Since the initial discovery that VEGF upregulation is a key driver of vascular leakage in an animal model of the diabetic retina [13], and subsequent confirmation of this in individuals with DME [14, 15], intravitreal anti-VEGF therapy has become the standard of care around the world [16,17,18,19,20]. This shift in the DME treatment paradigm was informed by several studies that demonstrated superior visual and anatomical improvements with anti-VEGF agents over existing laser therapies [21,22,23,24]. In August 2012, ranibizumab was the first anti-VEGF agent approved by the United States (US) Food and Drug Administration (FDA) for the treatment of DME, followed by aflibercept in July 2014, faricimab (dual VEGF-A/angiopoietin-2 inhibitor) in January 2022, brolucizumab in June 2022, and high-dose aflibercept in August 2023 [25]. In August 2022, ranibizumab-eqrn became the first FDA-approved anti-VEGF biosimilar for the treatment of DME [25]; several other biosimilars are approved or in development around the world [26], which will further expand the anti-VEGF treatment options available to patients and physicians.

Several real-world studies have found that the clinical benefits of anti-VEGF therapy for DME, as reported in landmark trials, have been difficult to replicate in clinical practice [6,7,8,9,10]. Table 1 summarizes the broad differences in results between RCTs and real-world data. In RCTs, mean best-corrected visual acuity (BCVA) gains of 6–13 Early Treatment Diabetic Retinopathy Study (ETDRS) letters were achieved with approximately 7–12 intravitreal anti-VEGF injections during the first year of treatment [21, 22, 27,28,29,30,31]. In contrast, the prospective real-world MERCURY study found that, on average, patients in Japan received 3–4 anti-VEGF injections during the first year of treatment, and consequently achieved inferior vision gains of approximately 4 ETDRS letters [9]. During the second year of treatment, patients in MERCURY received an additional 1–2 anti-VEGF injections on average, and achieved small additional vision gains of approximately 2.5 ETDRS letters [10].

Data from the US and Europe have shown that patients with DME are significantly more likely to cancel, or simply not attend, scheduled retina specialist appointments versus those with neovascular age-related macular degeneration (nAMD) [32]. Adherence to appointments may be particularly difficult for patients with DME, who are likely to be of working age [19], and who often have more medical appointments overall to address other complications of their diabetes than diabetic patients without DME [32]. Furthermore, a survey of Japanese retina specialists that sought to characterize real-world anti-VEGF treatment practices for DME found that the most common barriers to ongoing anti-VEGF therapy were the high costs of treatment (86% of respondents) and the need for frequent injections (24% of respondents) [33]. Similarly in the American Society of Retina Specialists (ASRS) 2023 Global Trends in Retina survey, specialists from around the world commonly identified “frequent loss to follow-up” and “limitations of patient access to retina care” as the greatest socioeconomic challenges faced when treating patients with DME [34].

The burden of frequent appointments and anti-VEGF injections may be addressed through newer agents with extended durability (e.g., brolucizumab, faricimab), which have shown that robust vision gains and anatomical improvements can be maintained with dosing up to every 12–16 weeks [30, 31]. In addition, the increasing availability of anti-VEGF biosimilars may alleviate the financial burden of treatment on patients and healthcare systems [26]; however, increased physician-patient communication and education may be required to overcome the potential reluctance of patients to accept treatment with a generic agent [35]. A growing number of anti-VEGF treatment options for DME may also confound clinical decision-making, thus highlighting the need to differentiate between individual anti-VEGF agents and identify patient populations most likely to benefit from each treatment.

Future research directions in anti-VEGF therapy for DME include extended-release intraocular devices and gene therapy. The Port Delivery System with ranibizumab (PDS) is a surgical ocular implant that provides continuous ranibizumab therapy into the eye [36]. In the phase 3 Archway trial of 418 patients with nAMD, the PDS (refilled every 24 weeks) had similar efficacy to monthly ranibizumab injections [36]. A phase 3 trial of the PDS in DME is currently ongoing (NCT04108156). Two gene therapy products (RGX-314 and 4D-150), that are administered as a one-time injection and allow endogenous expression of anti-VEGF, are currently undergoing phase 2 trials in patients with DME (NCT04567550 and NCT05930561), with estimated primary completion dates in 2024.

In addition to functional outcomes (i.e., visual acuity), clinical trials have consistently demonstrated the efficacy of anti-VEGF therapy using anatomical endpoints, including optical coherence tomography (OCT) measures of retinal thickness (e.g., central retinal thickness [CRT], central subfield thickness [CST]) and retinal fluid (e.g., subretinal fluid [SRF], intraretinal fluid [IRF]) [21, 22, 27,28,29,30,31, 37]. However, unlike vision gains, which are generally comparable across individual anti-VEGF therapies, comparative studies and meta-analyses have shown that anatomical responses to treatment can differ between agents [24, 29,30,31]. For example, a recent Cochrane review found no clinically important differences in 24-month BCVA gains between current anti-VEGF therapies for DME, but estimated that 24-month CRT reductions tended to favor brolucizumab and aflibercept over ranibizumab, bevacizumab, and ranibizumab plus prompt or deferred laser therapy [24].

Given that retinal thickness and fluid are important features of DME, international guidelines and clinical trials frequently use anatomical measures to monitor anti-VEGF treatment response and guide retreatment decisions. Current clinical guidelines recommend the use of OCT, in conjunction with fundus photography and fluorescein angiography, to diagnose DME based on morphological indicators, and advocate that patients can be monitored and assessed for anti-VEGF retreatment based on visual acuity and OCT findings [16,17,18, 20]. Similarly in clinical trials, OCT-based anatomical criteria are routinely used to select patients for study enrolment, and to determine anti-VEGF dosing intervals in personalized treatment regimens (e.g., pro re nata [PRN; as-needed], treat-and-extend) [21, 27, 29,30,31].

In line with clinical guidelines and trial protocols, OCT assessments of retinal morphology are widely used to guide the management of DME in current clinical practice [19, 33, 38]. Figure 1 illustrates typical OCT parameters assessed in everyday clinical practice; however, there is no universally agreed-upon approach for utilizing these parameters in a clinical context. A survey found that 68% of retina specialists agreed that OCT findings (particularly CRT and OCT retinal maps) are the most important considerations when initiating treatment for DME, followed by visual acuity (23% of respondents) and subjective symptoms (8% of respondents) [38]. In comparison, a similar cross-sectional study of ophthalmologists found that visual acuity was the most useful assessment to guide DME treatment initiation (44% of respondents) followed by OCT (31% of respondents) [33].

Representative examples of typical 3-D optical coherence tomography (OCT) parameters assessed in patients with diabetic macular edema (DME) in clinical practice. CRT central retinal thickness

CST, central subfield thickness; DRIL, disorganization of retinal inner layers; ELM, external limiting membrane.

Despite the routine use of anatomical measures in clinical trials and clinical practice, no consensus has been reached on how OCT findings should be interpreted and used to inform DME treatment decisions [39]. For example, there is no consistent CRT threshold to identify the presence or absence of DME, nor consistent definition of “stable” CRT to guide anti-VEGF retreatment.

Although clinical trials have shown that vision gains with anti-VEGF therapy are typically accompanied by reductions in retinal thickness, several studies have found that the correlation between these two endpoints is modest at best. For example, analyses of the Diabetic Retinopathy Clinical Research Network (DRCR.net) Protocol A and Protocol T trials aimed to characterize the relationships between retinal thickness and visual acuity, and between changes in these parameters after treatment, in patients with DME [40, 41]. Both studies estimated small-to-moderate correlations between retinal thickness and visual acuity before and after treatment, suggesting that measuring retinal thickness at a single time point, or as the change between two time points, may not be reliable surrogate markers for vision outcomes [40, 41].

However, recent studies exploring other measures of retinal fluid have progressively revealed associations between morphological indicators and visual acuity in DME. In a post hoc analysis of the DRCR.net Protocol I study, the duration and amount of residual edema after anti-VEGF treatment with ranibizumab were each significantly and negatively correlated with longer-term vision outcomes [42]. On average, smaller vision gains over 3 years of follow-up were observed among patients with more persistent residual fluid during the first year of treatment (based on the number of study visits with CRT ≥ 250 μm), and among those with higher levels of edema during the same period (based on the amount by which CRT exceeded 250 μm) [42]. Moreover, a post hoc analysis of DRCR.net Protocols T and V found that larger fluctuations in CST (i.e., less stable fluid control) during 1 year of anti-VEGF therapy, focal/grid laser treatment, or observation were associated with worse vision outcomes over 2 years [43]. When SRF and IRF volumes were quantified separately, another analysis of Protocol T data showed that for every 10-nL reduction of central IRF and SRF achieved with anti-VEGF treatment, BCVA was significantly improved by 0.15 and 0.34 ETDRS letters, respectively (both P < 0.001) [37]. In a subsequent retrospective cohort study that divided IRF into compartments of the inner nuclear layer (INL) and outer plexiform layer (OPL), fluid volume in the INL demonstrated a stronger correlation with visual acuity than fluid volume in the OPL, whole macular fluid volume, and CST [44]. Taken together, these data suggest that the persistence, stability, and location of retinal fluid may be stronger predictors of visual acuity than retinal thickness in patients with DME; and that treatments providing rapid, stable, and sustained fluid control may improve long-term vision outcomes.

Although retina specialists believe that high treatment costs and frequent injections are common barriers to ongoing anti-VEGF therapy in clinical practice [33], patient surveys suggest that vision is the most important consideration for those with DME, and that many patients would be willing to accept increased treatment burden in exchange for better vision outcomes [45, 46]. For example, in a US survey of patients receiving anti-VEGF therapy for DME or nAMD, achieving good vision was the most important factor in treatment decisions (40% relative importance), followed by low cost to the patient (23%), on-label drug status (21%), less frequent treatment intervals (12%), and low cost to the insurance provider (3%) [45]. Another survey from the US by Mason and colleagues further highlighted the importance of vision in patients with DME: 83% of respondents indicated that they would accept 15–16 intravitreal injections to gain 2 lines of Snellen visual acuity, 91% would sacrifice zero lines of vision in order to receive fewer treatments, and 76% were willing to attend 12 treatment visits per year in order to maintain their vision [46]. The willingness to accept increased treatment burden may have been influenced in these US surveys by patient health insurance status and level of affluence. In the survey by Mason and colleagues, who collected sociodemographic and healthcare insurance status data, laser treatment was preferred over injections in unemployed respondents compared with employed respondents [46]. The authors of the latter study concluded that patient demographics influenced their responses regarding their preferences for DME treatment [46].

Nevertheless, patients may be more willing to accept an increased treatment burden when they have a clear understanding of treatment requirements and expected outcomes; however, physician-patient communication around optimal diabetes management may be lacking in current clinical practice. In a global survey of physicians involved in the early treatment of patients with type 2 diabetes, most respondents (88%) agreed that conversations at the time of diagnosis can meaningfully impact a patient’s acceptance and self-management of their condition over time [47]. Despite this, almost all physicians (99%) reported communication challenges when discussing type 2 diabetes management with patients; these included their sense of disappointment with patient attitudes (e.g., difficulty adhering to treatment recommendations, failure to understand the seriousness of the condition), and their sense of frustration with the clinical setting (e.g., insufficient time or resources to support patients, difficulty responding to emotional responses from patients) [47]. In our view, the stigma of diabetes a patient may experience, which is well documented in adolescents and young adults with diabetes (especially in female patients and/or patients with elevated glycated hemoglobin levels or diabetic retinopathy) [48], may hinder patient-physician communication. It is possible that if the clinician discusses diabetes stigma with the patient as part of providing comprehensive diabetes care, overall patient-physician communication may be improved.

For patients with DME, strategies to improve physician-patient communication are needed to ensure that the long-term benefits of anti-VEGF therapy are realized. Early discussions should aim to educate patients that best-achievable responses to anti-VEGF therapy require frequent injections and close monitoring, and that DME is a chronic disease that requires lifelong management. Patients may also be encouraged by 5-year data from the DRCR.net Protocol I trial, which found that many patients were able to reach a state of “remission”, where they could maintain initial vision outcomes with very few additional anti-VEGF injections over time [49].

Patients should also be made aware that DME is a heterogeneous disease, and while some may be able to achieve vision gains with anti-VEGF therapy, a more realistic treatment goal for most patients is to delay disease progression and avoid further vision loss. This could be illustrated with OCT images showing that retinal thickness and fluid are improved with treatment, and an explanation