Intravitreal therapy—success stories and challenges

Today, injections with anti-VEGF for various indications such as nAMD, DME, RVO, myopic choroidal neovascularization, and proliferative diabetic retinopathy account for most IVI administered around the world. In the pre-anti-VEGF era, a then still-unknown factor had long been presumed to promote new vessel growth in hypoxic retinas. In 1994, Miller et al. identified this factor to be VEGF by inducing retinal ischemia through laser photocoagulation of retinal veins in monkeys, which eventually led to iris neovascularization. Vascular endothelial growth factor was then successfully isolated in the aqueous humor and levels changed proportionally to the severity of iris neovascularization. This suggested VEGF to be a retina-derived diffusible factor promoting angiogenesis and vascular permeability [12]. These findings were supported in a study by Aiello et al. during the same year, which found VEGF to be present in ocular fluid samples of patients with ischemic retinal diseases such as RVO and diabetic retinopathy [13].

It then took a decade to translate these basic scientific findings into clinically applicable medicine. The first anti-VEGF drug for the treatment of nAMD to be approved by the FDA in 2004 and the European Medicines Agency (EMA) in 2005 was pegaptanib (Macugen®, Eyetech Pharmaceuticals, Pfizer, New York, USA). This approval made pegaptanib the first anti-VEGF agent available for the treatment of ocular neovascularization [14]. Earlier the same year, bevacizumab (Avastin®, Genentech, Roche, California, USA), a humanized anti-VEGF antibody, received approval for treatment of colon cancer [15]. Soon after, ophthalmologists became interested in bevacizumab’s anti-VEGF properties and began successfully treating patients with nAMD by off-label administration of systemic and intravitreal bevacizumab [16, 17]. Bevacizumab remains the most widely used anti-VEGF agent around the world, despite its continuing off-label status.

After approval of the anti-VEGF agent ranibizumab (Lucentis®, Genentech, Roche/Novartis, California, USA) by the FDA in 2006 and the EMA in 2007, this drug was soon considered the gold standard for nAMD treatment due to its increased efficacy compared to pegaptanib [18, 19]. During the 2010s, indications for ranibizumab were broadened to RVO and DME [20‐23]. The efficacies of bevacizumab and ranibizumab treatments in nAMD were compared in a publicly funded randomized controlled trial and found to be equivalent, which further supported the continued use of bevacizumab in spite of the lack of formal approval (Fig. 1; [24]). During the same year, another anti-VEGF agent, aflibercept (Eylea®, Regeneron Pharmaceuticals, Bayer, New York, USA), was approved by the FDA and a year later by the EMA. With improved pharmacokinetics, i.e., a 100-fold increase in VEGF binding affinity, aflibercept 2 mg given every 2 months was found to be noninferior to ranibizumab every month [25]. Aflibercept also proven effective for the treatment of RVO and DME and allowed for higher intervals between injections to reduce patient burden [26, 27]. In 2024, aflibercept 8 mg (Eylea HD®, Regeneron Pharmaceuticals, Bayer, New York, USA) with a four-fold increase in dosage was approved by the FDA and the EMA for use in nAMD, DME, and RVO, with treatment intervals of up to 20 weeks [28, 29]. Brolucizumab (Beovu®, Novartis, Basel, Switzerland) was approved in 2019, but has lost popularity since then due to increased rates of intraocular inflammation (see below) [30, 31]. Faricimab (Vabysmo®, Genentech, Roche, California, USA), the first bispecific antibody targeting both VEGF and angiopoietin‑2, complements the approved anti-VEGF treatment options for nAMD, DME, and RVO, with approvals in 2022 and 2023 [32‐34].

Bild vergrößern

Ophthalmologists have been able to treat nAMD effectively with injections of anti-VEGF for almost two decades. However, for geographic atrophy (GA), the advanced disease stage of non-neovascular AMD characterized by subsequent atrophy of the outer retinal layers, there was no treatment available until recently. Progression of GA has been linked to the complement system, which is part of the innate immune response and comprises three different pathways. These pathways converge at factor C3 while also cascading through factor C5, which makes them promising targets for therapeutic intervention [35, 36]. The FDA approved pegcetacoplan (Syfovre®, Apellis, Massachusetts, USA), a C3 inhibitor, and avacincaptad pegol (Izervay®, Astellas Pharma, Illinois, USA), a C5 inhibitor, in 2023 for slowing progression of GA. This was based on the results of the four phase III trials, DERBY/OAKS [37] and GATHER1/2 [38, 39], respectively, which all showed slowing of GA progression but no effect on visual function.

Contrary to preceding approvals for nAMD, the FDA waived the requirement of improvement in functional endpoints such as visual acuity, scotoma size, or low luminance visual acuity and accepted slowing of GA expansion on fundus autofluorescence images as a primary anatomical endpoint for drug approval. These changed requirements as well as the only moderate effect on GA progression of pegcetacoplan and avacincaptad pegol (22% and 19% decreases of GA lesion growth with monthly injections over 24 months and 27% and 14% decreases with monthly injections over 12 months, respectively) raised questions concerning treatment burden, cost effectiveness, and patient benefit [40‐42]. In an initial review, the EMA refused approval of pegcetacoplan [43] on the grounds of a lack of benefit for visual function. The EMA also referred to the increased rate of new-onset nAMD for patients receiving pegcetacoplan monthly or every other month versus patients receiving sham injections, which was reported as 13% and 6% versus 4% in DERBY and 8% and 11% versus 2% in OAKS, respectively, at 24 months [37]. Therefore, it might be beneficial to primarily offer this treatment to fast-progressing patients in order to prevent overarching treatment burden in the face of an on-average moderate efficacy. However, this is the first time that a treatment has shown any success in slowing of GA progression in clinical trials, which provides hope to millions of affected people around the world.

In the early 2000s, the use of intravitreal steroids to treat retinal diseases such as DME, macular edema following RVO, uveitic macular edema, and pseudophakic cystoid macular edema (also known as Irvine–Gass syndrome) emerged. The rationale for using intravitreal steroids to treat retinal vascular diseases is the reduction of inflammation, which is often associated with abnormal cell proliferation, and the stabilization of the blood–retinal barrier to reduce vascular permeability [44]. The leading steroid to be administered intravitreally was triamcinolone and it provided excellent results, which at the time could already be monitored by optical coherence tomography (OCT), a noninvasive imaging technique that provides in vivo imaging of the retinal layers. Multiple trials for different indications that confirmed the effectiveness and safety of intravitreal triamcinolone were conducted in the following years and it accounted for most of the IVI administered in the early 2000s [45‐48].

To further decrease the burden caused by frequent IVI, two sustained-release implants of corticosteroids were developed. Ozurdex® (Allergan, AbbVie, California, USA), which is a bioerodible implant releasing dexamethasone over a 3–6 month period, was shown to be effective in the treatment of DME [49], macular edema secondary to RVO [50], and posterior uveitis [51]. Iluvien® (Alimera Sciences, Georgia, USA), a non-bioerodible implant which releases a continuous amount of fluocinolone over a period of 2–3 years, was shown to be effective in the treatment of DME [52] and posterior uveitis [53].

Intravitreal injections carry a small but significant risk of adverse events (AE). Complications range from harmless subconjunctival hemorrhages and ocular discomfort to sight-threatening retinal detachment and endophthalmitis. To prevent these complications, standard operating procedures are maintained as follows: All involved parties should wear surgical masks to prevent contamination of the needle tip and operating field with oral bacterial flora. After instillation of topical anesthesia, repeated ocular antisepsis with povidone-iodine (Fig. 2a) and periocular antisepsis, a lid speculum may be introduced as per the physician’s discretion or lid retraction can be achieved manually (Fig. 2b, d). A marker can then be used to identify the injection site 3.5 mm (pseudophakic) or 4 mm (phakic) behind the limbus (Fig. 2b) and povidone-iodine is instilled again. The 30G needle is then introduced perpendicularly and the agent is applied (Fig. 2c, d). Povidone-iodine is instilled again, and a lubricant is applied to reduce foreign-body sensation. Proper patient education concerning postinterventional behavior, i.e., no rubbing of the eye, further reduces the risk of complications [1].

Bild vergrößern

Intravitreal injections are mostly administered in vulnerable patient populations such as the elderly or patients with diabetes and vascular diseases. Therefore, possible systemic effects of IVI must be considered. It has been shown that intravitreally administered anti-VEGF agents enter the bloodstream and reduce systemic VEGF levels [78, 79]. This also serves as an explanation for the “fellow-eye effect,” in which contralateral eyes that did not receive an anti-VEGF injection also showed anatomical improvements on OCT [80]. However, other systemic effects remain unclear to date, and conflicting results have been published. Anti-VEGF injections have been linked to gastrointestinal perforation and bleeding, hypertension, kidney disease, myocardial infarction, stroke, and thromboembolic events. However, in an overview of 21 systematic reviews and meta-analyses in 2018, Thulliez et al. found that anti-VEGF treatment did not increase the risk of systemic adverse events (syAE) when compared to controls. This finding was independent of treatment intervals. Only ranibizumab may be associated with a slightly higher risk of non-ocular systemic hemorrhage, such as subdural hematoma and gastrointestinal bleeding, especially in elderly patients with nAMD [81].

There also seems to be no difference in the syAE incidence among different anti-VEGF agents, as a study comparing the onset of syAE between patients treated with bevacizumab, ranibizumab, and aflibercept found no differences in the incidence of myocardial infarction, acute cerebrovascular disease, major bleeding, or hospitalization rates in 87,844 patients over an 11-year period [82]. Furthermore, a recent meta-analysis conducted in 2024 found no statistically relevant differences between the incidence of systemic arterial and venous thrombotic events in patients receiving ranibizumab vs. sham and between patients that received different anti-VEGF agents (aflibercept vs. bevacizumab vs. ranibizumab) for any indication [83]. In a study comparing intravitreal dexamethasone and anti-VEGF agents, again, no differences in incidence of syAE could be found [84].

By contrast, in a study in 2024, Zafar et al. found anti-VEGF injections in patients with type 2 diabetes to be independently associated with an increased likelihood (odds ratio 1.8) of syAEs, with the 5‑year cumulative incidence of any syAE being 37.0% in the anti-VEGF group compared to 18.4% in the non-injection group [85]. These findings contradict the results of the previously mentioned reports and underline the importance of prospective clinical trials with an a priori design to elucidate systemic effects of intravitreal treatment.

It is estimated that over 7 million IVI per year are administered in the US, with the number projected to continue to increase further [3]. The cost of anti-VEGF agents, coupled with the need for repeated injections 6–12 times a year for a lifetime due to the indefinite nature of the disease, places a huge financial burden on healthcare systems. For example, from 2015 to 2019, the anti-VEGF agent aflibercept was the leading drug in overall Medicare Part B expenditure, which primarily accounts for outpatient/medical coverage. Costs for aflibercept, bevacizumab, and ranibizumab accounted for approximately 12% of all Medicare Part B drug expenditures, totaling over 4 billion USD in 2019 in comparison to 2.51 billion USD in 2014. This increase was largely driven by a trend toward more frequent use of aflibercept compared to other anti-VEGF agents. List prices for anti-VEGF agents range from 1850 USD and 1950 USD per dose for aflibercept and ranibizumab, respectively, to merely 60 USD on average per dose for bevacizumab [86].

Likewise, numbers of IVI keep increasing in Europe. Moorfield’s Eye Hospital London, one of the largest eye hospitals in Europe, reported a nearly 11-fold surge from 2009 to 2019, with roughly 45,000 IVI administered in 2019 and a forecast of 83,000 injections in 2029 (see Fig. 4). Neovascular AMD was the leading indication and aflibercept, with 87%, was the most commonly administered drug [87]. In Austria, 212,000 IVI were administered in 2022, which represents a 4.4-fold increase compared to 2012. Numbers are predicted to rise to 252,000–346,000 IVI in 2030 [88]. While some retina specialists are mostly free to choose between different anti-VEGF agents, sometimes step therapy is mandated by health insurance, and hospital providers in publicly funded healthcare systems may mandate using a specific anti-VEGF drug, i.e., bevacizumab, as a first-line treatment to decrease medication costs [89, 90].

Bild vergrößern

Aside from financial aspects, the treatment burden for patients cannot be overlooked. Repeated injections entail frequent visits to healthcare facilities, which can be physically taxing, especially for elderly and mobility-impaired patients, and reduce treatment adherence [91]. Other socioeconomic factors such as transportation costs, reduced quality of life, lost wages due to time off work, and caregiver burden remain. However, socioeconomic aspects also include benefits of anti-VEGF injections due to prevention of visual deterioration such as higher productivity, better quality of life, and decreases in long-term health costs such as home modifications and long-term care. Furthermore, effective treatment of retinal diseases helps patients to maintain their independence. This supports a healthier, more active aging population, which can contribute positively to the economy by reducing the demand for social and healthcare services.

The increasing workload of IVI also poses significant challenges for ophthalmologists and healthcare providers. The aging population and rising incidence of diabetes are driving up the prevalence of AMD and diabetic retinopathy. As new treatments for GA become available, the demand for IVI is expected to grow even further. Additionally, the anticipated retirement wave of ophthalmologists will exacerbate workforce shortage, placing increased pressure on the remaining professionals.

While newer drugs with longer intervals between injections have been approved recently, more research is needed to provide even more effective and longer-lasting treatment options to alleviate the burden on patients, caregivers, and healthcare providers.

Although IVI are highly effective, repeated injections place an enormous burden on patients, their relatives, and the healthcare system in general. To counter this issue, refillable implants with continuous drug release have long been desired. The most promising implant to date is the port delivery system (PDS) with ranibizumab called Susvimo® (Genentech, Roche, California, USA). It is a nondegradable and refillable implant, which is placed into the sclera at the pars plana by vitreoretinal surgery (Fig. 5). The concentration gradient between the port and the vitreous cavity results in passive diffusion of ranibizumab towards the vitreous. The pivotal Archway trial with 418 participants demonstrated noninferiority for patients implanted with a PDS (ranibizumab 100 mg/mL) and a refill exchange at week 24 compared to ranibizumab 0.5 mg every 4 weeks. Rescue therapy with ranibizumab 0.5 mg was possible but only needed by 1.6% of patients during the trial [92]. However, in October 2022, the PDS was recalled due to issues with septum dislodgement, which was recognized in 2.3% of patients after refilling the device in the Archway phase III trial and Portal extension study. In response to that, the device and refill needle have been remodeled and studies are currently ongoing [93].

Bild vergrößern

Gene therapy for retinal diseases offers a potentially long-lasting solution to conditions that currently require frequent and invasive interventions. To date, only an inherited retinal disease, namely Leber’s congenital amaurosis, can be treated effectively by gene augmentation with voretigene neparvovec (Luxturna®, Spark Therapeutics, Pennsylvania, USA) via adenovirus-associated virus vectors transporting the RPE65 transgene into target cells [94]. However, Luxturna® (Spark Therapeutics, Pennsylvania, USA) has to be administered through subretinal rather than intravitreal injection (Fig. 4), which requires a surgical approach. An intravitreal delivery of therapeutic agents via IVI would provide a less invasive way to place the transgene.

Research efforts in intravitreal anti-VEGF gene therapy focus on viral vectors that safely and efficiently deliver genes encoding anti-VEGF proteins directly to retinal cells. These vectors are engineered to carry genetic instructions that enable the continuous production of anti-VEGF agents, potentially reducing or eliminating the need for repeated IVI. However, there are certain limitations to intravitreal gene therapy that have to be overcome first, such as dilution of the vector, the distance the vector has to surpass, and the host’s immune response [95, 96].

One of the most promising candidates is ixoberogene soroparvovec (ixo-vec®, Adverum Biotechnologies, Delaware, USA), a single-dose gene therapy encoding for the anti-VEGF protein aflibercept that can be administered intravitreally. It showed favorable results in a phase I trial, improving anatomical outcomes in nAMD and reducing the need for anti-VEGF injections [97]. Another candidate for anti-VEGF gene therapy is RGX-314, potentially providing continuous VEGF suppression by expressing an anti-VEGF‑A antigen-binding fragment. It also demonstrated promising results in a phase I/IIa trial but has to be administered via subretinal injection [98]. Although still under development, intravitreal gene therapy holds significant potential to revolutionize the treatment of retinal diseases currently managed with anti-VEGF injections. The encouraging results from early clinical trials indicate a future where retinal diseases can be managed more effectively.