Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Predicting long-term functional anti-VEGF treatment outcomes in neovascular AMD in a real-world setting

Abstract

Purpose

To test to what degree retinal fluid (RF) after the loading phase and at the end of year 1 predicts long-term functional outcomes in neovascular macular degeneration (nAMD), as do macular (MA) atrophy, treatment density and treatment interval extension.

Methods

In this retrospective single-center cohort study, a consecutive series of eyes with treatment-naïve nAMD followed under a treat-and-extend (T&E) protocol followed over ≥2 years. Best-corrected visual acuity (BCVA), presence of retinal fluid (RF) and macular atrophy (MA) were registered along with central retinal thickness (CRT) and treatment density over time. The relationship between these variables was tested by regression analysis.

Results

A total of 433 eyes were followed for 4.9 ± 2.2 years. A series of univariate analyses were run to select the covariates for the final multivariate regression model. CRT after loading, time to dryness, intraretinal fluid and MA after one year were found to predict visual function over 2 to 5 years. A final regression model was adjusted for visual acuity (VA) at baseline and showed that CRT after loading was predictive only in the short term (2 years) and that MA had the greatest predictive value for VA after 2 to 5 years. Intraretinal fluid (IRF) significantly predicted VA only after 4 years. The final regression model explained 21 to 32% of the variation in VA.

Conclusions

In this large retrospective cohort, the presence of MA after one year was the strongest predictor of VA after 2 to 5 years, explaining a vision loss of 13 to 20 letters. The presence of IRF and SRF at any point of time had a comparably weak predictive potential for the outcomes over 5 and more years.

Citation: Pfister IB, Schild C, Garweg JG (2024) Predicting long-term functional anti-VEGF treatment outcomes in neovascular AMD in a real-world setting. PLoS ONE 19(11): e0314167. https://doi.org/10.1371/journal.pone.0314167

Editor: Tatsuya Inoue, Yokohama City University, JAPAN

Received: June 11, 2024; Accepted: November 5, 2024; Published: November 25, 2024

Copyright: © 2024 Pfister et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All data used for the analyses, as well as the R code for the regression analyses, can be found here: https://osf.io/7fgmk/ The DOI is: 10.17605/OSF.IO/7FGMK. All study findings can be replicated using this data.

Funding: The author(s) received no specific funding for this work.

Competing interests: None of the authors received direct or indirect support for this investigator-initiated study or have any conflicting interests with the data that are presented in this report.

Introduction

Several factors have been reported to contribute to the functional and anatomic prognosis of neovascular AMD (nAMD) patients receiving anti-VEGF treatment. Low baseline visual acuity (VA), preexisting structural damage to the macula, submacular hemorrhage and lesion size may be beyond the non-influenceable outcome predictors [1]. VA after the first year is closely linked to early diagnosis and treatment initiation [1–7]. Due to a ceiling effect, patients with a baseline VA above 20/40 are expected to have a lower visual gain but, on the other hand, will have the best final visual outcomes over time [8, 9]. Due to a ceiling effect, patients with excellent baseline VA may not gain but may even lose single letters, but attain an excellent functional outcome, whereas patients with very low visual acuity (below 20/100) may gain more than 15 letters; nevertheless, they do not reach reading or driving vision over 2 years [10] or up to 5 years [11]. On the other hand, the baseline VA may be less predictive of visual potential than the baseline VA after the loading phase [9, 12, 13].

Persistent disease activity, as evidenced by the presence of IRF and subretinal fluid (SRF), may result from insufficiently intensive early treatment [14] or incomplete treatment response. Persistent fluid after the loading phase is linked to a less favorable functional outcome even in eyes without significant structural damage [15]. Up to 66% of patients, depending on the anti-VEGF drug used, exhibit persistent disease activity after the loading phase in a real-world setting [16]. If complete control of disease activity is not achieved after the loading phase, a visual decline is expected within 12 to 48 months (PRN study arms of the Harbor [17], Catt [18–20], Montblanc [21]), whereas consequent suppression of VEGF activity and thereby prevention of lesion growth will result in visual stability for up to 5 years (fixed treatment arms in the Anchor [22], Marina [23], Harbor [24], CATT [20] and View studies [25]). However, disease stability (absence of intraretinal fluid), which is achieved in two-thirds of eyes after one year, will be maintained at only 53% over 2 years [26]. Independent of treatment, macular atrophy can develop or progress in virtually any patient at some point [22].

Given the aforementioned evidence, we wished to model the long-term functional outcomes of anti-VEGF therapy after correcting for baseline VA in a treatment-agnostic approach based on VA and morphological parameters in eyes receiving three loading injections and followed for at least two years.

Patients and methods

In this single-center retrospective cohort study, we included patients with treatment-naïve nAMD who began receiving intravitreal treatment (IVT) with ranibizumab (Ran) or aflibercept (Afl) between October 2012 and October 2021 under a T&E protocol that included three initial loading injections at monthly intervals and who were followed for at least two years. According to this protocol, intervals were adopted with 2-week in- or decrements based on OCT findings: Interval extension was allowed in the absence of intraretinal and subretinal fluid, or if subretinal fluid and PED had remained unchanged over 3 consecutive visits; shortening of the interval was requested in case of new or increasing fluid in any compartment or new retinal hemorrhage indicating disease activity. Data were collected between October 2023 and January 2024. Eyes with preexisting structural damage preventing a functional recovery and major submacular hemorrhages were excluded from the analysis. Furthermore, patients needed to have at least four clinical visits as well as at least five IVTs in the first year. Best-corrected distance visual acuity (BCVA) was transformed into early treatment of diabetic retinopathy score (ETDRS) values, where a Snellen decimal BCVA of 1.0 was defined as 85 ETDRS letters. Furthermore, OCT data including CRT in all eyes and the presence of IRF and/or SRF were retrospectively collected from the patient records at the following time points: at baseline, at three and 12 months and annually thereafter along with the treatment demand and intervals over time. MA at the end of year 1 was not quantified but was categorized as absent or present if incomplete (iRORA) or complete RPE and outer retinal atrophy (cRORA), incomplete (iORA) or complete outer retinal atrophy (cORA) were identified in the foveal center. These criteria were based on specific OCT findings: (1) a foveal region of hypertransmission of at least 250 μm in diameter, (2) a zone of attenuation or disruption of the RPE of at least 250 μm in diameter, (3) evidence of overlying photoreceptor degeneration, and (4) absence of scrolled RPE or other signs of an RPE tear [27]. Furthermore, the time to dryness was deducted from the clinical records, and the treatment intervals were recorded. From this, we calculated how many eyes could extend treatment to ≥10 weeks within the first two years of therapy. In contrast to the majority of clinical trials, treatment was considered extended if at least three consecutive intervals were ≥10 weeks.

The study was approved by the Ethics Committee of the Canton of Bern (KEK no. 2022–01850) and is fully compliant with the tenets of the Declaration of Helsinki in its latest version. Each participant provided a general written consent for the use of their coded clinical data.

All variables are measured at the between subject level. No outliers were eliminated for any of the analyses. Multiple regression analyses were applied to predict short- (2 years) and long-term (3–5 years) therapy responses. The data are presented as the mean ± standard deviation (SD), median and 25% and 75% interquartile ranges (IQRs), unless otherwise stated. The level of significance was set at p < 0.05. All the statistical analyses were performed using R (software version 4.3.2). We chose VA at baseline, change in VA from baseline to the end of the loading and to 1 year of therapy as possible control variables for the regression analyses. CRT after loading, change in CRT from baseline to the end of loading, and to 1 year, SRF after loading or after 1 year, IRF after loading or after 1 year, and IRF as well as SRF after loading or after 1 year, time to dryness, interval extension to > 10 weeks in the first year or in the second year, macular atrophy at baseline, after loading and after the first year and number of injections after the first year were considered as possible independent variables. Out of these parameters, for our multivariate analyses we chose the ones which maximally explained the variance in all cases.

All data used for the analyses, as well as the R code for the regression analyses, can be found here: https://osf.io/7fgmk/

The DOI is: 10.17605/OSF.IO/7FGMK. All study findings can be replicated using this data.

Results

A total of 433 eyes of 404 consecutive patients admitted to our clinic for the treatment of their neovascular AMD between 2012 and 2021 met the inclusion criteria. The demographic information of the participants is presented in Table 1. Of the 433 eyes, 228 (52.7%) were pseudophakic before their first anti-VEGF Treatment, 108 (24.9%) underwent a cataract surgery during follow up. Cataract surgery during follow up had no influence on the development of MA after 1 year (r = 0.052, p = 0.30), and did not impact visual acuity at any time point in these 108 eyes compared to the rest of the sample. A total of 39.5% of the sample was lost to follow-up. Of these, 15.3% moved away, 11% died, and 3% reached stability and decided to stop treatment. Another 3% paused treatment because of systemic illness, whereas 2% stopped treatment because of the loss of a visual potential. One percent of the patients wished to discontinue treatment, and the remaining 4.2% stopped treatment for unknown reasons. The findings for VA are displayed in Figs 1–3: Fig 1 shows the changes in VA from baseline, stratified in three groups based on baseline visual acuity (BCVA < 0.2, n = 82; BCVA 0.25–0.5, n = 217; BCVA > 0.5, n = 133). The first two groups experienced an increase in VA after the loading phase and remained stable thereafter. The group with the best baseline VA experienced a less pronounced VA increase after baseline, but remained stable over the entire follow up period. Group comparisons between the three groups (Kruskal-Wallis H test) showed that the change in VA was different between the three groups from baseline to years one, two and three (all p < 0.001). For the change from baseline to years four and five, a difference was found between group <0.2 and the other two groups (p < 0.05). Fig 2 displays eyes with IRF after one year with those who did not, regarding the change in VA over time. Fig 3 compares the change in VA over time in eyes without and with MA after one year. Eyes with MA experienced a decrease in VA over time, whereas eyes without MA remained stable over 5 years after the loading. The course of CRT and its change over time compared to baseline are displayed in Fig 4.

thumbnail
Download:
Fig 1. Change in visual acuity (VA, in ETDRS letters) compared to baseline, stratified by baseline VA (Decimal BCVA <0.2; 0.25–0.5; >0.5).

https://doi.org/10.1371/journal.pone.0314167.g001

thumbnail
Download:
Fig 2. Visual acuity over time stratified by the presence of intraretinal fluid (IRF) after the 1st year.

https://doi.org/10.1371/journal.pone.0314167.g002

thumbnail
Download:
Fig 3. Visual acuity over time stratified by the presence of macular atrophy after the 1st year (MA).

https://doi.org/10.1371/journal.pone.0314167.g003

thumbnail
Download:
Fig 4. Changes in central retinal thickness (CRT) during the treatment period up to 5 years of follow up.

https://doi.org/10.1371/journal.pone.0314167.g004

thumbnail
Download:
Table 1. Epidemiological data at baseline.

https://doi.org/10.1371/journal.pone.0314167.t001

Retinal fluid

The portion of eyes with retinal fluid at the different time points is displayed in Fig 5. After the loading phase, almost half of the sample had a dry retina (48.5%), 16.4% of the eyes showed IRF, 21.2% SRF, and 12.2% had both IRF and SRF.

thumbnail
Download:
Fig 5. Presence of retinal fluid during the treatment period up to 5 years of follow up.

https://doi.org/10.1371/journal.pone.0314167.g005

Treatment interval extension

Within the total sample of 433 eyes, a mean treatment interval extension to 11.1 ± 3.5 weeks (median: 11.0, IQR: 9–13) was registered. The mean time to reach ≥10 weeks was 15.5 ± 13.9 months (median: 10.1, IQR: 6.9–19.2; range: 8–83.9 months; S1 Fig). The treatment interval was extended to ≥10 weeks over three consecutive visits within the first two years of therapy in half of the sample (50.8%).

Multiple regression analysis

We first performed a series of univariate regression analyses to identify covariates with predictive potential for short- (2 years) and long-term (3–5 years) VA. We examined CRT levels after loading and compared them to the change in CRT from baseline to after loading or the change from baseline to the first year. We also examined the presence of IRF, SRF, and RF after the loading phase and after the first year, as well as the time to dryness and interval extension to ≥10 weeks within the 1st and 2nd years. Finally, we examined MA at baseline, after the loading phase and after one year as well as the number of injections in the first year as covariates for the prediction of VA (Table 2). Covariates with a significant predictive value over at least 2 timepoints were integrated into a multivariate model in the second step. In cases with several options for the same variable (CRT, IRF and SRF as well as MA), the timepoint with the most explanatory potential (R2) was chosen. Here, we found that CRT after loading, IRF after the 1st year, time to dryness and MA after 1 year had the strongest predictive potential for VA after 2 to 5 years. MA at baseline was not predictive for VA at any time point, explaining maximally 2% of the variation. MA after the loading was significant only for VA at 3 years, explaining 8% of the variation in VA, while MA at year 1 explained 18% of the variation. Interval extension to ≥10 weeks was predictive of short-term, but not long-term VA (up to 3 years). We therefore decided not to include this covariate in our model. The effect sizes of the variables which were included into the final model are displayed in S1 Table.

thumbnail
Download:
Table 2. Univariate regression analysis of prognostic parameters possibly predicting long term visual acuity.

Parameters with significant predictive power over at least 2 time points were included in the multivariate model.

https://doi.org/10.1371/journal.pone.0314167.t002

We also examined which was the best control variable, VA at baseline, compared to the change in VA from baseline to after loading or after one year. Here, we found that VA at baseline was a stronger predictor of long-term VA than change in VA.Therefore, we included VA at baseline as a control variable in our model.

In the second step, we constructed a model containing all covariates which had been identified in univariate analyses to have predictive potential for VA after control for VA at baseline (Table 3). The confidence intervals for the covariates in the models are reported in S2 Table. In this regression model, we found that interval extension to ≥10 weeks within the first 2 years of treatment was predictive for VA at years 2–4, but not thereafter. In the multivariate analysis, the presence of intraretinal fluid at year 1 was predictive for VA at 4 years but not before or thereafter. CRT and time to dryness did not predict VA at any of the time points (Table 3). The covariate with the strongest and consistent predictive value for VA at years 2 to 5 was MA at year 1: VA was worse in eyes with MA than in those without MA. The eyes of patients with MA had a mean of 13 letters lower in VA after 2 years, 16.6 letters after 3 years, and 20.1 letters after 5 years.

thumbnail
Download:
Table 3. Results of the multiple linear regression analyses including different predictors for long-term visual acuity (with visual acuity at baseline as control).

https://doi.org/10.1371/journal.pone.0314167.t003

CRT explained approximately 1 to 2% of the variation in VA, IRF as well as time to dryness explained 2 to 4% of the variation. The interval extension within the 1st or 2nd year moderately predicted between 1 and 4% of variability. All of these parameters represent significant, but small effect sizes. Finally, MA at year 1 after diagnosis reached a clinically relevant effect size, explaining 18% of the long-term visual outcomes. After inclusion of all of the aforementioned parameters, this model explainedwith 23–34% of the variation in VA a relevant effect size.

Discussion

According to the univariate model, CRT after loading, intraretinal fluid after 1 year and time to dryness, taken alone, predicted, to some extent, explaining between 1 and 8% of the variation in medium- and long-term visual outcomes (from 2 to 5 years). MA alone, in contrast, closely predicted visual outcomes, explaining 15% to 18% of the variation in VA. This effect persisted in the multivariate model for 2 to 5 years. Given the increasing effect of this parameter over time, it seems surprising that MA has not systematically been reported [1, 15, 28] and has not been introduced as an outcome predictor in prospective randomized clinical trials (Fig 3). It might be explained by the fact that MA regularly escapes detection at baseline because of overlay phenomena. Introducing MA at year 1 as a covariate in the regression model in our series demonstrated its strong impact on long-term VA development. AMD is known to progress to MA over time, and current evidence indicates that anti-VEGF treatment may partially contribute to its development [29]. In our study, we included MA that was present at the end of year 1, and did not take into account newly developing MA beyond year 1, which supports the robustness of this finding. Poor visual acuity at baseline, retinal angiomatous proliferation, presence of intraretinal fluid and absence of subretinal fluid, have been linked to MA in a recent prospective study [30].

This model may namely gain interest for the comparison of anti-VEGF drug effects after correcting for baseline functional characteristics, i.e. baseline VA. After correction for baseline VA, we were able to assess the impact of morphological OCT-derived structural parameters, such as the presence of IRF and SRF after the 1st year, as well as MA at the end of year 1, onto the long-term functional outcomes of treatment. If all of the above-mentioned parameters were recorded in a standardized fashion, it would be well conceivable to compare drug effects between studies independent of the treatment demand, including morphological treatment response and drying potential.

A strength of our study is its robust sample size, long follow-up period of 5 and more years, a treatment-agnostic approach, use of robust and easily identified discriminators, and correction for baseline VA. We aimed to exclude a selection bias induced by the exclusion of data from patients not followed for a longer time or due to treatment cessation because of an unsatisfactory functional response by comparing the predictive power for the whole cohort with that of only patients still under treatment, which revealed no difference. On the other hand, our results showed that IRF, but not SRF, after the loading phase and after 1 year is predictive of the visual outcome. However, the robustness of these qualitative findings should be independently confirmed, ideally through semiquantitative analyses based on central retinal fluid volume. The retrospective design, on the other hand, may have resulted in a reduced continuity of the data because not all patients came to their assigned study time points. Therefore, a tolerance of 20% for the choice of clinical examination time points had to be accepted to escape a major number of missing entries or the need to implement the last observation to be carried forward.

Conclusions

In conclusion, the presence of intraretinal, but not subretinal fluid at the end of the 1st year may predict functional treatment response. MA after the first year of treatment, but not–as might be expected—at baseline or after the loading phase was the strongest independent predictive biomarker for short- and long-term treatment response. This model does not aim to impact treatment strategies but offers a prognostic tool for predicting the long-term functional outcomes in the clinical setting and eventually also in and between studies. Hence, MA alone explains 18%, and after inclusion of further biomarkers in a treatment-agnostic multivariate regression model up to 32% of the variation in visual outcomes.

Supporting information

S1 Table. Effect sizes in univariate analyses.

https://doi.org/10.1371/journal.pone.0314167.s001

(DOCX)

S2 Table. Confidence intervals for the covariates included in the multiple regression model.

https://doi.org/10.1371/journal.pone.0314167.s002

(DOCX)

S1 Fig. Treatment interval distribution.

https://doi.org/10.1371/journal.pone.0314167.s003

(TIF)

References

  1. 1. Finger RP, Wickremasinghe SS, Baird PN, Guymer RH (2014) Predictors of anti-VEGF treatment response in neovascular age-related macular degeneration. Surv Ophthalmol 59: 1–18. pmid:24332379
  2. 2. Cohen SY, Mimoun G, Oubraham H, Zourdani A, Malbrel C, et al. (2013) Changes in visual acuity in patients with wet age-related macular degeneration treated with intravitreal ranibizumab in daily clinical practice: the LUMIERE study. Retina 33: 474–481. pmid:23266880
  3. 3. Muether PS, Hoerster R, Hermann MM, Kirchhof B, Fauser S (2013) Long-term effects of ranibizumab treatment delay in neovascular age-related macular degeneration. Graefes Arch Clin Exp Ophthalmol 251: 453–458. pmid:22573410
  4. 4. Bloch SB, la Cour M, Sander B, Hansen LK, Fuchs J, et al. (2013) Predictors of 1-year visual outcome in neovascular age-related macular degeneration following intravitreal ranibizumab treatment. Acta Ophthalmol 91: 42–47. pmid:22008284
  5. 5. Real JP, Luna JD, Urrets-Zavalia JA, De Santis MO, Palma SD, et al. (2013) Accessibility as a conditioning factor in treatment for exudative age-related macular degeneration. Eur J Ophthalmol 23: 857–864. pmid:23661541
  6. 6. Muether PS, Hermann MM, Koch K, Fauser S (2011) Delay between medical indication to anti-VEGF treatment in age-related macular degeneration can result in a loss of visual acuity. Graefes Arch Clin Exp Ophthalmol 249: 633–637. pmid:20865421
  7. 7. Real JP, Granero GE, De Santis MO, Juarez CP, Palma SD, et al. (2015) Rate of vision loss in neovascular age-related macular degeneration explored. Graefes Arch Clin Exp Ophthalmol 253: 1859–1865. pmid:25491161
  8. 8. Ying GS, Huang J, Maguire MG, Jaffe GJ, Grunwald JE, et al. (2013) Baseline predictors for one-year visual outcomes with ranibizumab or bevacizumab for neovascular age-related macular degeneration. Ophthalmology 120: 122–129. pmid:23047002
  9. 9. Ying GS, Maguire MG, Daniel E, Ferris FL, Jaffe GJ, et al. (2015) Association of Baseline Characteristics and Early Vision Response with 2-Year Vision Outcomes in the Comparison of AMD Treatments Trials (CATT). Ophthalmology 122: 2523–2531.e2521. pmid:26383996
  10. 10. Lee AY, Lee CS, Butt T, Xing W, Johnston RL, et al. (2015) UK AMD EMR USERS GROUP REPORT V: benefits of initiating ranibizumab therapy for neovascular AMD in eyes with vision better than 6/12. Br J Ophthalmol 99: 1045–1050. pmid:25680619
  11. 11. Gillies MC, Campain A, Barthelmes D, Simpson JM, Arnold JJ, et al. (2015) Long-Term Outcomes of Treatment of Neovascular Age-Related Macular Degeneration: Data from an Observational Study. Ophthalmology 122: 1837–1845. pmid:26096346
  12. 12. Reznicek L, Muhr J, Ulbig M, Kampik A, Mayer WJ, et al. (2014) Visual acuity and central retinal thickness: fulfilment of retreatment criteria for recurrent neovascular AMD in routine clinical care. Br J Ophthalmol 98: 1333–1337. pmid:24903670
  13. 13. Rasmussen A, Bloch SB, Fuchs J, Hansen LH, Larsen M, et al. (2013) A 4-year longitudinal study of 555 patients treated with ranibizumab for neovascular age-related macular degeneration. Ophthalmology 120: 2630–2636. pmid:23830760
  14. 14. Holz FG, Tadayoni R, Beatty S, Berger A, Cereda MG, et al. (2015) Multi-country real-life experience of anti-vascular endothelial growth factor therapy for wet age-related macular degeneration. Br J Ophthalmol 99: 220–226. pmid:25193672
  15. 15. Zandi S, Weisskopf F, Garweg JG, Pfister IB, Pruente C, et al. (2017) Pre-Existing RPE Atrophy and Defects in the External Limiting Membrane Predict Early Poor Visual Response to Ranibizumab in Neovascular Age-Related Macular Degeneration. Ophthalmic Surg Lasers Imaging Retina 48: 326–332. pmid:28419398
  16. 16. Tan CS, Lim LW, Ngo WK, Pannirselvam P, See C, et al. (2020) Predictors of persistent disease activity following anti-VEGF loading dose for nAMD patients in Singapore: the DIALS study. BMC Ophthalmol 20: 324. pmid:32762659
  17. 17. Sarraf D, London NJ, Khurana RN, Dugel PU, Gune S, et al. (2016) Ranibizumab Treatment for Pigment Epithelial Detachment Secondary to Neovascular Age-Related Macular Degeneration: Post Hoc Analysis of the HARBOR Study. Ophthalmology 123: 2213–2224. pmid:27566855
  18. 18. Martin DF, Maguire MG, Ying GS, Grunwald JE, Fine SL, et al. (2011) Ranibizumab and bevacizumab for neovascular age-related macular degeneration. N Engl J Med 364: 1897–1908. pmid:21526923
  19. 19. Martin DF, Maguire MG, Fine SL, Ying GS, Jaffe GJ, et al. (2020) Ranibizumab and Bevacizumab for Treatment of Neovascular Age-related Macular Degeneration: Two-Year Results. Ophthalmology 127: S135–s145. pmid:32200813
  20. 20. Maguire MG, Martin DF, Ying GS, Jaffe GJ, Daniel E, et al. (2016) Five-Year Outcomes with Anti-Vascular Endothelial Growth Factor Treatment of Neovascular Age-Related Macular Degeneration: The Comparison of Age-Related Macular Degeneration Treatments Trials. Ophthalmology 123: 1751–1761. pmid:27156698
  21. 21. Larsen M, Schmidt-Erfurth U, Lanzetta P, Wolf S, Simader C, et al. (2012) Verteporfin plus ranibizumab for choroidal neovascularization in age-related macular degeneration: twelve-month MONT BLANC study results. Ophthalmology 119: 992–1000. pmid:22424834
  22. 22. Rofagha S, Bhisitkul RB, Boyer DS, Sadda SR, Zhang K (2013) Seven-year outcomes in ranibizumab-treated patients in ANCHOR, MARINA, and HORIZON: a multicenter cohort study (SEVEN-UP). Ophthalmology 120: 2292–2299. pmid:23642856
  23. 23. Rosenfeld PJ, Brown DM, Heier JS, Boyer DS, Kaiser PK, et al. (2006) Ranibizumab for neovascular age-related macular degeneration. N Engl J Med 355: 1419–1431. pmid:17021318
  24. 24. Sadda SR, Tuomi LL, Ding B, Fung AE, Hopkins JJ (2018) Macular Atrophy in the HARBOR Study for Neovascular Age-Related Macular Degeneration. Ophthalmology 125: 878–886. pmid:29477692
  25. 25. Schmidt-Erfurth U, Kaiser PK, Korobelnik JF, Brown DM, Chong V, et al. (2014) Intravitreal aflibercept injection for neovascular age-related macular degeneration: ninety-six-week results of the VIEW studies. Ophthalmology 121: 193–201. pmid:24084500
  26. 26. Garweg JG, Gerhardt C (2021) Disease stability and extended dosing under anti-VEGF treatment of exudative age-related macular degeneration (AMD) - a meta-analysis. Graefes Arch Clin Exp Ophthalmol 259: 2181–2192.
  27. 27. Sadda SR, Guymer R, Holz FG, Schmitz-Valckenberg S, Curcio CA, et al. (2018) Consensus Definition for Atrophy Associated with Age-Related Macular Degeneration on OCT: Classification of Atrophy Report 3. Ophthalmology 125: 537–548. pmid:29103793
  28. 28. Xue K, Hua P, Maguire MG, Daniel E, Jaffe GJ, et al. (2023) Prediction for 2-Year Vision Outcomes Using Early Morphologic and Functional Responses in the Comparison of Age-related Macular Degeneration Treatments Trials. Ophthalmol Retina 7: 564–572. pmid:36803692
  29. 29. Foss A, Rotsos T, Empeslidis T, Chong V (2022) Development of Macular Atrophy in Patients with Wet Age-Related Macular Degeneration Receiving Anti-VEGF Treatment. Ophthalmologica 245: 204–217. pmid:34695835
  30. 30. Gemenetzi M, Lotery AJ, Patel PJ (2017) Risk of geographic atrophy in age-related macular degeneration patients treated with intravitreal anti-VEGF agents. Eye (Lond) 31: 1–9. pmid:27716750