thought that bevacizumab may have poor retinal and choroidal penetration, as well as the concern that the larger molecule size of bevacizumab might obstruct the trabecular meshwork and lead to intraocular pressure spikes. These initial concerns regarding bevacizumab were eventually proven false. ^{34 -36}

The VIEW (VEGF Trap-Eye: Investigation of Efficacy and Safety in Wet AMD) Clinical Trial Group studies were critical in FDA approval of aflibercept (Eylea) in 2011 after demonstrating the need for less frequent dosing and noninferiority to ranibizumab. Aflibercept (Eylea) is a chimeric fusion protein that, contrary to the prior medications, does not directly antagonize the VEGF receptor. Instead, it shares domains with the VEGF receptor, making it a decoy that sequesters VEGF. Because of this novel mechanism of action, it is often referred to as a “VEGF-trap.” Aflibercept was developed by Regeneron Pharmaceuticals of Tarrytown, New York, to improve the pharmacokinetics of contemporary VEGF medications by way of improving on half-life and binding affinity. ³⁷

Brolucizumab (Beovu) is an anti-VEGF medication by Novartis that was approved by the FDA in 2019 for the treatment of AMD. This medication, however, was not among the 100 most-cited articles involving anti-VEGF medications. This can likely be attributed to the release of its preliminary phase 3 data being as recent as 2017, when Novartis presented the HAWK (Efficacy and Safety of RTH258 versus Aflibercept – Study 1) and HARRIER (Efficacy and Safety of RTH258 versus Aflibercept – Study 2) trials at the American Academy of Ophthalmology conference. ³⁸ Within the Web of Science, the HAWK and HARRIER publication, which was published in January 2020, just 1 month before this review was written, had already been cited 12 times. This number extrapolated to a year is far higher than 37, the average annual number of citations among the top 100, suggesting that the HAWK and HARRIER publication has the potential to be among the 100 most significant articles.

Currently, AMD, DME, and RVO are the only ophthalmologic diseases for which the FDA has approved anti-VEGF medications. ³⁹ Vision loss in AMD is due to disruptive proliferation of vasculature below the retina, ⁴⁰ which was the first pathologic process targeted by anti-VEGF medications. ^{31 -35,37,38,41} The VISION trial (2004) was one of the first trials for anti-VEGF in AMD and demonstrated pegaptanib’s efficacy for subfoveal corneal neovascularization (CNV). ³¹ The MARINA (2006) and ANCHOR (2006) trials both were pivotal phase 3, multicenter, randomized trials that evaluated the efficacy of ranibizumab. The MARINA study found that ranibizumab was superior to sham for occult CNV, and the ANCHOR trial found that ranibizumab was superior to verteporfin for predominantly classic CNV. ^32,33 The CATT (The Comparisons of AMD Treatments Trials) study (2011) went on to show equivalent efficacies between Genentech’s bevacizumab and ranibizumab. ⁴¹ The VIEW studies (2012) evaluated the efficacy of aflibercept to ranibizumab and found equivalent efficacy. ³⁷ The IVAN (The Inhibition of VEGF in Age-related Choroidal Neovascularisation) trial (2012) showed equivalent efficacies between ranibizumab and bevacizumab. ³⁵

DME and the macular edema caused by RVO has been found to stem from leaky vasculature induced by VEGF. ⁴² Treatment of DME with anti-VEGF medication compared with sham was proven to be efficacious in the landmark studies RISE and RIDE (A Study of Ranibizumab Injection in Subjects With Clinically Significant Macular Edema With Center Involvement Secondary to Diabetes Mellitus) (2012). ⁴³ Further studies such as the DRCR.net Protocol I (2010) and RESTORE (Ranibizumab Monotherapy or Combined with Laser versus Laser Monotherapy for DME)/RESOLVE (Safety and Efficacy of Ranibizumab in DME) (2011) continued to prove the efficacy of ranibizumab compared with laser, ^44,45 whereas the VIVID (Intravitreal Aflibercept Injection in Vision Impairment Due to DME)/VISTA (Study of Intravitreal Aflibercept Injection in Patients With DME) (2015) studies found the superiority of aflibercept vs laser therapy. ⁴⁶ The DRCR.net protocol T (2015) found that aflibercept was superior to ranibizumab and bevacizumab in patients whose baseline Snellen VA from DME was 20/50 or worse; however, there was no significant difference between ranibizumab or bevacizumab when vision was 20/40 or better. ⁴⁷

Two landmark studies, the BRAVO (branch RVO) and CRUISE (central RVO) trials (2010), showed neutralization of VEGF with ranibizumab to be effective in improving VA in patients who have macular edema post RVO. ^48,49 Further studies, VIBRANT (Intravitreal Aflibercept for Macular Edema Following BRVO) (for branch RVO) (2015) and COPERNICUS (Intravitreal Aflibercept Injection for Macular Edema Due to CRVO)/GALILEO (VEGF Trap-Eye: Investigation of Efficacy and Safety in CRVO) (for central RVO) (2012/2013), showed the efficacy of aflibercept for macular edema in RVO. ^{50 -52} ROP, polypoidal choroidal vasculopathy, and CNV are other diseases treated with off-label use of VEGF inhibitors, and research evaluating their efficacy is among the top 100 list of cited articles. ^{53 -55}

Genentech had a senior author in 21 of the articles among the top 100. The involvement of technology companies in medical research raises ethical concerns, namely conflicts of interest. The outcomes of research that are portrayed can depend greatly on the support that it receives. The widespread use of bevacizumab and ranibizumab likely comes from Genentech’s promotion of its product through research and marketing, not only because of its efficacy. Industrial affiliation, on the other hand, undoubtedly spawns advancement and notably has been shown to increase the quality of research and the progress of medicine. ^56,57 Although conflicts of interest undoubtedly exist, large-scale and often complex clinical trials often rely on the monetary assistance of large pharmaceutical companies.

There are several limitations to this study. First, the search methods used (Web of Science’s “ophthalmology” category, Lancet, NEJM, and JAMA Ophthalmology), had the intention of selectively identifying articles based on relevance to ophthalmology. However, it is possible that other journals not included in our search could have produced articles on anti-VEGF medications and ophthalmology that would have been in the top 100 for citations. Second, despite using a datasheet produced by Web of Science for cross-reference, many of these data were manually extracted individually from all 100 articles, which presents opportunities for error. Although multiple cross-analyses and reviews resulted in the rechecking of these data, this does not completely rule out the effect of human error. Also, owing to inconsistencies among different journals’ bibliographic data, some data could represent erroneous assumptions. For example, certain articles display only 1 institution but multiple authors. In this instance we assumed that all authors were from the same institution.

This article is a bibliographic analysis of the top 100 most-cited articles involving anti-VEGF medications in ophthalmology (Supplemental Tables 1 and 2). We analyzed trends across the 100 most-cited papers with primary outcomes of obtaining the most frequently cited articles, which journals had articles with the most citations, as well as the type of drug and disease process studied. Although the NEJM had the most citations per publication (1714 citations), Ophthalmology published the greatest number of articles among the top 100 (48 articles). Ranibizumab (56 articles) and bevacizumab (48) were by far the most common medications studied. Finally, AMD was the most common disease process studied (42 articles). This study demonstrates the impact on ophthalmic research and large number of citations that anti-VEGF medication studies have had over the past 20 years.