Predicting retinal pigment epithelium remodelling and its functional impact.

PURPOSE To identify predictive factors for RPE tear remodelling and its correlation with functional and morphological outcomes. METHODS Retrospective longitudinal study of patients with retinal pigment epithelium (RPE) tears secondary to age-related macular degeneration (AMD). Imaging was performed using spectral-domain optical coherence tomography (SD-OCT) and fundus autofluorescence (FAF). RPE layer integrity in the RPE-denuded area was examined with SD-OCT, and variation in the RPE-denuded homogeneous hypofluorescent area was examined with FAF over time for each case (eye). Patients were divided in two groups, according to the presence (Rem) or absence (No Rem) of evidence of RPE tear remodelling. Data were collected at three different time points: at baseline (at diagnosis of exudative AMD), at RPE tear diagnosis, and at the last available follow-up. Using SD-OCT, the following parameters were evaluated: type of CNV, type of PED and its dimensions, presence of subretinal (SRF) or intraretinal (IRF) fluid, central retinal thickness (CRT), presence and location of hyperreflective dots, and dimension and location of RPE tear. RESULTS This study included 32 eyes from 31 patients (19 female and 12 male), with RPE tears secondary to AMD. RPE remodelling after tear development was evident in 17 (53.1%) eyes after 7 [1-59] months. Anatomical recovery was associated with a younger age at RPE tear diagnosis (73 ± 7 vs. 81 ± 7 years old, p=0.01), smaller and narrower retinal pigment epithelial detachment (PED) at tear diagnosis (height 369 vs. 602 μm, p=0.02; width 2379 vs. 3378 μm, p=0.04), and the presence of SRF at tear diagnosis (94% vs. 53%, p=0.02). After adjusting for other covariates, a younger age at RPE tear diagnosis maintained significant association with RPE tear remodelling. RPE tear remodelling did not correlate with a better visual outcome at last follow-up (43 ± 22.8 vs. 34 ± 23.8 ETDRS letters, p=0.30). Final VA was directly proportional to VA at tear diagnosis (r= 0.654; p<0.001) and correlated negatively with PED width at tear diagnosis (r = -0.388; p=0.03). CONCLUSION RPE remodelling was evident in half of our sample and was associated with a younger age, smaller and narrower PED at RPE tear diagnosis, and presence of SRF also at tear diagnosis. Nevertheless, this structural recovery did not result in a better functional outcome.