Myopic choroidal neovascularization: To load or not to load

Abstract

There is far less known about how vascular endothelial growth factor inhibitors are best given for choroidal neovascularization (CNV) that occurs in eyes with myopia than in those with macular degeneration. Myopic CNV (mCNV) tends to occur in younger eyes with thinner chorioretinal complexes that are more likely to be Asian. These features may lead to treatment requirements that are different from those that have become established for neovascular agerelated macular degeneration (nAMD) in Caucasians and also to different outcomes. The SMILE study (Treatment and assessment Strategy for MyopIc CNV with LucEntis) by Li et al of Clinical and Experimental Ophthalmology provides useful evidence to address this gap in knowledge. Myopic CNV usually requires a much less intense treatment regimen than CNV in nAMD, which requires long-term injections that typically begin with an induction phase consisting of monthly injections, usually at least three, followed mostly by a less intensive maintenance phase. The REPAIR study administered ranibizumab using a “pragmatic treatment algorithm” after the first injection had been given for mCNV. Only one injection was enough in 21% of patients, 15% required three more for an overall mean gain of 14 logarithm of minimul angle of resolution (logMAR) letters for the entire cohort after 1 year of treatment. In addition to finding benefits of intravitreal ranibizumab over photodynamic therapy with verteporfin for mCNV, the RADIANCE study found similar improvements in mean visual acuity (VA) as REPAIR, with a median of around three injections after the first or the first and second injections 4 weeks apart using an “individualized treatment” regimen. The MYRROR study, a randomized, shamcontrolled, phase III clinical trial, treated eyes with mCNV with aflibercept or photodynamic therapy with verteporfin using a pro re nata (prn) approach, with monthly assessments after the first injection. A mean gain of around 14 letters over the first year was again found, with a median of two (range 1-3) injections of aflibercept in the first 3 months, decreasing to zero (0-3) injections in the following 9 months. There was no consensus prior to the study by Li et al on whether treatment of mCNV should start with three consecutive monthly loading doses or one single initial dose. A nonrandomized prospective trial reported that bevacizumab with one initial treatment, followed by prn, achieved the same visual and anatomical outcomes with fewer injections but had a higher rate of recurrences, similar to treatment with 3-monthly loading injections followed by prn. Kung et al found, in a retrospective study of 46 patients, that those who had just one loading dose of ranibizumab received an average of 2.3 injections over the first year of treatment, whereas those who started with three injections a month apart received 3.6, so the second group received more but had fewer recurrences. Both groups had similar VA increases. The SMILE study, a prospective single-blinded clinical trial, randomised 50 patients 1:1 to either one or three initial ranibizumab treatments followed by prn injections. The primary outcome of number of injections at 12 months was significantly lower in the 1+ prn group, which received 2.0 ± 1.2 injections, than in the 3+ prn group, with 3.6 ± 0.7 injections. Mean VA, a secondary outcome, improved after the first injection similarly for both groups and tended, if anything, to increase further until month 12. Similar proportions of eyes obtained ≥15 letters or lost 0 to 5 letters in each group, whereas nobody lost more than five letters. CNV thickness, CNV area and central retinal thickness (CRT) all decreased similarly in both groups. Only seven of the 26 (27%) patients in the 1+ prn group vs four of 24 (17%) (P = 0.38) patients in the 3+ prn experienced any recurrence, which was defined as occurring at least 3 months after a previous injection, over the entire follow-up period. These recurrence rates were not significantly different between the two groups, but Cox regression analysis found that the risk of recurrence was 3.6 times higher in patients over 55 years of age, 5.3 times higher if female, 4.4 times higher if CRT was >300 μm at presentation and 2.6 times higher if only 1+ loading dose was given. These characteristics are easily ascertainable and might be worth bearing in mind. The SMILE study found no serious systemic adverse events in this younger population as it was underpowered to do so, but one patient did develop a retinal detachment (from the 3+ prn group), highlighting the increased risks of intravitreal treatment in myopic eyes. The SMILE study suggests that eyes with mCNV treated with 1+ prn ranibizumab followed by prn dosing do just as well functionally and anatomically over 12 months as those treated with a 3-monthly loading dose. Fewer injections are DOI: 10.1111/ceo.13473