Abstract
Purpose :
To describe the relationship between fluctuations in central foveal thickness (CFT) and vision outcomes in the HARBOR trial of ranibizumab for neovascular age-related macular degeneration.
Methods :
In HARBOR (NCT00891735), treatment-naïve patients were randomized to monthly or as-needed (PRN) intravitreal ranibizumab (0.5 mg or 2.0 mg) for 24 months. Post hoc analyses included patients with ≥ 16 evaluable OCT images from month 3 (after 3 monthly loading doses), and best-corrected visual acuity (BCVA) data at baseline and month 24. After month 3, CFT data were grouped into 3-month intervals through month 24, and “bounces” were counted for every ≥ 50-µm difference between the minimum and maximum CFT value within an interval, and between 2 consecutive intervals. Associations between bounces in CFT and vision outcomes at month 24 were compared between patients randomized to monthly versus PRN ranibizumab (0.5 mg and 2.0 mg arms pooled).
Results :
Analyses included 849 patients randomized to monthly (n = 427) or PRN (n = 422) ranibizumab. By design, PRN-treated patients exhibited more bounces in CFT from months 3 through 24 than those treated monthly (Figure 1; mean [95% CI], 8.5 [8.0, 9.0] vs 3.4 [2.9, 3.9]). After adjusting for baseline BCVA, mean (95% CI) vision gains from baseline at month 24 were similar between monthly- and PRN-treated patients with 0–3 bounces in CFT (+10.4 [8.6, 12.2] letters vs +10.3 [7.7, 12.9] letters). In patients with 4–21 bounces in CFT from months 3 through 24, adjusted mean (95% CI) BCVA gains in those who received monthly or PRN treatment decreased to +9.5 (7.1, 12.0) letters and +7.4 (5.7, 9.1) letters, respectively. When patients were stratified by bounce quartile, adjusted mean vision gains were smallest among the monthly and PRN subgroups with 11–21 bounces in CFT (Figure 2).
Conclusions :
This post hoc analysis demonstrated a trend for numerically smaller vision gains among patients who exhibited greater fluctuations in CFT during the HARBOR trial. Further studies should assess whether early bounces in CFT may predict longer-term treatment responses, and may be used to optimize injection intervals and vision outcomes.
This is a 2020 ARVO Annual Meeting abstract.