Occurrence, progression and risk factors of macular atrophy in neovascular age‐related macular degeneration according to neovascularization subtypes
Amandine Moulin, Geoffrey Zeender, Pauline Rizzuto, Pierre Gascon, Kevin Allaix, Audrey Feldman, Flore De Bats, Benjamin Wolff, Pierre Pradat, Philippe Desmangles, Laurent Kodjikian, Thibaud MathisAbstract
Objective
To assess the occurrence, risk factors, and progression of macular atrophy (MA) according to macular neovascularization (MNV) subtype in neovascular AMD (nAMD).
Methods
This multicentre retrospective observational cohort study included 416 eyes with nAMD, followed for at least 12 months and treated with anti‐VEGF. MA was measured annually on near‐infrared imaging. Occurrence and risk factors were analysed by MNV subtype using Kaplan–Meier survival and Cox regression models. Atrophy enlargement rate was assessed in mm 2 /year and after applying square root transformation. MNV subtypes were divided into subepithelial (Type 1), intraretinal (Type 3) and MNV with subretinal component (Type 2 and mixed Type 1+2).
Results
Over a mean follow‐up of 55.4 ± 22.9 months, 47.4% of eyes developed or presented with MA, with the highest prevalence in Type 3 MNV (82%). Kaplan–Meier analysis confirmed a significantly greater risk in Type 3 lesions ( p < 0.0001). Independent risk factors included Type 3 MNV, subretinal drusenoid deposits, thinner choroid and fewer visits with subretinal fluid. After multivariable adjustment, the cumulative number of anti‐VEGF injections was not significantly associated with MA occurrence. The mean MA enlargement rate was 1.0 ± 1.1 mm 2 /year overall, ranging from 0.9 ± 1.1 mm 2 /year in Types 1 to 1.5 ± 1.1 mm 2 /year in MNV with subretinal component, p = 0.002.
Conclusions
Type 3 MNV carries the highest risk of MA, underscoring the need for careful monitoring, whereas MA progression was faster in lesions with a subretinal component. No independent association between cumulative anti‐VEGF injection number and MA occurrence was observed; however, this finding should be interpreted cautiously given the potential for confounding by disease activity, MNV subtype, treatment regimen and follow‐up intensity.