Prospective Real-World Outcomes After Switching to Aflibercept 8 mg in Neovascular Age-Related Macular Degeneration with High Treatment Burden

Abstract

Introduction

This study aimed to evaluate functional, anatomical, and safety outcomes after switching from aflibercept 2 mg to aflibercept 8 mg in patients with neovascular age-related macular degeneration (nAMD) requiring short re-treatment intervals in a real-world setting.

Methods

This single-center prospective observational study included patients with nAMD insufficiently responsive to prior anti-vascular endothelial growth factor (anti-VEGF) therapy requiring re-treatment every 4–6 weeks. All eyes had received four consecutive intravitreal injections of aflibercept 2 mg prior to switching to aflibercept 8 mg as part of routine clinical management. Functional outcome was assessed by distance-corrected visual acuity (DCVA). Anatomical outcomes included central retinal thickness (CRT), central subfield thickness (CST), and optical coherence tomography (OCT) features. Injection burden, responder status, discontinuation, and safety outcomes were evaluated. Follow-up occurred at 3, 6, 9, and 12 months.

Results

Fifty patients (50 eyes) were included (mean age 78.4 ± 6.9 years; 60% female). Median baseline DCVA was 0.19 logMAR (IQR 0.12–0.32; 75.5 ETDRS letters), median CST 272 µm, and median CRT 338 µm. DCVA improved significantly at months 3, 6, and 9 (p ≤ 0.046) and remained stable at month 12. CRT and CST decreased early and remained improved. Over 12 months, 24 patients (48%) completed follow-up, while 26 (52%) discontinued early. Intraocular inflammation (IOI) occurred in 10 eyes (20%), corresponding to 3.0% per injection, and was temporally associated with off-label aliquoted preparation; all cases resolved without permanent visual loss. No further IOI events were observed after transition to on-label preparation. During safety evaluation, seven patients discontinued aflibercept 8 mg as a result of precautionary safety measures. Seventeen (34%) required ≤ 8 injections per year.

Conclusions

In chronically treated patients with nAMD requiring short re-treatment intervals, switching to aflibercept 8 mg was associated with modest anatomical improvement and stable visual function. A reduction in treatment burden was observed in a subset of patients, although overall durability remained heterogeneous and discontinuation rates were high. A cluster of sterile inflammatory events occurred during off-label aliquoted preparation, highlighting the importance of appropriate drug handling. Careful patient selection and strict adherence to on-label preparation procedures are essential to optimize safety. These findings should be interpreted cautiously given the exploratory design, lack of a comparator group, and high discontinuation rate.

Trial Registration

ClinicalTrials.gov identifier NCT07390253.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40123-026-01394-1.

Key Summary Points

Why carry out this study?

Neovascular age-related macular degeneration (nAMD) often requires frequent intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections, and many chronically treated patients still require re-treatment intervals of 4–6 weeks, resulting in a high treatment burden.

Aflibercept 8 mg has been developed to provide more sustained VEGF suppression and potentially extend treatment intervals.

This prospective real-world study investigated outcomes after switching from aflibercept 2 mg to aflibercept 8 mg in patients with nAMD requiring frequent re-treatment.

What was learned from the study?

Switching to aflibercept 8 mg was associated with early and sustained anatomical improvements and modest visual acuity gains, with overall stable vision at 12 months.

Treatment durability remained heterogeneous, with 38% of eyes achieving treatment intervals of ≥ 8 weeks and only a minority reaching longer intervals.

A cluster of intraocular inflammation (IOI) events occurred during off-label aliquoted preparation, while no further events were observed after transition to on-label administration.

IOI incidence was 3.0% per injection. All events occurred during off-label aliquoted preparation (6.8%), highlighting the importance of proper drug handling.

Overall, these findings are exploratory and should be interpreted with caution in this high-burden, previously treated patient population.

Introduction

Age-related macular degeneration (AMD) is a leading cause of irreversible visual impairment in older adults and represents a major and growing global health burden [1–3]. Epidemiological data indicate a high prevalence of AMD in Europe, with pooled estimates of approximately 8%, and demographic projections suggest a substantial increase in affected individuals as a result of population aging [2, 3]. Among the clinical phenotypes of AMD, neovascular AMD (nAMD) accounts for most cases associated with severe vision loss and is characterized by pathological macular neovascularization, vascular leakage, and progressive disruption of retinal architecture [1, 3].

The pathophysiology of nAMD is primarily driven by dysregulated vascular endothelial growth factor (VEGF) signaling, resulting in the formation of fragile neovascular membranes and subsequent accumulation of intra- and subretinal fluid [1, 3, 4]. If inadequately controlled, these processes lead to photoreceptor damage, retinal scarring, and irreversible visual decline. The introduction of intravitreal anti-VEGF therapy has fundamentally changed the natural history of nAMD, enabling stabilization and, in many patients, improvement of visual acuity [5–7].

Currently available anti-VEGF agents, including bevacizumab, ranibizumab, and faricimab aflibercept, constitute the standard of care for nAMD and have demonstrated robust efficacy in randomized clinical trials as well as real-world clinical practice [5–7]. Nevertheless, nAMD remains a chronic disease that typically requires long-term, repeated intravitreal injections to maintain disease control and prevent recurrence of exudation [5, 8].

A major limitation of current treatment strategies is the substantial treatment burden associated with frequent injections. Many patients, particularly those with persistent or recurrent disease activity, require re-treatment intervals as short as 4–6 weeks [8–10]. Such intensive treatment schedules pose logistical challenges, reduce adherence, and increase pressure on healthcare systems [8]. Real-world studies consistently demonstrate lower injection frequencies and inferior visual outcomes compared with clinical trial settings, underscoring the importance of treatment durability and sustained VEGF suppression [9].

In response to these challenges, considerable efforts have focused on developing therapeutic approaches that maintain efficacy while extending treatment intervals. Two principal strategies have emerged: the modulation of additional angiogenic pathways and the escalation of the intraocular dose of VEGF inhibition [5]. Faricimab, a bispecific antibody targeting VEGFA and angiopoietin-2, has demonstrated extended dosing intervals of up to 16 weeks in treatment-naïve patients with nAMD while maintaining visual outcomes comparable to standard therapies [6]. Similarly, aflibercept 8 mg, representing a fourfold increase in dose compared with aflibercept 2 mg, has recently been approved for the treatment of nAMD [11].

The rationale for aflibercept 8 mg is supported by pharmacokinetic considerations, as higher intraocular drug concentrations may result in more sustained VEGF inhibition and improved control of residual disease activity [5, 12]. In pivotal randomized trials such as PULSAR, aflibercept 8 mg demonstrated non-inferior visual outcomes compared with aflibercept 2 mg while allowing extended dosing intervals of up to 12 or 16 weeks in treatment-naïve patients [11]. Similar principles have been demonstrated for aflibercept 8 mg in other retinal diseases, such as diabetic macular edema, further supporting the concept of increased durability with higher dosing [13].

However, patients treated in routine clinical practice frequently differ from those enrolled in clinical trials. Chronically treated patients who continue to require short re-treatment intervals despite regular anti-VEGF therapy represent a clinically challenging subgroup. These so-called suboptimal responders or non-responders often exhibit persistent or recurrent retinal fluid despite intensive treatment and account for a disproportionate share of treatment burden [5, 9, 12]. Proposed mechanisms underlying limited treatment durability include pharmacokinetic variability, lesion characteristics, tachyphylaxis, and interindividual differences in drug clearance or disease biology [5, 12].

Evidence guiding treatment decisions in this population remains limited. Most real-world studies of aflibercept 8 mg have focused on treatment-naïve patients and relatively short follow-up periods, limiting their applicability to chronically treated high-burden populations [14–18]. Retrospective analyses and small real-world studies suggest that higher-dose aflibercept may improve anatomical disease control and treatment durability in suboptimal responders to standard-dose therapy, although visual gains may be modest and heterogeneous [12, 19–21]. Importantly, most randomized trials of aflibercept 8 mg have focused on treatment-naïve populations, leaving a relevant knowledge gap regarding the effectiveness of high-dose therapy in chronically treated patients with high treatment demand [11, 12].

In addition, real-world safety data for aflibercept 8 mg are still emerging. While pivotal trials reported a safety profile comparable to standard-dose aflibercept, post-marketing observations highlight the importance of evaluating safety and tolerability in broader, more heterogeneous patient populations [11–13]. Recent retrospective real-world studies have reported outcomes following a switch to aflibercept 8 mg in previously treated patients with nAMD and diabetic retinopathy, demonstrating improvements in anatomical parameters and reduced treatment burden; prospective data remains limited [20, 22].

Taken together, there remains an unmet need for prospective real-world data examining the effects of switching chronically treated patients with nAMD from standard-dose aflibercept to aflibercept 8 mg anti-VEGF therapy. It is unclear whether such a switch can meaningfully prolong treatment intervals while maintaining functional and anatomical stability in patients requiring re-treatment every 4–6 weeks under routine clinical conditions.

The present prospective, observational real-world study was therefore designed to evaluate functional, anatomical, and treatment-related outcomes following a direct switch from aflibercept 2 mg to aflibercept 8 mg in patients with nAMD requiring short re-treatment intervals. By focusing on a treat-and-extend regimen without an additional loading phase, this study aims to provide clinically relevant evidence on treatment durability, injection burden, and disease control in a patient population that is insufficiently represented in randomized clinical trials.

Methods

Study Design

This prospective, single-center, real-world observational study assessed functional, anatomical, and treatment-related outcomes in patients with nAMD who showed a suboptimal response or non-response to prior anti-VEGF therapy and were switched to aflibercept 8 mg.

Given the exploratory nature of this pilot study and limited prior data in this specific clinical setting on this research question, no formal sample size calculation was performed; a pragmatic sample size of approximately 50 patients was deemed adequate for this exploratory real-world analysis.

The study was approved by the Ethics Committee of the City of Vienna (EK 24-105-0824, NCT07390253) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. No identifying participant information is included in this manuscript.

Patients and Eligibility

Patients with a confirmed diagnosis of nAMD who were scheduled to switch to aflibercept 8 mg were prospectively enrolled. All included patients had a history of extensive prior intravitreal anti-VEGF treatment and were considered insufficient responders (suboptimal or non-responders), defined by persistent or recurrent intraretinal and/or subretinal fluid on OCT despite regular treatment and/or inability to extend treatment intervals beyond 4–6 weeks without recurrence of disease activity.

Although patients had received different anti-VEGF agents during their disease, the treatment sequence immediately preceding the switch was standardized. Specifically, all patients had to have received a minimum of four consecutive intravitreal injections of aflibercept 2 mg directly prior to initiation of aflibercept 8 mg for inclusion into the study, ensuring a homogeneous treatment exposure before the switch. These injections were not administered as part of a study-specific loading phase but reflected routine clinical management in patients requiring frequent re-treatment intervals of 4–6 weeks because of persistent disease activity.

Patients were excluded if aflibercept 8 mg was discontinued as a result of safety-related exclusion criteria before outcome assessment or if baseline clinical data were unavailable. Additional exclusion criteria included extensive macular fibrosis, geographic atrophy involving the fovea, vitreous hemorrhage, active intraocular inflammation (IOI), other significant retinal pathologies (e.g., retinal vein occlusion), diabetic retinopathy, dense media opacities, hypersensitivity to the study drug, prior aflibercept 8 mg treatment, and pregnancy. Only one study eye per patient was included in the analysis; in patients with bilateral disease, the eye that was switched first by the treating physician in routine clinical practice was selected.

Treatment Regimen

Before switching, patients had been treated with intravitreal anti-VEGF agents, including aflibercept 2 mg, bevacizumab, and/or faricimab, according to individualized treatment regimens in routine clinical practice. The decision to switch to aflibercept 8 mg was made by the treating physician based on insufficient response or high treatment burden under previous therapy.

After the switch, aflibercept 8 mg was administered intravitreally according to a treat-and-extend approach: The minimum treatment interval was 4 weeks and the maximum interval was 16 weeks, in accordance with the approved label. Interval adjustments were performed in steps of 2 weeks. Interval extension was only allowed in eyes showing complete absence of both intraretinal and subretinal fluid on OCT. In the presence of recurrent disease activity, including intraretinal or subretinal fluid, the treatment interval was shortened by one step (2 weeks). Trace subretinal fluid was not tolerated for interval extension. Pigment epithelial detachment alone did not preclude interval extension in the absence of intraretinal or subretinal fluid.

Injection dates were prospectively recorded, and injection burden was normalized to injections per year based on the individual follow-up duration for each patient to account for variable observation periods, including patients with early discontinuation. For this purpose, the total number of injections received during follow-up was divided by the individual observation time (in months) and extrapolated to a 12-month period.

The medications were previously prepared in the hospital pharmacy under sterile conditions in a certified cleanroom. Preparation was performed by fully trained pharmacy personnel in accordance with institutional protocols and applicable regulatory standards for sterile compounding. The pharmacists had extensive experience in aliquoting medications for intravitreal injections. This preparation approach represents off-label handling of aflibercept 8 mg. There were no changes in medication batches, syringes, or any other materials compared with the preceding months. The same experienced pharmacy personnel were involved throughout the entire period. Later in the study, aflibercept was administered in accordance with the approved label, either drawn directly from a vial in the operating room or using a prefilled syringe.

Data Collection and Outcome Measures

Demographic characteristics, disease duration, prior treatment history, and clinical outcomes were prospectively collected during routine clinical visits and documented in the electronic medical record. Functional outcomes included distance-corrected visual acuity (DCVA), measured under standardized distance refraction conditions using the Precision Vision chart, recorded in logarithm of the minimum angle of resolution (logMAR), and converted to Early Treatment Diabetic Retinopathy Study (ETDRS) letter scores for analysis. Anatomical outcomes comprised central retinal thickness (CRT) and central subfield thickness (CST), measured in micrometers (µm). Central subfield thickness (CST) was defined as the mean retinal thickness within the central 1-mm ETDRS subfield, whereas central retinal thickness (CRT) refers to the central point thickness at the foveal center as provided by the OCT device.

Assessments were performed at baseline and at prospectively defined follow-up visits according to the individualized treatment interval after switching to aflibercept 8 mg. The primary outcomes were changes in DCVA and CST over time. Secondary outcomes included changes in CRT, qualitative optical coherence tomography (OCT) features, injection frequency, treatment interval distribution, responder status (≤ 8 vs. > 8 injections/year), and study discontinuation. For each follow-up time point, the measurement closest to the nominal visit time (3, 6, 9, and 12 months after baseline) was selected for analysis. Analyses were performed on an available-data basis.

Optical Coherence Tomography Analysis

Spectral-domain optical coherence tomography (SD-OCT) images were acquired using the Heidelberg Spectralis system (Heidelberg Engineering, Heidelberg, Germany). Images were evaluated for the presence of subretinal fluid (SRF), intraretinal fluid (IRF), pigment epithelial detachment (PED), retinal scarring, and integrity of the ellipsoid zone (IS/OS junction), both overall and within the central 1-mm subfield. OCT image quality was graded on a four-point scale (1 = good, 2 = noisy, 3 = insufficient, 4 = not gradable) and only scans of sufficient quality were included in the qualitative analysis.

Safety Assessment and Study Disposition

Adverse events (AE) and serious adverse events (SAE) were prospectively documented throughout follow-up. Study discontinuation was categorized as due to adverse events, serious adverse events, safety-related exclusion, or non-safety-related discontinuation, such as administrative reasons or planned treatment changes. Safety-related exclusion refers to precautionary discontinuation of treatment in patients without an individual adverse event whose scheduled injections fell within the time window of the inflammatory event cluster under investigation. Patient disposition and reasons for discontinuation were summarized descriptively.

Statistical Analysis

Continuous variables are presented as median and interquartile range (IQR) or mean ± standard deviation (SD), depending on data distribution, while categorical variables are presented as counts and percentages. Normality was assessed using the Shapiro-Wilk test. Comparisons between baseline and follow-up visits were performed using the Wilcoxon signed-rank test for paired continuous variables. Comparisons between independent groups were conducted using the Mann-Whitney U test for continuous variables and Fisher’s exact test for categorical variables.

A predefined sensitivity analysis was performed in patients who completed 12 months of follow-up to assess the robustness of the primary outcomes. Owing to safety-related discontinuations and informative censoring, analyses were based on available data at each visit without imputation. In addition, injection burden was analyzed in the subgroup of 12-month completers by evaluating the absolute number of aflibercept 8 mg injections administered between baseline and the 12-month visit. Given the exploratory nature of the analyses, no adjustment for multiple testing was applied, and p values should be interpreted descriptively; p < 0.05 was considered statistically significant. The statistical approach relied on repeated pairwise comparisons at predefined time points and did not include longitudinal modeling of repeated measures. Given the substantial dropout and the potential for informative censoring, this approach may be susceptible to bias. In addition, assumptions required for longitudinal models, such as missing at random, may not be fulfilled in this setting.

Statistical analyses were performed using IBM SPSS Statistics (version 29.0; IBM Corp., Armonk, NY, USA). Graphs and visualizations were generated using GraphPad Prism (version 10.0), Microsoft Excel (version 16.77), and Matplotlib (Python 3.11), and finalized for figure preparation using Microsoft PowerPoint (version 16.77).

Results

A total of 50 patients (50 eyes) with nAMD were prospectively included. Baseline demographic and clinical characteristics are summarized in Table 1. The mean age was 78.4 ± 6.9 years, and 30 patients (60%) were female. All eyes had undergone extensive prior intravitreal anti-VEGF treatment and were classified as insufficient responders, requiring short treatment intervals of 4–6 weeks before switching therapy. Although various anti-VEGF agents had been used during the disease course, all eyes had received at least four consecutive intravitreal injections of aflibercept 2 mg immediately prior to switching to aflibercept 8 mg. Overall, patients exhibited a very high prior treatment burden. On average, eyes had received 30.6 ± 19.6 prior anti-VEGF injections, including the four consecutive aflibercept 2 mg injections before the switch. The median number of prior injections was 25 (IQR 16–43), corresponding to a treatment frequency of 11.1 injections per year over a median disease duration of 2.76 years (IQR 1.44–4.97). Before switching to aflibercept 8 mg, injection burden was high: after annualization to account for variable observation windows 91.7% required > 8 injections per year, whereas only 8.3% required ≤ 8 injections per year (Fig. 3a).

Table 1.

Baseline characteristics

Variable	Value
Eyes/patients, n	50/50
Age, years (mean ± SD)	78.4 ± 6.9
Female, n (%)	30 (60%)
Male, n (%)	20 (40%)
Right eye, n (%)	27 (54%)
Left eye, n (%)	23 (46%)
Disease duration, median (IQR), years	2.76 (1.44–4.97)
Disease duration, mean ± SD, years	3.46 ± 2.46
Previously treated eyes, n (%)	50 (100%)
Prior aflibercept, n (%)	50 (100%)
Prior bevacizumab, n (%)	37 (74%)
Prior faricimab, n (%)	2 (4%)
≥ 2 prior switches, n (%)	2 (4%)
Baseline DCVA logMAR, median (IQR)	0.19 (0.12–0.32)
Baseline CST (µm), median (IQR)	272 (248–293)
Baseline CRT (µm), median (IQR)	338 (308–378)
SRF present, n (%)	30 (60%)
IRF present, n (%)	14 (28%)

DCVA distance-corrected visual acuity, logMAR logarithm of the minimum angle of resolution, IQR interquartile range, CST central subfield thickness, CRT central retinal thickness, SRF subretinal fluid, IRF intraretinal fluid, SD standard deviation

Fig. 3.

Injection frequency a before and b after switch to aflibercept 8 mg: proportion of eyes requiring ≤ 8 injections per year versus > 8 injections per year after switching to aflibercept 8 mg

At baseline, median DCVA was 0.19 logMAR (IQR 0.12–0.32; n = 50), corresponding to 75.5 ETDRS letters (IQR 69–79). Median CST was 272 µm (IQR 248–293), and median CRT measured 338 µm (IQR 308–378). SRF was present in 60% (30/50) of eyes, while 28% (14/50) exhibited IRF at baseline (Table 1). OCT image quality was predominantly sufficient throughout follow-up. At baseline, 80% of scans were graded as good or noisy (grades 1–2).

Functional outcomes over time are summarized in Table 2 and shown graphically in Fig. 1a. Median DCVA improved significantly from baseline to 0.16 logMAR (IQR 0.10–0.26; n = 49) at month 3 (p = 0.012) and further to 0.14 logMAR (IQR 0.08–0.27; n = 44) at month 6 (p = 0.046). This improvement was maintained at month 9, with a median DCVA of 0.14 logMAR (IQR 0.065–0.29; n = 30; p = 0.004). At month 12, median DCVA was 0.16 logMAR (IQR 0.10–0.32; n = 25) and no longer differed significantly from baseline (p = 0.132) (Table 2). On the basis of the last available visit, 24% of eyes gained ≥ 5 ETDRS letters and 12% gained ≥ 10 letters, while 12% lost ≥ 5 letters and 2% lost ≥ 10 letters. The median change in ETDRS letters from baseline to the last available visit was + 1 letter (95% CI 0 to + 3).

Table 2.

Functional and anatomical outcomes over time

Outcome	Baseline	Month 3	Month 6	Month 9	Month 12
DCVA logMAR	0.19 (0.12–0.32)	0.16 (0.10–0.26)	0.14 (0.08–0.27)	0.14 (0.065–0.29)	0.16 (0.10–0.32)
	n = 50	n = 49, p = 0.012	n = 44, p = 0.046	n = 30, p = 0.004	n = 25, p = 0.132
DCVA ETDRS	75.5 (69–79)	77 (72–80)	78 (71.5–81)	78 (70.5–81.8)	77 (69–80)
	n = 50	n = 49, p = 0.023	n = 44, p = 0.041	n = 30, p = 0.004	n = 25, p = 0.136
CRT (µm)	338 (308–378)	314 (291–336)	321 (288–356)	317 (282–359)	307 (281–348)
	n = 50	n = 48, p < 0.001	n = 44, p = 0.028	n = 30, p = 0.003	n = 25, p = 0.005
CST (µm)	272 (248–293)	252 (228–278)	266 (231–288)	256 (232–289)	245 (230–271)
	n = 50	n = 48, p < 0.001	n = 44, p = 0.109	n = 30, p = 0.016	n = 25, p = 0.054

Data are reported as median (IQR)

DCVA distance-corrected visual acuity, ETDRS Early Treatment Diabetic Retinopathy Study, CRT central retinal thickness, CST central subfield thickness

Fig. 1.

a DCVA over time after switch to aflibercept 8 mg: median DCVA (logMAR) at baseline and during follow-up after switching to aflibercept 8 mg. Shaded areas indicate the IQR. b Central retinal thickness (CRT) over time: median CRT in µm at baseline and during follow-up after switch to aflibercept 8 mg. Shaded areas represent the IQR. c Central subfield thickness (CST) over time: median CST in µm at baseline and follow-up visits after switch to aflibercept 8 mg. Shaded areas indicate the IQR

Anatomical outcomes demonstrated early and sustained changes and are summarized in Table 2 and Fig. 1b, c. Median CRT decreased from 338 µm (IQR 308–378, n = 50) at baseline to 314 µm (IQR 291–336; n = 48) at month 3 (p < 0.001). Significant reductions were maintained at month 6 (321 µm; p = 0.028), month 9 (317 µm; p = 0.003), and month 12 (307 µm; p = 0.005). Median CST declined from 272 µm (IQR 248–293) at baseline to 252 µm (IQR 228–278; n = 48) at month 3 (p < 0.001). Although CST remained numerically reduced thereafter, the reduction reached statistical significance at month 9 (256 µm; p = 0.016), but not at month 12 (245 µm; p = 0.054).

Qualitative OCT findings are summarized in Table 1 and Fig. 2. The proportion of eyes with SRF decreased from 60% (30/50) at baseline to 36% (15/43) at month 6 and 24% (6/25) at month 12. Similarly, IRF declined from 28% (14/50) at baseline to 18% (8/42) at month 6 and 8% (2/24) at month 12. No consistent changes were observed for pigment epithelial detachment, retinal scarring, or ellipsoid zone integrity. At baseline, pigment epithelial detachment was present in 90% (45/50) of eyes, retinal scarring in 26% (13/50), while an absolute intact ellipsoid zone was rare both overall (2%, 1/50) and within the central 1-mm subfield (16%, 8/50).

Fig. 2.

Subretinal and intraretinal fluid over time. Proportion of eyes with subretinal fluid (SRF) and intraretinal fluid (IRF) at baseline, month 6, and month 12 after switching to aflibercept 8 mg. Percentages are based on available eyes at each visit (baseline n = 50; month 6 n = 43/42; month 12 n = 25/24 for SRF/IRF, respectively)

Injection burden varied considerably between patients. When annualized to injections per year based on individual follow-up duration, 17 eyes (34%) required ≤ 8 injections per year, whereas 33 eyes (66%) required > 8 injections per year (Fig. 3b). Among the 12-month completers (n = 24), 8 eyes (33%) required ≤ 8 aflibercept 8 mg injections over 12 months, whereas 16 eyes (67%) required more than 8 injections. Treatment interval distribution was therefore analyzed as an additional marker of treatment durability. Baseline demographic, functional, and anatomical characteristics did not differ significantly between responder and non-responder groups. The distribution of last available treatment intervals after switching to aflibercept 8 mg demonstrated substantial heterogeneity, with most eyes remaining within shorter intervals and only a small proportion achieving extended dosing intervals (Fig. 4). On the basis of the distribution of last available treatment intervals, 38% of eyes achieved intervals of ≥ 8 weeks, 14% reached ≥ 10 weeks, 8% ≥ 12 weeks, and 2% ≥ 16 weeks.

Fig. 4.

Distribution of last available treatment intervals after switching to aflibercept 8 mg. Intervals were categorized as < 8, 8–11, 12–15, and ≥ 16 weeks. The last available interval was defined as the time between the two most recent injections per eye. Data are shown as percentages

Over the 12-month observation period, only 24 patients (48%) completed follow-up, whereas 26 patients (52%) discontinued early. Discontinuations were predominantly safety-related and comprised 13 adverse events (26%), 4 serious adverse events (8%), and 7 precautionary safety-related exclusions (14%). Only 2 patients (4%) discontinued for non-safety-related reasons. Baseline demographic and clinical characteristics, including age, sex, and anatomical parameters, were comparable between patients who completed 12 months of follow-up and those who discontinued early (Supplementary Table S1). Patient disposition and reasons for discontinuation are illustrated in Fig. 5 and detailed in Supplementary Table S2. IOI occurred in 10 eyes (20%) within a narrowly defined time window of 7 days (January 13–20, 2025), typically within 1–4 days after injection. All cases resolved without permanent visual loss.

Fig. 5.

Study discontinuation and patient disposition: patient disposition during the 12-month follow-up period. Discontinuations were categorized as adverse events (AE), serious adverse events (SAE), precautionary safety-related exclusions, or discontinuation for non-safety reasons

During the safety evaluation following the outbreak, seven patients were excluded from treatment with aflibercept 8 mg since they needed treatment and were switched to another medication. Following the analysis of the situation, the preparation process was modified, and intravitreal aflibercept 8 mg was drawn up in the operating room directly from the vial and administered, and later manufacturer-prepared prefilled syringes were used in all patients. Under these modified preparation and administration protocols, no further inflammatory reactions were observed. A total of 336 intravitreal aflibercept 8 mg injections were administered during the study period. IOI occurred in 10 eyes, corresponding to an overall incidence of 3.0% per injection. All IOI events occurred during the phase in which aflibercept 8 mg was prepared by aliquoting (10/147 injections; 6.8%, 95% CI 3.3–12.2), whereas no further events were observed after transition to on-label preparation using single-use vials or prefilled syringes (0/189 injections; 0.0%, 95% CI 0.0–1.9%).

Of the 10 IOI cases, 8 were classified as mild noninfectious inflammation and were managed with topical corticosteroids. Two cases were severe and required pars plana vitrectomy. These severe cases were initially treated as suspected infectious endophthalmitis; however, microbiological cultures were negative and infection was subsequently ruled out.

In addition, seven patients were excluded from further aflibercept 8 mg treatment for precautionary safety reasons during the safety evaluation period. Detailed characteristics are provided in Table 3 and Supplementary Table S4.

Table 3.

Characteristics of intraocular inflammation events

Case	Sex/age (years)	PID	Injection date	Days to IOI	AC inflammation	Vitreous haze	Pain	Hyperemia	Treatment	Vitrectomy	Culture	VA pre (logMAR)	VA at onset	VA final
1	M/83	36	2025-01-13	2	Cells++, flare+	++	No	No	Local steroids + AB	No	NA	0.1	0.32	0.2
2	M/70	44	2025-01-13	1	Cells++, flare+	++	No	No	Local steroids + AB	No	NA	0.4	HM	0.4
3	F/85	39	2025-01-13	4	Cells+	+++	Mild	No	Local steroids + AB	No	NA	0.2	FC	0.1
4	F/77	35	2025-01-13	2	Cells+	+++	No	No	Local steroids	No	NA	0.0	0.3	0.0
5	F/89	38	2025-01-20	2	Cells++, flare++	+++	Yes	Yes	Steroids + IV AB	Yes	Negative	0.1	0.32	0.4
6	F/75	21	2025-01-20	2	Cells++, flare+	++	Yes	No	Steroids + IV AB	Yes	Negative	0.2	0.33	0.0
7	F/82	10	2025-01-20	4	Cells++, flare+	±	No	No	Steroids	No	NA	1.0	1.0	0.0
8	F/83	37	2025-01-20	1	Cells++, flare+	++	No	No	Steroids	No	NA	0.1	FC	0.7
9	F/82	5	2025-01-20	2	Cells+	+	No	No	Steroids	No	NA	0.1	0.67	0.2
10	F/85	24	2025-01-20	2	Cells+++	++	No	No	Steroids	No	NA	0.2	NA	0.3

IOI intraocular inflammation, AC anterior chamber, AB antibiotics, IV intravenous, VA visual acuity, logMAR logarithm of the minimum angle of resolution, HM hand movements, FC finger counting, NA not available

A predefined sensitivity analysis restricted to 12-month completers (n = 24) confirmed the robustness of the primary outcomes. Among completers, significant improvements from baseline to month 12 were observed for DCVA, ETDRS letter scores, CST, and CRT, consistent with the main analysis (Supplementary Table S3).

No unexpected safety signals were observed beyond the events leading to documented study discontinuation.

Discussion

In this prospective real-world switching study, we investigated functional, anatomical, and treatment-related outcomes following a direct switch from aflibercept 2 mg to aflibercept 8 mg in patients with nAMD who required short re-treatment intervals of 4–6 weeks despite ongoing anti-VEGF therapy. This cohort represents a clinically challenging population characterized by long disease duration, extensive prior treatment exposure, and high treatment burden. Given the exploratory nature of the study, the findings should be interpreted as descriptive and hypothesis-generating rather than confirmatory.

Overall, switching to aflibercept 8 mg was associated with early and sustained anatomical improvements, reflected by reductions in CST and CRT parameters and a decreased prevalence of IRF and SRF. Functional outcomes showed modest but statistically significant improvements during early and intermediate follow-up, with visual acuity remaining stable at 12 months. These findings are consistent with the pharmacological rationale of higher intraocular VEGF suppression achieved through dose escalation and support the concept that improved anatomical disease control may be achievable even in chronically treated eyes with limited durability under standard-dose therapy [5, 12]. However, these findings should be interpreted cautiously given the lack of a comparator group. In addition, prior disease stabilization under frequent aflibercept 2 mg treatment may represent a potential confounding factor, as intensive treatment alone can improve disease control and facilitate subsequent interval extension under a treat-and-extend approach.

Our results complement data from pivotal randomized clinical trials such as PULSAR, which demonstrated non-inferior visual outcomes and improved durability with aflibercept 8 mg compared with aflibercept 2 mg in treatment-naïve patients [11]. However, important differences must be acknowledged. PULSAR enrolled treatment-naïve patients treated after a loading phase under fixed dosing regimens, demonstrating noninferior visual outcomes with extended treatment intervals. In contrast, our cohort comprised chronically treated, refractory eyes managed under a treat-and-extend approach without a new loading phase; therefore, direct comparisons of efficacy and durability between these settings are limited. Unlike trial populations, our cohort consisted exclusively of previously treated patients with persistent disease activity and short re-treatment intervals. In such patients, the potential benefit of high-dose therapy may be reflected less in large visual acuity gains and more in anatomical stabilization and treatment burden-related outcomes, a pattern that has also been observed in other real-world and switch studies [9, 12, 21]. Most currently available real-world studies of aflibercept 8 mg have focused on treatment-naïve patients or short-term follow-up, which limits their comparability to chronically treated high-burden populations [14–18]. As a single-arm observational study without a comparator group, all outcomes represent within-group changes and do not allow causal attribution to the treatment switch.

When compared with recent real-world switch studies, the stabilization of visual acuity and improvement in anatomical parameters observed in our cohort are broadly consistent with previously reported findings [12, 20, 21]. However, the achieved treatment durability appears somewhat less favorable, as only a subset of patients reached extended treatment intervals. This may in part be explained by the strict extension criteria applied in this study. In addition, the discontinuation rate in our cohort was higher than in most published switch studies and was largely driven by the cluster of IOI events, rather than reflecting limited treatment efficacy alone.

Injection burden remained heterogeneous across patients, with approximately one-third meeting the responder criterion of requiring ≤ 8 injections per year, while the majority continued to require more frequent re-treatment. This heterogeneity highlights the multifactorial nature of suboptimal response in nAMD, which has been attributed to factors such as pharmacokinetic variability, lesion characteristics, tachyphylaxis, and differences in disease biology [5]. Previous studies have similarly reported that injection frequency correlates with visual outcomes in real-world settings, emphasizing that insufficient durability remains a key limitation of current anti-VEGF therapy in a subset of patients [9]. However, this may in part be explained by the strict extension criteria applied in this study, as interval prolongation was only permitted in the complete absence of both intraretinal and subretinal fluid. Such a conservative treat-and-extend approach may have limited the achievable treatment intervals and should therefore be considered a protocol-driven factor rather than solely a limitation of the drug effect. This is further reflected in the distribution of last available treatment intervals, which illustrates that only a small proportion of patients achieved very extended dosing intervals (Fig. 4). These findings highlight the importance of treatment interval distribution as a clinically relevant measure of durability in switching studies.

Beyond treatment efficacy and durability, safety outcomes represent a critical aspect of real-world implementation of aflibercept 8 mg. A key secondary finding of this study was the occurrence of a cluster of IOI events, which contributed to a high discontinuation rate over 12 months, with more than half of patients discontinuing treatment early, predominantly for safety-related reasons. These events were temporally associated with off-label pharmacy aliquoting of aflibercept 8 mg and therefore represent an important safety observation in this real-world setting. Taken together, the high discontinuation rate and the safety signal related to drug preparation need to be carefully considered when interpreting both efficacy outcomes and estimates of treatment burden, as these factors may have influenced the observed results.

This observation introduces a potential confounding-by-technique effect, as it remains difficult to disentangle whether the inflammatory response was related to the drug formulation itself or to preparation and handling procedures. Although aflibercept 8 mg represents a higher concentration formulation compared with aflibercept 2 mg, there is currently no definitive evidence that intrinsic physicochemical differences alone would account for an increased inflammatory risk, and preparation and handling procedures likely represent an important contributing factor.

Although the study was designed as a switching study, this safety signal represents a clinically important observation that may have immediate implications for real-world practice.

The observed rate appears higher than those typically reported in randomized clinical trials; however, extensive analyses did not identify a specific underlying cause. The medication was prepared under sterile conditions in a certified hospital cleanroom, with unchanged storage conditions and identical medication batches, syringes, and materials compared with the preceding months. Notably, the same highly experienced pharmacy personnel were involved throughout the entire period, and no deviations in preparation, handling, storage, or batch-related factors could be identified. In addition, microbiological cultures of the aflibercept 8 mg batches used during the aliquoting phase were performed by the institutional hygiene team and yielded no microbial growth.

As a precautionary measure during the safety evaluation, seven patients were excluded from treatment. Subsequently, the preparation protocol was modified, with aflibercept 8 mg drawn directly from the vial and administered using single-use syringes. Under this modified preparation and administration approach, using single-use vials and later manufacturer-prepared prefilled syringes, no further inflammatory reactions were observed. In addition, higher event rates may be conceivable in real-world cohorts of elderly, multimorbid patients with neovascular age-related macular degeneration and long treatment histories, which differ substantially from populations typically enrolled in randomized clinical trials [8].

Importantly, the high dropout rate limits the interpretability of treatment-burden endpoints over a fixed 12-month period. The high attrition rate represents a major limitation and introduces a risk of attrition bias, as patients discontinuing early may differ systematically from those completing follow-up. Although normalization to injections per year partially accounts for variable follow-up duration, informative censoring cannot be fully excluded, as discontinuation was frequently related to adverse events or safety considerations. Long-term outcomes may preferentially reflect patients with more favorable responses or better tolerability. Baseline characteristics, including demographic and anatomical parameters, were largely comparable between completers and non-completers, suggesting that attrition was not primarily driven by baseline disease severity, although residual bias cannot be excluded. Estimates of treatment burden based on annualized injection frequency should be interpreted with caution, as early discontinuation may lead to an underestimation of true treatment requirements.

Consequently, late follow-up outcomes should be interpreted with caution and in conjunction with sensitivity analyses among 12-month completers. In this context, it should be acknowledged that completer analyses may overestimate treatment effects, as patients with more favorable response or tolerability are more likely to remain under follow-up. Conversely, available-case analyses in the presence of informative dropout may also be biased, as patients discontinuing early because of adverse events or insufficient response are underrepresented at later time points.

Safety outcomes warrant particular attention in the context of aflibercept 8 mg. A subset of discontinuations in our cohort was related to inflammatory adverse events, including IOI. The clinical presentation, management, and outcomes of these events are summarized in Supplementary Table S1. Recent real-world reports, including a large single-center analysis from a tertiary referral center, have described higher-than-expected rates of sterile IOI following aflibercept 8 mg compared with pivotal trial data, although most cases resolved without permanent visual sequelae [23]. Additional case series and multicenter observations have similarly reported clusters of IOI events shortly after injection, underscoring the importance of post-marketing surveillance for new high-dose formulations [22, 24].

In our study, an early cluster of inflammatory events occurred during a phase in which aflibercept 8 mg was prepared by subdividing a single vial for use in multiple patients under sterile conditions in a certified cleanroom by trained hospital pharmacists. While such vial-splitting practices are relatively common in routine care for economic reasons, they represent off-label handling of the drug. Following the occurrence of clustered IOI events, the preparation protocol was changed to on-label administration using either a single vial per patient or manufacturer-provided prefilled syringes. Notably, no further cases of IOI were observed after this change and all events resolved without permanent visual loss. Although causality cannot be established in an observational study, this temporal association suggests that preparation and handling procedures may contribute to inflammatory risk, in line with mechanistic considerations discussed in the recent IOI literature [23, 24].

Importantly, this safety signal should not be generalized to all aflibercept 8 mg use. The observed IOI events occurred within a narrowly defined time window, were temporally associated with off-label aliquoting of the drug, and were not observed after transition to on-label preparation. These findings suggest that preparation and handling procedures may have played a critical role and are consistent with emerging real-world data indicating that IOI rates may vary depending on formulation and preparation conditions [25–27].

These observations emphasize that real-world safety outcomes may differ from those reported in controlled clinical trials, where drug handling, patient selection, and monitoring are highly standardized [11, 13]. They also highlight the importance of strict adherence to on-label preparation and administration procedures, particularly when introducing higher-dose or new intravitreal therapies into routine clinical practice.

Several limitations of this study should be acknowledged. As a single-center, single-arm observational study without a comparator group, all outcomes represent within-group changes and do not allow causal attribution to the switch to aflibercept 8 mg and limit generalizability. No formal sample size calculation was performed as a result of t limited prior data in this specific clinical setting, and the study should therefore be considered exploratory. In addition, multiple statistical comparisons were conducted without adjustment for multiplicity. Moreover, the statistical approach relied on repeated pairwise comparisons at predefined time points and did not include longitudinal modeling of repeated measures. Given the substantial dropout and the potential for informative censoring, this approach may be susceptible to bias and should be interpreted with caution. Longitudinal methods, such as mixed-effects models, may provide a more robust analytical framework in future studies with larger cohorts and more complete follow-up. Detailed classification of lesion subtype (e.g., type 1/2 MNV, RAP, or PCV) was not consistently feasible because of long prior treatment histories and limited availability of baseline imaging prior to treatment initiation. Qualitative OCT grading was performed in a real-world clinical setting and was not conducted in a masked or multigrader framework; therefore, intergrader agreement was not assessed. The high discontinuation rate reduces the number of patients available for long-term analysis and introduces a risk of informative censoring and survivorship bias; the calculation of annualized injection burden in patients with early discontinuation may introduce uncertainty, and these estimates should be interpreted with caution. Moreover, the observed reduction in treatment burden was modest and should be interpreted cautiously. Nevertheless, a key strength of this study is its prospective design, allowing for standardized data collection and systematic follow-up in a real-world clinical setting. To our knowledge, this is among the first prospective real-world studies evaluating aflibercept 8 mg in a previously treated high burden nAMD population.

Conclusion

Our findings suggest that switching to aflibercept 8 mg in chronically treated patients with nAMD requiring short re-treatment intervals may be associated with modest anatomical improvement, improved disease control, and stabilization of visual function. A reduction in treatment burden was observed in a subset of patients, although overall conclusions remain limited considering the high discontinuation rate and observed safety findings.

Importantly, the occurrence of clustered IOI events associated with off-label aliquoted preparation highlights the critical importance of appropriate handling and administration procedures for aflibercept 8 mg in clinical practice. Careful patient selection, close monitoring, and strict adherence to on-label preparation and administration procedures are essential to optimize safety and clinical outcomes.

Overall, these findings should be interpreted as descriptive and hypothesis-generating. Further multicenter studies with larger cohorts and longer follow-up periods are needed to better define the role of aflibercept 8 mg in real-world treatment algorithms for nAMD.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Conceptualization: Oliver Findl, Christoph Spartalis; Methodology: Christoph Spartalis, Marlies Ullrich, Oliver Findl; Formal analysis and investigation: Christoph Spartalis, Marlies Ullrich; Data collection: Christoph Spartalis, Marlies Ullrich, Silvia Winkler, Christoph Leisser, Manuel Ruiss, Caroline Pilwachs; Writing: original draft preparation: Christoph Spartalis; Writing: review and editing: Christoph Spartalis, Marlies Ullrich, Silvia Winkler, Christoph Leisser, Manuel Ruiss, Caroline Pilwachs, Oliver Findl; Resources: Oliver Findl; Supervision: Oliver Findl; Funding acquisition: Oliver Findl.

All authors read and approved the final manuscript.

Funding

This investigator-initiated study was financially supported by Bayer, which also provided aflibercept 8 mg for use in the study. The sponsor had no role in the study design; data collection, analysis, or interpretation; manuscript preparation; or the decision to submit the manuscript for publication. The journal’s Rapid Service Fee was funded by the authors.

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Christoph Spartalis reports consultancy for Bayer. Oliver Findl is a scientific advisor to Carl Zeiss Meditec AG, Croma, and Johnson & Johnson. Marlies Ullrich, Silvia Winkler, Christoph Leisser, Manuel Ruiss, and Caroline Pilwachs declare that they have no competing interests.

Ethical Approval

The study was approved by the Ethics Committee of the City of Vienna (EK 24-105-0824, NCT07390253) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. No identifying participant information is included in this manuscript.

Thanking Patient Participant(s)

The authors thank the patients who participated in this study for their contribution to clinical research.

Patient Involvement

Patients were not involved in the design, conduct, reporting, or dissemination of this research.