Seven-Year Outcomes of Aflibercept in Neovascular Age-Related Macular Degeneration in a Teaching Hospital Setting

Abstract

Background: In clinical practice, visual outcomes with anti-VEGF therapy may be worse than those observed in clinical trials. In this study, we aim to investigate the long-term outcomes of neovascularization treated with intravitreal aflibercept injections (IAI) in a teaching hospital setting. Methods: This is a retrospective, single-center study including 81 nAMD patients (116 eyes), those both newly diagnosed and switched from ranibizumab. All patients had a follow-up duration of at least seven years. Treatment involved three monthly injections followed by either a pro re nata (PRN) or treat and extend regimen. Follow-up care was primarily conducted by training physicians. The primary endpoint was the change in best-corrected visual acuity (BCVA) over seven years. Secondary endpoints included central retinal thickness changes, qualitative OCT parameters, macular atrophy progression, injection frequency, and treatment adherence. Results: Among the 116 eyes, 52 (44.8%) completed the seven-year follow-up. Visual acuity improved by +2.1 letters in the overall population (+6.3 letters in treatment-naive eyes) after the loading phase but gradually declined, resulting in a loss of −12.3 letters at seven years. BCVA remained stable (a loss of fewer than 15 letters) in 57.7% of eyes. Central retinal thickness (CRT) decreased significantly during follow-up in both naive and switcher eyes. Macular atrophy occurred in 94.2% of eyes, progressing from 1.42 mm² to 8.55 mm² over seven years (p < 0.001). The mean number of injections was 4.1 ± 1.8 during the first year and 3.7 per year thereafter. Advanced age at diagnosis was a risk factor for loss to follow-up, with bilaterality being a protective factor against loss to follow-up (p < 0.05). Conclusions: This study highlights the challenges faced by a retina clinic in a teaching hospital. Suboptimal functional and anatomical outcomes in real life may derive from insufficient patient information and inconsistent monitoring, which contributes to undertreatment and affects long-term visual outcomes. It also raises concerns about supervision in a teaching hospital which needs to be improved.

1. Introduction

Neovascular age-related macular degeneration (nAMD) is a chronic progressive disease of macula with a poor long-term visual prognosis [1]. The three subtypes of neovascular degeneration are polypoidal choroidal vasculopathy, retinal angiomatous proliferation or type 3 choroidal neovascularization (CNV), and typical nAMD including CNV confined to the sub-RPE space (type 1) or growth through the RPE (type 2) [2]. The Consensus Nomenclature for Reporting Neovascular Age-Related Macular Degeneration Study Group added OCTA as an imaging modality for a more detailed classification of nAMD [3].

Anti-vascular endothelial growth factor (VEGF) therapy is the gold standard treatment for nAMD. The efficacy of ranibizumab, aflibercept, and brolucizumab has been proven through clinical trials [4,5,6]. However, these pivotal studies do not reflect results found in real-life practice.

Treatment decisions typically rely on visual acuity assessment, spectral-domain optical coherence tomography (SD-OCT), and angiography OCT (OCT-A). Variations in treatment regimens and morphological characteristics are believed to influence VA outcomes in anti-VEGF therapy. Additionally, patient compliance and adherence to treatment—factors closedly tied to age and comorbidities—play a crucial role in achieving sucessful coutcomes. For instance, the SEVEN-UP study, which assessed the seven-year outcomes of intravitreal injections of ranibizumab, administrated initially via a fixed regimen over 2 years, followed by a pro re nata (PRN) regimen, revealed a decrease in visual acuity over time.

In this real-world study, we investigated the functional and anatomic seven-year outcomes in nAMD patients treated with intravitreal aflibercept injections (IAI), in both naives and switcher eyes, using an individualized regimen. Secondary outcomes were injection frequency, macular atrophy progression, and loss to follow-up rates in a teaching hospital setting.

2. Materials and Methods

2.1. Study Design and Ethics

This is a retrospective single-center study conducted in a French university-based hospital. Local IRB approval was obtained at Picardie Jules Verne University (reference PI2024_843_0125, approval date 22 November 2024). This study was carried out in accordance with the Declaration of Helsinki.

2.2. Patients

This study included patients diagnosed with neovascular AMD, either newly diagnosed (treatment-naive) or previously treated (switched from ranibizumab due to suboptimal response), who had started with 2 mg/0.05 mL aflibercept intravitreal injection between January 2013 and June 2017. The main reason for switching was suboptimal response during the follow-up, defined as intraretinal fluid (IRF) or subretinal fluid (SRF) persistence 1 month after injection.

The exclusion criteria were age < 50 years, other retinal diseases (uveitis, diabetic retinopathy, vein retinal occlusion), ocular surgery in the 3 months prior injection, or other causes of choroidal neovascularization.

2.3. Study Cohort

Diagnosis of neovascular AMD was established using multimodal retinal imaging (structural OCT, OCT-A, Fluorescein angiography, and Indocyanine Green angiography (HRA2; Heidelberg Engineering GmbH, Heidelberg, Germany)) [7]. The type of choroidal neovascularization (CNV) was graded as type 1 (sub-RPE), type 2 (subretinal), and type 3 or polypoidal choroidal vasculopathy [8,9].

During the follow-up, assessments included best corrected visual acuity using the Early Treatment Diabetic Retinopathy Study (EDTRS) scale, measurement of intraocular pressure, slit lamp examination, and a spectral domain OCT. The measurement of central retinal thickness and macular volume were calculated automatically by the software (Heidelberg Eye Explorer, version 1.10.2.0, Heidelberg, Germany).

2.4. Organization of Retina Clinic and Treatment Protocol

All patients followed a structured treatment protocol, starting with a loading dose of 3 monthly intravitreal aflibercept injections. After this, treatment decisions were based on disease activity using a reactive “pro re nata” (PRN) protocol from 2013 to 2018. The proactive “treat and extend“ regimen with a ±2 weeks extension or reduction interval was applied thereafter.

In the reactive protocol, visits were scheduled every month including VA assessment, funduscopy, and OCT. Fluorescein and Indocyanine Angiography were performed if any doubt. Disease activity was defined as visual loss of more than 5 letters (not attributed to macula atrophy or fibrosis), and/or new macular hemorrhage on funduscopy, and/or detection of any fluid on OCT, or increased central foveal thickness. In the absence of any sign of CNV activity, further visits were scheduled at least every month. In cases of activity, retreatment was indicated, and intravitreal injection was performed as soon as possible.

In the treat and extend protocol, if there were no signs of recurrence, a new injection was administered and the next injection was extended by 2 weeks, up to a maximum interval of 16 weeks. If there was disease activity, treatment interval was thereby shortened by 2 weeks at a time.

Treatment and follow-up schedules were entirely at the discretion of the treating physicians. Intervals between decision of treatment and injection varied from the same day to 2 weeks, depending on resource availability. As the study population was treated in a university-based hospital, follow-up care was mainly provided by trainee physicians, who rotated every 6–12 months, under the supervision of senior retinal specialists.

2.5. Outcome Measures

Outcome measures were best corrected visual acuity, quantitative (Subfoveal Central thickness), and qualitative OCT parameters (intra and subretinal fluid distribution). Early Treatment Diabetic Retinopathy Study (ETDRS) score letters were used in our clinical practice. Number of intravitreal injections were collected at each endpoint. Anti-VEGF therapy change was also reported.

Macular atrophy (MA) was defined as alterations in the outer layers of the retina, associated or not with disruption of retinal pigment epithelium, according to the Classification of Atrophy Meeting group [10]. MA was evaluated using near-infrared imaging and SD-OCT. The MA area was measured before aflibercept administration, at 1 year, 2 years, and at 4 years and 7 years. Angiograms and OCT images were annotated by two experienced graders (T.T.H.C and A.B.) [11].

Treatment compliance during follow-up was analyzed on the basis of non-adherence, non-persistence, and discontinuation using the consensus criteria described by Okada et al. [12]. Non-adherence was defined as missing 2 or more treatment or visits over 12 months, with a visit considered missed if it exceed more than 2 weeks from the recommended date. Non-persistence was defined as non-attendance of any treatment or follow-up visit within the last 6 months. Discontinuation was defined as intentionally “planned discontinuation”, and “transfer of care” was also recorded.

Loss to follow-up in AMD treated with anti-VEGF was defined as the presence of at least 1 interval longer than 6 months without any visit (follow-up or injection). Patients were not considered LTFU if they had died within 6 months of the last visit or had been referred to another site.

Data were collected from the electronic medical records and entered into a computerized table at initiation of aflibercept, 3 months, 1 year, 2 years, 4 years, and 7 years.

Statistical analysis was performed using IBM SPSS Statistics Software or Windows (version 26, SPSS, Chicago, IL). Descriptive data are presented as mean, standard deviation (SD) and 95% confidence interval, range, minimum, maximum, and percentage when appropriate. The paired Student’s t-test was used to compare paired continuous variables at different time points. Correlations between visual change at 7 years and initial visual acuity, visual change at 3 months, and the number of IVTs were performed using Pearson’s correlation coefficient. One way ANOVA was conducted to study the relationship between VA change and different parameters. Binary logistic regression was used to examine different factors in loss to follow-up or non-adherent groups, with univariate and multivariate analysis. A p value of <0.05 was considered statistically significant.

3. Results

3.1. Study Population and Baseline Characteristics

Demographic and clinical data of the study population are summarized in

Table 1. Clinical and demographic characteristics at inclusion.

	Overall Population	Naive Subgroup
Eyes number (patients)	116 (81)	41 (33)
Mean age ± SD (years) (min-max)	80.5 ± 7.1 (59–95)	82.1 ± 5.7 (72–94)
Sex, n (%)   Male   Female	41 (35.3) 75 (64.7)	28 (6.3) 13 (31.7)
Bilateral desease, n	34	8
Previous treatments, n   Ranibizumab (%)   Previous IVT number ± SD (min-max)   PDT + Ranibizumab (%)	75 (64.7) 10.1 ± 7.8 (2–50) 3 (2.6)	0 (0)00 (0)
BCVA (EDTRS) ± SD (min-max)	56.2 ± 21.2 (5–85)	52.8 ± 21.8 (5–80)
CNV type, n (%)   Type 1   Type 2   Type 3	78 (67.2) 19 (16.4) 19 (16.4)	30 (73.2) 6 (14.6) 5 (12.2)

n, number; BCVA, best corrected visual acuity; PDT, photodynamic therapy; CNV, choroidal neovascularization; SD, standard deviation.

and Figure 1. Of the 118 eyes treated with aflibercept injections for neovascular AMD between November 2013 and November 2016, 116 met the inclusion criteria. The mean age of patients at diagnosis was 80.5 ± 7.1 years (range: 59–95), with 75 females (64.7%) and 41 males (35.3%) (Table 1). The mean follow-up duration was 5.2 ± 2.2 years (min–max, 0.25–7).

Among the study eyes, 41 (35.3%) were treatment-naive, and 75 (64.7%) had been switched from ranibizumab to aflibercept due to suboptimal response. The mean number of prior anti-VEGF injections in the switched group was 10.1 ± 7.8 (range: 2–50). The mean BCVA at inclusion was 56.2 ± 21.2 EDTRS letters (range, 5–85). A total of 52 eyes completed seven years of follow-up (44.8%) (Figure 1).

3.2. Visual Outcome

BCVA at diagnosis in the overall population was 56.2 ± 21.2 letters (n = 116), 58.1 ± 20.4 at 3 months (n = 114), 55.4 ± 22.9 at 1 year (n = 106), 51.6 ± 22.6 at 2 years (n = 103), 51.4 ± 22.7 (n = 77) at 4 years, 47.1 ± 25.2 at 7 years (n = 52). There was a visual gain of 2.1 ± 13.6 letters (p = 0.93) at 3 months, but it was followed by progressive declines: a loss of −2.5 ± 14.8 letters (p = 0.094) at 1 year which did not reach statistical significance. Changes in BCVA became significant, with visual loss of −5.2 ± 15.9 letters (p = 0.001) at 2 years, −8.1 ± 19.9 letters at 4 years (p < 0.001), and −12.3 ± 25.4 letters (p < 0.001) at seven years.

In the naive eye subgroup, the BCVA was 56.2 ± 22.8 at diagnosis (n = 41), 58.1 ± 21.8 at 3 months (n = 41), 57.1 ± 20 at 6 months (n = 37), 56.5 ± 20.9 at 1 year (n = 35), 54.1 ± 21.9 at 2 years (n = 34), 53.4 ± 22.7 at 4 years (n = 29), and 51.3 ± 21.4 at 7 years (n = 16). BCVA improved significantly by +6.3 ± 14.5 letters at 3 months (p = 0.008). However, the visual gains decreased over time: from a non-significant change of 0.4 ± 17.5 letters at one year (p = 0.89) to −1.5 ± 20 letters (p = 0.67) at two years, −4 ± 24.6 letters at four years (p = 0.39), and −10.6 ± 32.6 letters at seven years (p = 0.21) (Figure 2).

At seven years, 15.4% of eyes achieved a gain of 15 or more letters, 57.7% maintained stable vision (loss < 15 letters), and 38.5% experienced significant visual loss (>15 letters). The primary causes of severe vision loss (>15 letters) were macula atrophy (17 cases), pigment epithelial tears involving the fovea (two cases), and sub-retinal hemorrhage (one case). VA change ranged from a maximum gain of 55 letters to a maximum loss of 50 letters.

There was a significant correlation between final VA at seven years and initial VA (r = 0.341; p = 0.013), but no correlation was found with BCVA at three months (r = 0.061; p = 0.666), or the total number of injections (r = 0.006; p = 0.966) or the CNV type (p = 0.693).

3.3. Anatomical Response to Aflibercept

The central retinal thickness in the entire cohort was 316 ± 98 μm (n = 116) at diagnosis, which decreased to 265 ± 72 μm (n = 114) at 3 months, 294 ± 95 μm at 1 year (n = 106), 286 ± 66 at 4 years (n = 77), and 267 ± 70 (n = 52) at 7 years. CRT reductions were significant at all time points: −52 ± 98 μm at 3 months (p < 0.001), −20 ± 19.9 μm at 1 year (p = 0.05), −30 ± 102 μm (p = 0.017) at 4 years, and −46 ± 102 μm (p = 0.003) at 7 years (Figure 3).

In the treatment-naive subgroup, CRT decreased from 343 ± 110 μm (n = 41) at diagnosis to 242 ± 46 µm at seven years, with significant reductions noted at three months (−76 ± 106 μm, p < 0.001) and seven years (−94 ± 96 μm, p = 0.001).

3.4. Distribution of Fluid and Qualitative SD-OCT Analysis

At diagnosis, 100% of eyes exhibited fluid on OCT-SD. At the last visit (n = 52), fluid was present in 51.9%, distributed as follows: intraretinal fluid only in 16 eyes (30.8%), subretinal fluid only in three eyes (5.8%), both intra and subretinal fluid in six eyes (11.5%) (

Table 2. Anatomical outcomes at last visit, overall study population versus naive eyes.

Outcome	Overall Population (n = 52)	Naive Subgroup (n = 16)	p Value *
CRT, mm
Mean ± SD	267 ± 70	242 ± 46	0.184
Mean change ± SD	−45.7 ± 103	−94 ± 96	0.099
Macular volume, mm3
Mean	8.10 ± 1.09	7.87 ± 0.55	0.420
OCT fluid; n (%)
Absent	27 (51.9)	9 (52.9)	0.942
Present	25 (48.1)	8 (47.1)
IRF	22 (42.3)	7 (41.2)	0.935
SRF	9 (17.3)	1 (5.9)
Macular atrophy, size mm2
Mean ± SD	8.71 ± 8.95	8.74 ± 8.36	0.99

CRT Central Retinal Thickness, SD standard Deviation, n Number, IRF Intraretinal Fluid, SFR Subretinal Fluid. * p < 0.05.

).

In the naive population, fluid resolution was achieved in 25/41 eyes at 3 months (61%), in 12/35 eyes at 1 year (34%), in 17/34 eyes at 2 years (50%), and in 12/24 eyes at 5 years (50%).

3.5. Progression of Atrophy

Macular atrophy (foveal sparing and foveal involving) was observed in 94.2% (49/52 eyes) at 7 years. Macular atrophy area size increased significantly over time: +1.42 ± 2.36 mm² (p < 0.001) at year 1; +2.92 ± 3.56 mm² (p < 0.001) at year 2: +5.64 ± 6.93 mm² at year 4 (p < 0.001); +6.16 ± 7.56 mm² at year 5 (p < 0.001); +7.57 ± 8.15 mm² at year 6 (p < 0.001); and +8.55 ± 9.4 mm² at year 7 (p < 0.001) (Figure 4).

There was no significant association between macular atrophy progression over years and CNV type (p = 0.9).

3.6. Injection Frequency and Treatment Compliance

The number of injections was 4.1 ± 1.8 (min–max, 1–10) in the first year, decreasing to 3.7 injections per year thereafter. Over the seven years, patients received a total of 18.6 ± 14.1 injections on average (range, 1–57).

In the treatment-naive subgroup, 33/41 eyes (80.5%) completed the loading phase of 3 monthly injections. These patients received an average 4.7 ± 1.9 injections during the first year.

Among naive patients, 33/41 eyes (80.5%) received monotherapy through follow-up. Among these, 8/41 eyes (20.5%) were switched to ranibizumab after a mean interval of 2.9 ± 1.8 years. Three months after switching from aflibercept to ranibizumab, visual acuity (VA) had fallen by −2.5 ± 5.3 letters (p = 0.227), CRT had increased by 34.1 ± 77.0 μm, and all eyes (8/8) still had fluid. Subsequently, 4/8 eyes (50%) switched back to aflibercept.

Using one-way ANOVA, we did not find a relationship between CNV type and number of injections in the overall population (p = 0.234). The number of intravitreal injections was 25 ± 14.5 injections for type 1 (n = 33); 24.7 ± 18.6 injections for type 2 (n = 7); and 29.5 ± 10.4 injections for type 3 (n = 12). In the naive population, we were unable to analyze the influence of CNV type as a function of the number of injections, due to the small number of eyes for each subgroup.

3.7. Compliance and Adverse Events

Adherence to treatment was assessed through loss of sight, non-adherence, and non-persistence.

Thus, 40 eyes (34.5%) were lost to follow-up and did not reach seven years of follow-up: 18 eyes (15.5%) due to death and six (5.2%) due to transfer of care. Univariate analysis revealed that age and unilateral involvement were risk factors for visual loss in our cohort (p < 0.05), whereas gender, treatment-naive status, and initial BCVA did not appear to have any influence. These parameters were confirmed by multivariate analysis using binary logistic regression. The model showed that the probability of being lost to follow-up increased by 19.7% with each additional year of age at diagnosis (p < 0.001). Patients with bilateral involvement were 91% less likely to be lost to follow-up than those with unilateral involvement (p < 0.001) (

Table 3. Risk factors for visual loss in patients with neovascular AMD.

	Lost to Follow Up	Univariate Model p-Value (Odds Ratio; 95% CI)	Multivariate Model p-Value (Odds Ratio; CI 95%)
N	40
Age (years)	83.5	0.001 * (1.123; 1.046–1.206)	<0.001 * (1.157; 1.068–1.252)
Sex
male	15	Reference	Reference
female	25	0.725 (0.867; 0.391–1.922)	0.457 (0.707; 0.284–1.763)
Involved eye
unilateral	21	Reference	Reference
bilateral	19	0.036 * (0.497; 0.202–0.597)	0.012 * (0.304; 0.120–0.770)
Inital BCVA (EDTRS)	56.4	0.731 (1.001; 0.983–1.019)	0.636 (1.005; 0.985–1.026)
Treatment status
switched	24	Reference	Reference
naïve	16	0.447 (1.360; 0.615–3.006)	0.780 (0.844; 0.257–2.773)

CI Confidence interval, N Number, BCVA Best corrected visual acuity, * Indicate p-value < 0.05.

).

In total, 32/116 eyes (27.6%) experienced non-adherence to treatment, primarily due to health deterioration (50%), COVID-19 related disruptions (34%), or treatment burden. Non-persistence was observed in 5.2% of patients.

Using binary linear regression, no statistically significant association was found between non-adherence and age, gender, bilaterality, initial BCVA, and treatment-naive status.

No adverse ocular events (endophthalmitis, retinal detachment, glaucoma) were reported during the 7-year follow-up period.

4. Discussion

Anti-VEGF agents, such as ranibizumab, aflibercept, and brolucizumab, can inhibit the growth of neovascular lesions, improve vision, and resolve retina edema in clinical trials [4,5,6]. Ranibizumab, a humanized monoclonal antibody fragment that neutralizes all of the soluble isoforms of vascular endothelial growth factor (VEGF)-A, is primarily indicated for neovascular AMD [4] according to ANCHOR [13] (Antibody for the Treatment of Predominantly Classic Choroidal in Age related Macular Degeneration) and MARINA [14] (Minimally Classic/Occult Trial of the anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular Age-Related Macular Degeneration) studies. Visual gain was 11.3 and 7.2 letters, respectively, at one year post ranibizumab.

Aflibercept is a fusion protein which binds to VEGF-A, VEGF-B, and placental growth factor (PIGF). After the loading phase, intravitreal aflibercept injections (IAI) every 2 months has been shown to be as safe and effective as ranibizumab monthly injections in phase III of VIEW 1 and VIEW 2 studies at one year [15]. Both drugs were shown to be equally effective in maintaining visual acuity (VA) [4] at year 2. Under aflibercept therapy, mean BCVA improved by 8.9 ETDRS letters at 52 weeks from baseline and visual acuity was maintained in 95.4%. At 2 years, visual gain was 7.6 letters using an as-needed regimen and mandatory dosing at least every 12 weeks. The total number of injections was 7.5 at 52 weeks and 11.2 at the 96 weeks endpoint [4]. Ranibizumab has been available since 2007, whereas aflibercept has been available since November 2013 in France. Both ranibizumab and aflibercept are approved for nAMD and fully reimbursed.

In this study, we investigated the seven-year real-world outcomes of treatment-naive and switched eyes treated with intravitreal aflibercept injection (IAI) for neovascular AMD in a teaching hospital. Differing from clinical trials that do not extend for a long period of time, this study provides insights into the long-term progression of treated nAMD in real-life conditions. Over the study period, treatment regimens evolved from ranibizumab to aflibercept in switched eyes, and from PRN to the treat and extend protocol. This study also highlighted the challenges of managing a retina clinic in a teaching hospital environment, where resource limitations and reliance on trainee physicians may lead to suboptimal treatment organization.

The key findings are as follows: (1) Visual gain of +6.3 letters obtained after the loading dose dropped rapidly to +0.4 letters at one year in naive eyes and decreased gradually to −10.6 letters at 7 years without the plateau phase. (2) Central retinal thickness significantly decreased thorough the follow-up in both naive and non-naive eyes. At the seven-year endpoint, fluid resolution was achieved in half of the cases. (3) Macular atrophy growth was progressive, affecting 94.2% of eyes at seven years. (4) The number of IAI was low during the first year (mean 4.1 injections), then remained stable at around 3.7 IAI per year. (5) Rate of loss to FU was high, reaching 55.2% over 7 years. (6) Non-adherence was observed in 27.6%, while non-persistence occurred in 5.2%.

4.1. Visual Outcomes

In our real-life study, BVCA loss following the loading phase was substantial, with an overall decrease of −12.3 letters at seven years (−10.6 letters in naive eyes). Using a cut-off of 15 ETDRS letters, BCVA improved in 15.4% eyes, was stable in 57.7%, and declined in 38.5%. These findings are comparable to the SEVEN-UP study (including eyes treated with 2-year fixed regime, followed by PRN ranibizumab) which reported a decline of −8.6 letters over 7.3 years, with 34% of the eyes losing more than 15 letters [16]. In Hama‘s report including naive eyes of typical nAMD and Polypoidal Choroidal Vasculopthy treated with a 1-year fixed regimen followed by PRN aflibercept, BCVA changed from 0.2 logMAR to 0.29 logMAR over the 7-year observational period; it was stable in 68% and declined in 19%. The difference in visual outcomes between our report and those above-mentioned may be attributable in treatment protocols: the SEVEN-UP study used a fixed regimen for two years, followed by PRN treatment, while our study predominantly employed a PRN approach from the outset.

Inadequate visual outcomes in our cohort could be partially explained by organizational challenges, including extended intervals between monitoring visits and treatment delays. Firstly, the visits were not scheduled monthly, and there was a tendency to extend the visit interval for up to 3 months. Secondly, retreatment decisions was often postponed up to 10 days, while French experts recommend a maximum delay of 4 days [17]. Real-life studies in France using aflibercept demonstrated 5.3 letters gain [18] to 6 letters gain at 1 year in naive eyes [19] with an average of six aflibercept injections. In switcher eyes, our study showed a VA gain of +2.1 letters at month 3 and loss of −5.2 letters at 2 years whereas previous studies showed +3.2 letters at month 3 and no visual change at 2 years [20]. Additionally, 19.5% of naive eyes in our study did not complete the required 3 monthly injections during the loading phase. These factors contributed to undertreatment which likely impacted the visual outcome negatively.

4.2. Morphological Outcomes

In terms of morphological changes over 7 years, the CRT decreased from 316 µm to 267 µm (−18%) in the overall cohort, and from 343 µm to 242 µm (−41%) in naive eyes. Reduction of CRT in a previous study using aflibercept in naive eyes including both typical nAMD and PCV was −54% [21]. While these reductions align with prior studies using aflibercept, which reported a −54% reduction in CRT, our cohort experienced an increase between the third and the sixth months, indicative of reactivation during the observation phase.

Overall, the anatomical gain remained significant over time until the 7-year endpoint, as in previous real-world observational prospective studies using aflibercept for naive nAMD [15,19]. Fluid resolution was achieved in half of eyes at year seven, which is similar to the rate of Gayadine et al. [22]. This proportion is notably higher than that observed in the CATT study which reported fluid resolution only in 17% of eyes [23].

4.3. Macular Atrophy

Macular atrophy was present in 94% of eyes by the end of the study. Hama et al. observed a rate of foveolar atrophy of 22% at 7 years in naive patients treated with aflibercept with fixed PRN at a 7-year regimen [21]. In the SEVEN-UP prospective study of nAMD treated with ranibizumab, a macular atrophy rate of 98% was found after 7.3 years of treatment, which is similar to ours [16].

We did not find a difference in 7-year MA area among CNV types; whereas other reports found a higher rate of MA in CNV type 3 [24] and a lower rate in CNV type 1 [25]. A systematic literature review from clinical trials showed that atrophy occurs in the context of MNV treated with anti-VEGF therapy. However, it is not clear whether anti-VEGF treatment is causative of atrophy versus being associated with atrophy development since clinical trials were not designed or powered to assess atrophy as a primary outcome. It is important to recognize that adequately treating exudative MNV remains the best option to optimize vision outcomes in patients with nAMD, particularly given the risk of vision loss with undertreatment observed in the real world.

4.4. Compliance and Lost to Follow-Up

Lost to follow-up (LTFU) affected 34.5% of patients in our cohort, with advanced age identified as a significant risk factor. Bilateral disease appeared to protect against LTFU, potentially due to patients’ heightened concern for preserving vision in the second eyes. These findings align with those of Okada et al. [26], who reported conflict regarding bilaterality as a protective factor. Gender and initial visual acuity did not seem to be a factor in LTFU in our study, whereas other studies reported association between male gender and low initial acuity as a risk of LTFU. In a retrospective and prospective case controls study, Kusenda et al. [27] found that the four most common reasons for LTFU were general health worsening (21.8%), patient-missed appointments (16.7%), COVID-19-related issues (14.9%), and treatment dissatisfaction (8.6%).

Over the 7-year duration, only 5.2% of patients experienced a period of non-persistence and 32 eyes (27.6%) of non-adherence. Multiple factors determine non-adherence and non-persistence, including age at the condition, therapy, patient, social/economic, and health systems/healthcare team levels. The percentage of non-adherence found in our study seems to be significantly lower than that reported in the systematic review by Okada who found a non-adherence rate ranging from 32 to 95% of patients [26]. The high rate of adherence in our study is likely due to France’s healthcare system, which provides full reimbursement for treatment and transportation, but it would be higher if patients were followed up by the same physician instead of rotated trainee physicians. After analyzing the results of this study, we have reorganized our retina clinic. Patients received an interview to receive information about treatment protocol and a notebook on follow-up with scheduled appointments according to the prescribed interval injection. The retina clinic was handled by both senior and trainee physicians, with immediate feedback on OCT interpretation and treatment decisions before leaving.

Our study has the strength of a long-term, real-world design, and the homogeneity of the French healthcare system, which minimizes socioeconomic disparities. However, limitations include the retrospective nature of the study and its bias: data collection bias, loss to follow-up bias, and variability in clinical evaluation and treatment protocols. The frequent turnover of trainee physicians and resource constraints further contributed to heterogeneity in patient care. Our study also highlights that organization, leading to a longer interval injection than required, may affect the visual outcome, suggesting a need to change the mindset that contributes to undertreatment.

4.5. Perspectives

To improve outcomes, future efforts should focus on addressing organizational barriers to reduce undertreatment. Emerging therapies [28] with longer retreatment intervals may enhance compliance and alleviate the treatment burden for patients and healthcare systems.

5. Conclusions

In conclusion, managing nAMD in a teaching hospital is challenging, where resource limitations and non-uniform care protocols can lead to undertreatment. The suboptimal functional and anatomic outcomes in real-life may derive from insufficient patient information, and inconsistent monitoring. Our study also highlights the need for better supervision in a hospital teaching environment. Understanding and addressing these barriers are critical for improving long-term visual outcomes in real-world settings.