Background: Due to its progressive nature, early evaluation and timely prediction of legal blindness are important in patients with neovascular age-related macular degeneration (nAMD). We examined the association between early-stage variation in central retinal thickness (CRT) and long-term visual outcomes in patients with nAMD.
Methods: We included 103 nAMD patients who were administered anti-vascular endothelial growth factor (anti-VGEF). Participants were considered qualified if they were: 1) 50 years and older, 2) treatment-naïve, 3) received standard anti-VEGF treatment and had complete baseline information. We further excluded patients with less than one-year follow-up data and those who experienced best-corrected visual acuity (BCVA) ≤35. Early-stage variability in CRT was measured as the first-year coefficient of variability (CV) of CRT. Patients were then classified into the high-variability and the low-variability groups according to the X-tile. A product-limit plot was used to demonstrate the differences and tested with the log-rank test. The association between first-year variability and visual outcomes was quantified using Cox regression models. Time-to-event primary endpoint was the overall visual preservation (OVP) rate, defined as the time from the first injection to legal blindness, i.e., BCVA ≤ 35 Early Treatment Diabetic Retinopathy Study (ETDRS) letters.
Results: A threshold of 20% of first-year CV in CRT was used to categorized 76 qualified patients into high-variability (35, 46.1%) and low-variability (41, 53.9%). The 5- and 10-year OVPs for patients with high- vs. low-variability were 76% vs. 48%, and 59% vs. 22%, respectively. High early-stage CRT variability showed a significantly higher risk of legal blindness. Even after adjusting for the demographic and clinical features, the variability remained significant (HR=2.39, 95% CI: 1.20 to 4.78).
Conclusions: First-year variability of CRT was predictive of long-term visual outcomes in patients with nAMD, and 20% of the variability could be used as a clinically convenient threshold to qualitatively classify patients into high- and low-variability groups. The current study is important for identifying high-risk populations and for long-term disease management.