FACTORS INFLUENCING NEED FOR RETREATMENT AND LONG-TERM VISUAL OUTCOME AFTER INTRAVITREAL BEVACIZUMAB FOR MYOPIC...

To evaluate the efficacy and the predictive factors associated with the need for retreatment and long-term visual outcome after intravitreal bevacizumab for myopic choroidal neovascularization (CNV).

Retrospective cohort study of 93 eyes with subfoveal or juxtafoveal myopic CNV treated initially with either 3-monthly or single intravitreal bevacizumab injections followed by pro re nata retreatment. The efficacy was evaluated by the best-corrected visual acuity (BCVA) during follow-up visits. Backward stepwise multiple linear regression analyses were performed to evaluate the potential predictive factors on final BCVA, change in BCVA, and number of injections. Multiple logistic regression was performed to evaluate the potential predictive factors for retreatment.

The mean follow-up duration was 25.12 ± 11.18 (SD) months. The mean logMAR BCVA at baseline was 0.72 ± 0.58 logMAR (20/100 Snellen equivalent) and was maintained at 0.39 ± 0.46 logMAR (20/50 Snellen equivalent) at the last follow-up (P < 0.001). The mean number of injections was 3.53 ± 1.70 (range, 3-10), and a total of 25 eyes (26.9%) received retreatment. Patients who received single loading injection had significantly lower mean total number of injections (1.50 ± 0.73 vs. 3.96 ± 1.53). Both subfoveal and juxtafoveal myopic CNV eyes had significant improvement in BCVA (0.28 ± 0.43 vs. 0.22 ± 0.32 [20/40 vs 20/30 Snellen equivalent], P = 0.506), and juxtafoveal myopic CNV eyes had significantly better BCVA at baseline and at the last follow-up than the subfoveal group. Treatment-naive eyes had significant improvement from baseline BCVA, and the amount of improvement was significantly more than those who received previous photodynamic therapy (0.31 ± 0.43 vs. 0.06 ± 0.11 [20/40 vs 20/25 Snellen equivalent], P < 0.001). Multivariate stepwise regression analysis showed that the baseline CNV size (P < 0.05), baseline BCVA (P < 0.001), and duration of symptoms (P < 0.05) were significant predictive factors for final BCVA, and BCVA improvement. Multiple logistic regression analysis identified that CNV size (P = 0.014) and follow-up duration (P = 0.017) were significant predictive factors for retreatment. No significant association was found for number of injections.

Intravitreal bevacizumab seems to be an effective treatment for both subfoveal and juxtafoveal myopic CNV in the long term. Patients presented with shorter duration of symptoms and smaller CNV size before treatment as significant prognostic factors that predict better visual outcome. Eyes with longer follow-up duration and larger baseline CNV size may have higher risk for retreatment.