Purpose To report the long-term efficacy of patients with neovascular age-related macular degeneration (nAMD) treated with anti-vascular endothelial growth factor (VEGF) in Changhua Christian Hospital in Taiwan.

Method Retrospective case series of patients with nAMD that were treated with intravitreal injection of anti-VEGF and had a minimum follow up of 48 months. Every patient was initially treated with 3 loading doses of either bevacizumab or ranibizumab, followed by a loose treat and extend regimen. Eyes were divided into 2 groups according to whether aflibercept was later used as a rescue therapy (group 2) or not (group 1). Patients underwent best-corrected visual acuity (BCVA) testing, optical coherence tomography, and ophthalmic examination at baseline and all the scheduled follow-up visits.

INTRODUCTION Neovascular age-related macular degeneration (nAMD) is one of the most common disorders that cause irreversible blindness in the senior population in the developed world. If left untreated the patients may experience gradual central visual impairment and metamorphopsia which could severely affect their quality of life. However, with the development and application of intravitreal injection of anti-vascular endothelial growth factor (VEGF), the visual prognosis of nAMD has been greatly improved.Currently there are three commonly prescribed anti-VEGF agents. Aflibercept (Eylea; Regeneron Pharma, Tarrytown, New York, USA; Bayer, Basel, Switzerland), ranibizumab (Lucentis; Novartis Pharma, Basel, Switzerland; Genentech, South San Francisco, Chicago, USA), and bevacizumab (Avastin; Genentech). Generally speaking, the binding affinity of VEGF is stronger in aflibercept than in ranibizumab and bevacizumab. Previous studies have shown that aflibercept may be effective in eyes resistant to ranibizumab or bevacizumab.

Though the studies of the 2 years results of anti-VEGF agents showed great efficacy in improving vision in nAMD, the visual outcome of the long term treatment were inconsistent.Besides, studies about the long-term efficacy in Asians were limited.In this retrospective study, we aim to investigate the 4 year result of nAMD treated with anti-VEGF agents in Taiwan.

MATERIALS AND METHOD A retrospective, interventional case series analysis o of patients with nAMD treated at the Ophthalmology department at Changhua Christian Hospital for at least 4 years was conducted. The study was approved by the institutional review board of Changhua Christian Hospital and was adhered to the tenets of the Declaration of Helsinki. Only patients who were both treatment naive upon the initial visit and fulfilled the reimbursement criteria set by Taiwan NHIB for subsidized ranibizumab were included.

Ranibizumab has been reimbursed for the treatment of nAMD by the Taiwan National Health Insurance Bureau (NHIB) since 2012, with a total reimbursement of 3 or 7 injections for each eye depending on the clinical response of the patient in his or her lifetime, after which, the treatment is paid by the patients themselves. In order to be reimbursed by the NHIB, patients must be at least 50 years old at the time of diagnosis, and had nAMD confirmed by fluorescein angiography, with subretinal fluid or macular edema demonstrated by optical coherence tomography (OCT). In addition, there should be no macular scars on the fovea, and the best correct decimal visual acuity (BCVA) should be ranging between 0.05 to 0.5. The waiting period for the permission of reimbursement of ranibizumab was around 4 weeks, thus some patients would have a self-paid bevacizumab injection during the waiting period.

All eyes included in this study received an initial loading dose of 3 monthly intravitreal injections of either ranibizumab and/or bevacizumab. Afterwards, the treatment was made on a loose treat and extend protocol based on funduscopic examination, OCT findings and BCVA. The loose treat and extend protocol is as follows: if under examination there is stabilization of macular edema and resolution of subretinal fluid by OCT, plus no retinal hemorrhage by biomicroscopy with 90D lens and a stabilized BCVA for 2 successive visits, the next treatment and follow up were extended by 1 month. If the examination showed any signs of recurrence and the deterioration of VA equal or more than 1 line, the treatment injection interval was then shortened by 1 month. Most, but not all, of our patients followed the above loose treat and extend strategy since some patients would insist on following up and treatment with less frequency in spite of disease activity. Generally speaking, the patients' follow up interval and dosing schedule are made to be equal to or less frequent than a bi-monthly basis after 1 year, even though there are signs of neovascular activity; either because of the economic burden, or the fear of the surgical procedures of intravitreal injection by the patients. Aflibercept was considered when the patient developed persistent subretinal fluid, or needed more frequent injections than a bi-monthly dosing. Patients had BCVA measured, along with funduscopic examination with 90 D lens and OCT imaging with Spectralis-OCT (Heidelberg Engineering, Heidelberg, Germany) at each visit.

Patients were divided into 2 groups based on whether aflibercept was later used as a rescue therapy. Group 1 received bevacizumab and ranibizumab only. Group 2 received bevacizumab, ranibizumab, and later aflibercept rescue therapy. Outcome measures include BCVA at 1 year, 2 years, 3 years and 4 years of follow up, and the average number of injections per year.

Statistical analysis The BCVA obtained during each visit was converted from decimal to logarithm of the minimum angle of resolution (logMar) for calculation. Chi-square test were used to Compare the baseline data including age, sex, eye. Mann-Whitney-U test was used to compare the difference of age, and average number of injections per year between the 2 groups. Mann-Whitney-U test, Wilcoxon test and independent t-test were used for the comparison of BCVA at different time points in each group. Patients were also divided according to their initial BCVA (>1.0 or <=1.0 in logMar), or age (>70 or <=70 years of age) to see whether initial BCVA or age play a role in the visual changes. A p-value of less than 0.05 was accepted as statistically significant. All the analyses were performed using SPSS software package (ver. 23.9; SPSS Inc., Chicago, illinois, USA).

Inclusion Criteria:

• Patients with nAMD that were treated with intravitreal injection of anti-VEGF and had a minimum follow up of 48 months

Exclusion Criteria:

• Loss follow over 1 year.

Retrospective case series of patients with nAMD that were treated with intravitreal injection of anti-VEGF and had a minimum follow up of 48 months. Every patient was initially treated with 3 loading doses of either bevacizumab or ranibizumab.