To report and compare 10-year treatment outcomes of vascular endothelial growth factor (VEGF) inhibitors for neovascular age-related macular degeneration (nAMD) from Australia and New Zealand (ANZ) and Switzerland.
Retrospective, comparative, interventional case series.
We analyzed 712 treatment-naive eyes (ANZ, n = 474; Switzerland, n = 321) starting anti-VEGF for nAMD in routine clinical practice between January 1, 2006, and December 31, 2008, tracked in the prospectively designed observational database, the Fight Retinal Blindness! registry. The primary outcome was mean change in visual acuity (VA [in logMAR letters]) in eyes that completed 10 years of treatment.
The mean VA in 132 eyes (28%) from ANZ patients who completed 10 years of treatment dropped by 0.9 letters from baseline (95% confidence interval [CI], −4.9 to 3.1; P = 0.7) with 42% achieving ≥20/40, whereas the 37 eyes (12%) from Swiss subjects lost 14.9 letters (95% CI, −24 to −5.7; P < 0.001) with 35% achieving ≥20/40. Eyes from ANZ patients received more injections than eyes from Swiss subjects over 10 years (a median of 53 vs 42, respectively) from fewer visits with better disease control (proportion of visits with active disease: 38% vs 69%, respectively), suggesting a treat-and-extend regimen versus a pro re nata regimen (treatment given only when the lesion is active). Macular atrophy and subretinal fibrosis were the main reasons for 10 letter loss in the subset of eyes analyzed retrospectively. The mean VA of eyes from both regions that discontinued treatment within 10 years had fallen below the baseline at their final visit.
Eyes with nAMD may achieve satisfactory long-term visual outcomes if they receive adequate treatment. Central macular atrophy does not develop universally in eyes receiving long-term treatment with VEGF inhibitors as previously feared. Visual outcomes were better in eyes from ANZ patients, likely because they received more injections.
Visual outcomes at 10 years were reasonably good, with adequate treatment.
Disease control was better using the treat-and-extend regimen.
More patients with eyes with initially good vision completed 10 years of treatment.
Vascular endothelial growth factor inhibitors were well tolerated over 10 years.
The incidence of neovascular age-related macular degeneration (nAMD), a leading cause of blindness in people 50 years of age or older in the developed world, is likely to increase with the progressively ageing population. The pivotal phase 3 studies of intravitreal injections of vascular endothelial growth factor (VEGF) inhibitors that were completed over 10 years ago demonstrated remarkable clinical benefits for neovascular nAMD at 2 years. Many practitioners continue to treat eyes indefinitely, but there are limited data for the long-term outcomes.
Progression of the non-neovascular components of AMD such as development of macular atrophy may compromise long-term outcomes. Some studies have suggested that more intensive treatment regimens may increase the risk of developing macular atrophy. One of the few reports of 7-year outcomes of treatment of nAMD found macular atrophy had developed in over 90% of eyes. A larger report, which found approximately 40% of eyes had lost central vision due to macular atrophy after 7 years of treatment, suggested that under-treatment was a more significant cause of poor long-term outcomes. The development of subretinal fibrosis may also irreversibly damage vision.
The present study sought to describe 10-year outcomes of treatment with VEGF inhibitors in eyes with nAMD. Preliminary analysis revealed different treatment patterns and outcomes from 2 different regions, Australia and New Zealand (ANZ) and Switzerland, so their outcomes were analyzed separately.
Subjects and Methods
Data from a prospectively designed, internet-based treatment outcomes registry, the Fight Retinal Blindness! (FRB!) project were analyzed.
The FRB! project AMD module database tracked outcomes of treatment of eyes with nAMD. The data analyzed in this study were from participants who were treated for nAMD from practices in Australia, New Zealand, and Switzerland. Institutional ethics approval was obtained from the ethics committees of the University of Sydney, the Royal Australian and New Zealand College of Ophthalmologists, and the Cantonal Ethics Committee Zurich. Ethics committees in Australia and New Zealand approved the use of an “opt out” patient consent. The FRB! study conformed to the provisions of the Declaration of Helsinki.
Data Sources and Measurements
Details of the principles and design of the FRB! registry have been published previously. The number of letters read on a logarithm of the minimum angle of resolution (logMAR) VA chart (best of uncorrected, corrected, or pinhole), activity of the choroidal neovascular membrane, treatment given (if any), and ocular adverse events were recorded at each visit. Treatment decisions, including choice of treatment and visit schedules, were made at the discretion of the treating physician in consultation with the patient, thereby reflecting real world practice.
Participants and Variables
Treatment naïve eyes with nAMD starting treatment with intravitreal injections of VEGF inhibitors between January 1, 2006, and December 31, 2008, allowing a potential observation period of at least 10 years, were considered for the analysis.
The main outcome measurement was the mean change in VA from the initial treatment visit (baseline) to 10 years in patients who were followed for at least 10 years. The proportion of eyes with VA of ≥70 letters (20/40 Snellen equivalent, driving vision) and ≤ 35 letters (20/200, legally blind) at 10 years and those that gained ≥10 letters or lost ≥10 letters during the same period were also evaluated.
Other prospectively defined secondary outcomes were the number of injections and the number of visits over the 10-year period. The VA outcome of eyes with good initial vision (≥70 letters [20/40]) and eyes with poor initial vision (≤35 letters [20/200]) were compared with those eyes with a baseline VA of 36 to 69 letters. The VA trends for eyes of patients who were lost during the 10-year follow-up period also were analyzed. These outcomes were assessed in patients from 2 regions, ANZ and Switzerland. The cause of vision loss of ≥10 letters in eyes that received 10 years of treatment was assessed by a selection of practitioners who contributed the most patients based on clinical notes, optical coherence tomography images, which were usually obtained at each visit, as well as fundus autofluorescence images where available.
Descriptive data included mean ± SD, median (first and third quartiles, Q1 and Q3), and percentage where appropriate. Change in VA from baseline was analyzed for each year in eyes of patients who completed follow-up visits (completers), in eyes of patients who did not complete follow-up visits (noncompleters), and in all eyes. Noncompleters were grouped according to the year of noncompletion. Outcomes were also investigated in eyes divided by their baseline visual acuity: ≥70 letters (20/40) and ≤35 letters (20/200). Paired t -tests were used to determine whether changes in VA at 10 years from baseline were significant. Time to study dropout was analyzed using Kaplan-Meier survival curves.
A longitudinal generalized additive model, including visits from all the eyes regardless of whether they completed 10 years of observations or not, was used for predicting VA over 10 years. This model was used to plot VA change from the baseline to 10 years. The longitudinal VA over 10 years was also compared between ANZ and Swiss patients by using these models. A P value of 0.05 was considered statistically significant.
All analyses were conducted using R version 3.4.4 software (Vienna, Austria; www.R-project.org/ ), using the MGCV application (version 1.8-24) for generalized additive (mixed) model computation and the SURVIVAL application (version 2.38) for dropout analysis.
A total of 795 treatment-naïve eyes of 690 patients starting treatment with a VEGF inhibitor between January1, 2006, and December 31, 2008, were identified in the FRB! project registry. The mean age of the patients at their first eye injection (“baseline”) was 79.2 ± 7.4 years old, and the mean presenting VA was 53.8 ± 17.9 letters (20/80 Snellen equivalent). Table 1 shows the demographic characteristics of the eyes from the 2 regions. More than half of the eyes (58%) received treatment as monotherapy for the entire period: 411 eyes (52%) with ranibizumab and 51 eyes (6%) with bevacizumab. A total of 187 eyes (24%) exchanged between ranibizumab and bevacizumab, 1 eye (0.12%) between aflibercept and bevacizumab, 87 eyes (11%) between ranibizumab and aflibercept, and 58 eyes (7%) between all 3 anti-VEGF agents for the treatment of nAMD during the period.
|Australia and New Zealand||Switzerland|
|Number of eyes||474||321|
|Number of patients||417||273|
|Mean ± SD age, y||79.1 ± 7.7||79.3 ± 6.9|
|Female patients||247 (59.2)||189 (69.2)|
|Right eye||232 (48.9)||163 (50.8)|
|Mean ± SD baseline VA letters (logMAR) (Snellen equivalent)||54.4 ± 19.3 (20/80)||52.9 ± 15.5 (20/80)|
|Lesion type, % a|
Outcomes at 10 Years
Of the 795 eyes identified, 169 eyes (21%) of 150 patients completed 10 years of continuous treatment. The mean VA of these 169 eyes at baseline was 60.9 ± 16.4 letters (20/60 Snellen equivalent). The mean VA in eyes from ANZ had dropped by 0.9 letters at 10 years ( P = 0.7), although there was a drop of 14.9 letters in eyes from Switzerland ( P < 0.001) ( Table 2 ). The proportion of eyes with VA of ≥70 letters (20/40) improved at 10 years in both regions ( Table 2 ). The proportion of eyes with VA of ≤35 letters (20/200) increased in both regions at 10 years ( Table 2 ).
|Outcomes Stratified by Regions|
|Australia and New Zealand||Switzerland|
|Completers||132 (28)||37 (12)|
|Mean ± SD baseline VA letters (logMAR) (Snellen equivalent)||60.7 ± 17 (20/60)||61.6 ± 14 (20/60)|
|Mean ± SD final VA letters (logMAR) (Snellen equivalent)||60.1 ± 20.7 (20/60)||46.8 ± 28.8 (20/125)|
|Mean change in VA letters (logMAR) (95% CI)||−0.9 (95% CI: −4.9 to 3.1)||−14.9 (95% CI: −24 to −5.7)|
|Proportion with ≥10-letter gain||34%||19%|
|Proportion with ≥10-letter loss||27%||49%|
|VA ≥70 letters (20/40 Snellen equivalent), baseline/final||36%/42%||24%/35%|
|VA ≤35 letters (20/200 Snellen equivalent), baseline/final||7%/14%||5%/38%|
|Proportion of active visits||38%||69%|
|Median injections (Q1, Q3)||53 (35, 69)||42 (17, 71)|
|Median visits (Q1, Q3)||66 (48, 80)||78 (60, 91)|