Morphological Predictors of Short-Term Response to Intravitreal Bevacizumab in Macular Edema Due to Retinal Vein Occlusion

Abstract

INTRODUCTION: Our purpose was to identify morphological predictive factors of short-term macular functional and anatomical outcomes after monthly intravitreal bevacizumab for the treatment of macular edema (ME) due to central (CRVO) and branch retinal vein occlusion (BRVO).

METhODS: Retrospective study of patients with ME secondary to CRVO or BRVO under monthly treatment with intravitreal injections of bevacizumab. Only treatment naïve patients, with center-involved ME of ≥305 μm in women and ≥320 μm in men on baseline Spectral-domain OCT (SD-OCT) (Heidelberg Spectralis OCT; Heidelberg Engineering, Heidelberg, Germany) were included. Resolution of ME was defined as central macular thickness (CMT) less than 300 μm, no subretinal and no intraretinal fluid. Demographic and clinical parameters, best-corrected visual acuity (BCVA) in ETDRS logarithmic scale and SD-OCT images were reviewed at baseline and at 4 months. SD-OCT morphologic features in the central 1.0-mm diameter circle were checked for disorganization of the retinal inner layers (DRIL), ellipsoid zone (EZ) and external limiting membrane (ELM) disruption, presence and location of intraretinal hyperreflective foci (HRF) cysts, subretinal and intraretinal fluid, and vitreoretinal interface status.

RESULTS: The study enrolled 61 eyes of 61 patients, 29 (47.5%) with CRVO and 32 (52.5%) with BRVO. At 4 months, patients had received a mean number of 3.2±2.7 bevacizumab injections. Mean BCVA was 32±27 ETDRS letters at baseline and improved to 44±27 at 4 months (p<0.001). BCVA improvement was similar in CRVO and BRVO eyes (p=0.68). A greater BCVA improvement was correlated with a worse baseline value (r=-0.45, p<0.001). CMT reduced significantly from 592 ± 223 μm at baseline to 327 ±117 μm after loading dose (p<0.001) and twenty-three (37.7%) patients presented a complete resolution of ME at the 4th month timeline. The number of eyes with ME resolution were similar between those with CRVO and BRVO (p=0.590). The BCVA at the 4th-month follow-up was significantly lower in patients who presented at baseline with DRIL (39±27 vs 64±17 ETDRS letters, p=0.006), disrupted EZ (40±26 vs 64±21 ETDRS letters, p=0.016) and disrupted ELM (40±2 vs 64±20 6 ETDRS letters, p=0.016). Patients who presented DRIL at baseline have less 25.1 letters in BCVA at 4-months than patients who did not (95% confidence interval [CI] 8.1 – 42.3; p=0.004). Similarly, EZ and EML disruption predicted a decrease of 24.5 letters in final BCVA comparing to patients with integrity of these layers (EZ 95% CI 5.6 – 43.5; p=0.010 MLE 95% CI 5.6 – 43.5; p=0.010). None of the analyzed baseline morphological factors were predictive of ME resolution. However, absence of DRIL (p=0.003), presence of HRF in the inner retinal layers (p<0.001) and preserved EZ (p=0.030) and ELM (p=0.004) were significantly more frequent among those with ME resolution.

CONCLUSION: Intravitreal injection of bevacizumab for ME due to CRVO and BRVO resulted in a significant functional and anatomical improvement. In our study, patients with DRIL and disrupted EZ and ELM at baseline presented a significant lower BCVA at the end of the follow-up. Identification of baseline biomarkers for ME poor response to anti-VEGF will enable disease stratification and prognosis and improve treatment decisions.