
With the extension of life expectancy and population aging, age-related macular degeneration (AMD) has become one of the major diseases causing vision impairment in the elderly [1-3]. AMD can be categorized into two types: dry AMD and neovascular AMD (nAMD). While dry AMD accounts for approximately 90% of cases, nAMD tends to progress more rapidly and cause severe vision loss, resulting in a more significant impact on patient quality of life [2-4]. Despite various treatments being attempted, the current firstline treatment involves anti-vascular endothelial growth factor (anti-VEGF) injections [2,3].
Anti-VEGF injections inhibit the growth of immature neovascularization in the choroid and prevent f luid or blood leakage into the subretinal, intraretinal, or sub-retinal pigment epithelium (sub-RPE) space [2,4]. Anti-VEGF injections not only reduce disease activity, but also lead to meaningful improvements in vision [5]. The choroid plays a crucial role in metabolism by supplying energy and nutrients to photoreceptors [2]; consequently, choroid thickness may be associated with a functioning RPE, outer retina, and good vision [2]. There are reports of significant reductions in sub-foveal choroidal thickness (SFCT) following anti-VEGF injections [4,6-10]; however, there are also studies indicating no significant changes in choroidal thickness after such treatment [11,12]. Furthermore, research into the long-term effects of anti-VEGF on choroidal thickness is lacking.
Therefore, the authors investigated the long-term effects of anti-VEGF injection therapy on the choroidal thickness in nAMD eyes. Additionally, the relationship between SFCT and visual acuity (VA) was investigated.
Patients diagnosed with nAMD and naïve to treatment between March 2014 and June 2017 were retrospectively analyzed using medical records. Exclusion criteria were as follows: any history of vitrectomy, anti-VEGF therapy or photodynamic therapy, evidence of end-stage AMD such as subfoveal fibrosis or atrophy, evidence of other retinal diseases including central serous chorioretinopathy and other neovascular maculopathies, glaucoma, and poor imaging data quality. In addition, cases in which SFCT could not be measured due to difficulty in accurately determining the choroidal-scleral boundary were excluded from the study. Polypoidal choroidal vasculopathy (PCV) was diagnosed based on the presence of polypoidal lesions with or without branching vascular networks [13,14]. Cases that exhibited retinal-retinal or retinal-choroidal anastomoses were classified as type 3 neovascularization (retinal angiomatous proliferation, RAP) [13,14]. The remaining patients who were not diagnosed with either PCV or RAP were classified as having typical nAMD with type 1 or type 2 choroidal neovascularization (CNV) [15].
During the study period, all patients received injections of one of three drugs: bevacizumab (Avastin®; Genetech, Inc.), ranibizumab (Lucentis®; Novartis AG and Genetech Inc.), or aflibercept (Eylea®; Regeneron and Bayer HealthCare). All patients were given 0.05 cc intravitreal anti-VEGF injections monthly for three months. Thereafter, patients received additional intravitreal injections based on the PRN regimen according to the doctor's judgment over the following fiveyear follow-up period if persistent or recurrent subretinal or intraretinal fluid was evident, new macular hemorrhage developed, or the extent of pigment epithelial detachment increased.
Best-corrected visual acuity (BCVA) was measured using the manifest refraction test and then converted to the logarithm of the minimal angle of resolution (logMAR). Intraocular pressure (IOP, NT-530P; Nidek) and spectral domain optical coherence tomography (SD-OCT, version 5.3.2.0; Heidelberg Engineering) were examined at baseline; at one, three, and six months; and then annually until five years after injection. The BCVA, IOP, and SFCT were compared between the nAMD and fellow eyes at every follow-up visit.
Enhanced-depth imaging (EDI)-OCT using the Heidelberg Spectralis (Heidelberg Spectral Domain Optical Coherence Tomography; Heidelberg Engineering) platform was used. SFCT was measured at the subfoveal point extending from the outer border of the pigment epithelium to the choroidal scleral boundary using the built-in caliper software. For better accuracy, horizontal and vertical macular B-scans were measured, and the average thickness, measured independently by two authors, was used for analysis.
This study was approved by the Institutional Review Board of Hangil Eye Hospital (IRB number: IRB-23006). All study procedures adhered to the tenets of the Declaration of Helsinki. The requirement for informed consent was waived because of the retrospective nature of the study, and all clinical data used in the analysis were anonymized.
Continuous variables are expressed as mean ± standard deviation, and categorical variables are described as proportions. The SFCT was measured in the nAMD and fellow eyes and compared using a paired t-test, which was also used to analyze differences between baseline and post injection values within each group. A multiple regression model was used to evaluate factors affecting final VA. Baseline VA, age, sex, injection number, nAMD subtype, baseline CMT, baseline SFCT, final CMT, and final SFCT were corrected.
All data were entered into an Excel spreadsheet (Microsoft Corp.) and analyzed using SPSS software (version 23; IBM Corp.). Statistical significance was determined when p-values were less than 0.05.
Fifty-seven eyes of 57 patients with nAMD were included in this study. The baseline characteristics of the patients are summarized in Table 1. The mean age of the patients was 66.46 ± 8.64 years. Typical nAMD was the most frequent nAMD subtype (36 eyes, 63.1%), followed by PCV (18 eyes, 31.6%) and RAP (3 eyes, 5.3%). The mean number of anti-VEGF injections was 12.44 ± 90. The percentages of the injection drugs used at the first injection was 56% (25 eyes) for ranibizumab, 28% (16 eyes) for aflibercept, and 28% (16 eyes) for bevacizumab. In addition, 29 eyes (51%) received one drug intravitreal injection for five years, 24 eyes (42%) received two drugs, and four eyes (7%) received three drugs. We further analyzed the change in SFCT by dividing the group into subgroups of those that did not switch injection drugs and those that did. There was no significant difference between the group that received only one type of drug and did not switch that drug (29 eyes, mean SFCT change 45.90 ± 48.93 μm) and the group that switched drugs (28 eyes, mean SFCT change 45.5 ± 40.00 μm; p = 0.462). The mean IOP of the AMD eyes was 15.72 ± 2.68 mmHg at baseline. The BCVA improved significantly in the nAMD eyes up to one year after injection, but then started to deteriorate. Compared to the fellow eyes, nAMD eyes showed significantly poorer BCVA over the entire follow-up (Table 2).
In nAMD eyes, CMT decreased significantly at each follow-up visit compared to baseline (p < 0.001 at every visit). nAMD eyes showed significantly higher CMT than that of the fellow eyes at baseline (421.2 ± 147.9 vs. 265.2 ± 66.3, p < 0.001); from four years after injection, the CMT showed no significant difference between nAMD and fellow eyes (p = 0.135, 0.063, Table 2).
In nAMD eyes, SFCT decreased significantly at each follow-up visit compared to baseline (p < 0.001 at every visit, Fig. 1). The SFCT of nAMD eyes showed significant reductions between baseline and one year, between two and three years, and between four and five years. The SFCT of fellow eyes showed a tendency to decrease during the five-year follow-up (p < 0.05). The SFCT of nAMD and fellow eyes showed no significant difference at baseline and after one year. However, a significant difference was observed from the second year onwards, and the SFCT of nAMD eyes was significantly thinner at the final five years (Table 2).
Since there were only three RAP cases among all nAMD subjects, it was difficult to perform quantitative comparisons among the three groups, but changes in SFCT choroidal thickness in each group were as follows. The SFCT of PCV (18 eyes) was 193.94 ± 100.49 μm at baseline and 153.33 ± 95.88 μm at five years after injection, and the change in SFCT was 40.61 ± 39.61 μm. In the case of RAP (3 eyes), the baseline was 134.00 ± 18.38 μm, the final time point was 110.00 ± 28.28 μm, and the change was 24.00 ± 46.67 μm. In other typical nAMD (36 eyes) excluding PCV and RAP, the baseline was 213.33 ± 92.00 μm, the five-year follow-up time point was 170.86 ± 87.89 μm, and the SFCT change was 47.22 ± 45.84 μm.
A correlation analysis was performed to determine the factors affecting the final VA. Final VA had a significant relationship only with baseline VA (standardized coefficient beta = 0.380, p = 0.011), whereas other factors including age, sex, nAMD subtype, baseline CMT, baseline SFCT, final CMT, final SFCT, and injection number were not significantly related (Table 3).
This study aimed to investigate the long-term effects of anti-VEGF injections on choroidal thickness in treatment-naïve nAMD patients and the association with VA. The results show that, under the PRN regimen, choroidal thickness in nAMD eyes significantly decreased each year compared to baseline. Furthermore, from the second year onward, choroidal thickness in nAMD eyes became significantly thinner than in fellow eyes. The final VA after five years did not show a significant difference from baseline VA and was not associated with age, sex, number of injections, nAMD subtype, baseline CMT, SFCT, final CMT, or SFCT but was significantly correlated with baseline VA.
Several previous studies reported a decrease in choroidal thickness following anti-VEGF injections, similar to our results. Yeom et al. observed a significant reduction in CMT, choroidal thickness, and PED height in patients receiving brolucizumab injections under the treat-and-extend (TNE) protocol after switching therapy in recurrent nAMD eyes [16]. Minnella et al. [4] conducted a study on treatment-naïve nAMD eyes and observed significant choroidal thinning compared to baseline after three consecutive monthly aflibercept injections, with the reduction occurring two weeks after the third injection. In a study by Pellegrini et al. [9], significant reductions in choroidal thickness and vascularity were observed after three consecutive monthly aflibercept injections in treatment-naïve nAMD eyes. Similarly, Inan et al. [8] reported a significant reduction in choroidal thickness 14 months after three consecutive ranibizumab injections at one-month intervals, followed by additional injections based on the TNE protocol in treatment-naïve nAMD eyes.
Conversely, there are studies showing no significant reduction in choroidal thickness after anti-VEGF injections. For instance, some studies reported no significant reduction in choroidal thickness after three consecutive ranibizumab injections in treatment-naïve nAMD eyes [11]. In addition, Sizmaz et al. [12] examined changes in choroidal thickness after a single anti-VEGF injection in treatment-naïve nAMD eyes and found no significant decrease in choroidal thickness one month after the injection, regardless of injection of ranibizumab or bevacizumab. Similarly, another study observed changes following aflibercept injections in treatment-naïve nAMD eyes using the modified TNE method, but no significant change in choroidal thickness was seen for up to two years after injection [17].
Unlike these earlier studies that focused mainly on relatively short-term changes, this study investigated the longterm effects of anti-VEGF therapy on choroidal thickness in nAMD eyes over a period of up to five years. The results demonstrated that choroidal thickness in nAMD eyes continued to significantly decrease compared to baseline for up to five years after the first anti-VEGF injection. Furthermore, this study analyzed changes in choroidal thickness in the fellow eyes that did not receive anti-VEGF treatment. Choroidal thickness in these untreated eyes also decreased compared to baseline, which is presumed to be due to aging. In this study, patients initially received monthly anti-VEGF injections for three months, and subsequent injections were administered based on a PRN regimen at the doctor’s discretion. This differs from the TNE method commonly used in most existing studies and reflects real-world clinical outcomes when the PRN regimen is applied. Since the pattern of choroidal thickness changes may vary depending on the injection regimen, additional studies investigating long-term changes in choroidal thickness according to treatment regimen are needed.
Inan et al. [8] reported no significant relationship between final VA and the mean change in choroidal thickness 14 months after ranibizumab injections in treatment-naïve nAMD eyes. They also identified that disciform scars, geographic atrophy, and baseline VA were predictors of final VA [8]. According to a cross-sectional analysis conducted by Manjunath et al. [18], there was no significant correlation between choroidal thickness, number of injections, time since nAMD diagnosis, and VA in nAMD eyes. On the other hand, Kang et al. [19] found that VA at six months after ranibizumab injections in treatment-naïve nAMD eyes was significantly associated with baseline choroidal thickness, baseline VA, and baseline CNV size. In the present study, only baseline VA was significantly associated with final VA five years after anti-VEGF injections in treatment-naïve nAMD eyes. Age, sex, number of injections, nAMD subtype, baseline CMT, baseline SFCT, final CMT, and final SFCT were not significantly related to final VA. The differences in study results could be attributed to variations in patient group characteristics, such as age, sex, nAMD subtype, stage of nAMD progression, and the type of anti-VEGF used in each study. Additionally, variations in the methods used to measure choroidal thickness could contribute to different outcomes. Furthermore, the length of the study period may also influence the results.
This study has several limitations. First, this was a retrospective study and the number of subjects was insufficient, with only 57 eyes. Therefore, detailed analysis by nAMD subtype could not be conducted. Studies on changes in choroidal thickness and VA by subtype with a larger number of subjects are necessary. Second, the specific type of anti-VEGF injection could not be distinguished. Due to frequent switching between anti-VEGF drugs during the longterm follow-up, it was not possible to analyze the effects of each injection type separately. Since the drugs may have different effects on the anatomical and functional changes of nAMD, long-term follow-up studies are necessary to evaluate the impact of each drug type. Third, we did not include other factors that may affect VA, such as CNV size and presence of new geographic atrophy or disciform scars. Last, this study was conducted only on one racial group, Koreans, making it difficult to generalize the results to other racial or ethnic groups.
In summary, this retrospective study investigated the longterm effects of anti-VEGF injections on choroidal thickness and VA in treatment-naïve nAMD eyes. We found that choroidal thickness significantly decreased each year after anti-VEGF injections, with nAMD eyes showing thinner choroidal thickness than fellow eyes from the second year onward. VA improved significantly for up to one year after injections but then deteriorated, remaining significantly worse than in fellow eyes throughout the follow-up period. Final VA significantly correlated with baseline VA. Larger-scale studies are needed to further investigate the effects of nAMD subtype and different anti-VEGF drug on choroidal thickness and VA outcomes.
The authors declare no conflicts of interest relevant to this article.
Conception (D.D.H.); Design (D.D.H.); Data acquisition (D.D.H.); Analysis (A.Y.L., D.D.H.); interpretation (A.Y.L., D.D.H.); writing (A.Y.L., D.D.H.); review (D.D.H.); Final approval of the article (All authors)
![]() |
![]() |