A prospective interventional case series study was conducted from October 2015 to November 2016 after approval of the Ethical Committee of the Faculty of Medicine, Assiut University, Egypt (264-15). All procedures were carried out under the tenets of the Helsinki Declaration. Written consent was provided by all participants after discussing the procedure, alternative treatment plans, follow-up schedules, and possible benefits and risks.

Patients suffering from previously diagnosed macular edema secondary to type 1 or 2 diabetes mellitus were included. For transition to aflibercept 2 mg/0.05 mL, a diagnosis of resistant DME was required. Patients fulfilling one or more of the following criteria were considered to have resistant DME after at least 3 consecutive monthly bevacizumab 1.25 mg or ranibizumab 0.5 mg injections in the previous 6 months: central macular thickness greater than 300 µm by spectral-domain optical coherence tomography (SD-OCT), reduction of retinal thickness by less than 10% of baseline retinal thickness, or suboptimal visual improvement (failure to gain at least 3 lines on the Snellen chart). Exclusion criteria were unwillingness to participate, significant cataract or corneal opacity, DME associated with proliferative diabetic retinopathy, history of laser treatment in the previous 6 months (either focal or macular grid laser photocoagulation), history of steroid injection in the previous 6 months (peri-ocular or intra-vitreal injection/implant), co-existing retinal pathology (e.g., retinal vascular occlusion, age-related macular degeneration), history of cataract surgery in the previous 12 months, associated optic nerve disorders (e.g., ischemic optic neuropathy), glycosylated hemoglobin higher than 8% at the time of participation, ischemic heart disease, or previously complicated intra-vitreal injection of anti-VEGF. Participants were excluded from the study if they had fewer than 3 consecutive aflibercept intra-vitreal injections.

All participants underwent thorough ophthalmic and systemic evaluation. Ophthalmic evaluation included detailed history-taking, best-corrected visual acuity (BCVA) by Snellen chart (converted subsequently to equivalent logarithm of the minimum angle of resolution [logMAR] values for statistical analysis), slit-lamp examination, and intraocular pressure (IOP) measurement before papillary dilatation (Goldmann applanation tonometer), followed by dilated fundus examination. Fundus fluorescein angiography (Topcon fundus camera; Topcon, Tokyo, Japan) and SD-OCT (Heidelberg Engineering, Heidelberg, Germany) were performed on every participant at initial evaluation.

In SD-OCT assessment, a 30 degree by 30 degree macular grid scan acquisition was carried out with 64 horizontal lines, spaced 50 µm and averaged to 30 frames. Early Treatment Diabetic Retinopathy Study macular thickness map was then measured automatically with Heidelberg software. Poor OCT images (signal strength value below 10) were excluded. Retinal thickness was measured as the perpendicular line between the internal limiting membrane and Bruch's membrane. Central 1 mm macular thickness was recorded with additional manual adjustment for measurement lines. Analysis and interpretation of OCT images were performed by an experienced ophthalmologist (ZE).

All participants received pre-injection prophylactic topical gatifloxacin 0.3% eye drops and topical anesthetic benoxinate 0.4% eye drops. After digital ocular massage to lower the IOP, the intra-vitreal injections were carried out under strict aseptic conditions in the operating room. The injection was performed in the temporal sclera 3.5 mm from the limbus after sweeping conjunctiva with a 5% Povidone Iodine-soaked sterile sponge. Assessment of pupillary response, detection of hand movement, and fundus examination (if needed) were carried out to exclude possible post-injection IOP spikes. Intra-vitreal injection of aflibercept was performed by a single experienced ophthalmologist (WI).

All patients received topical gatifloxacin 0.3% eye drops post-injection for 5 days. Follow-up visits were scheduled at 1, 2, and 3 months after the initial injection. BCVA, slitlamp examination, IOP measurement, and follow-up SD-OCT were performed at each injection.

DME edema was stratified into spongiform macular edema, cystoid macular edema with or without the presence of subretinal fluid, and/or vitreo-macular interface abnormality (VMI). Subretinal fluid was defined as accumulation of fluid between the neuro-sensory retina and retinal pigment epithelium. VMI abnormality was tagged by the presence of high reflective epi-retinal membrane (ERM), antero-posterior vitreo-macular traction, or thickened posterior hyaloid. Retinal morphological analysis was performed at baseline and at 3 months post-injection by SD-OCT and was interpreted by an experienced ophthalmologist (ZE).

The primary outcome measure was mean change in visual acuity after 3 consecutive monthly injections of aflibercept. Secondary outcome measures were mean change in retinal thickness, correlation between baseline visual acuity/baseline retinal thickness and final 3-month visual acuity/retinal thickness, and retinal architectural changes after switching to aflibercept.

Quantitative data were presented as mean ± standard deviation. For comparison between mean BCVA and CMT change pre- and post-aflibercept injection, we used paired sample t-test. Comparison of means among subgroups was carried out by Mann-Whitney test. We utilized Pearson correlation coefficient for correlation between mean post-injection BCVA and central macular thickness (CMT). Statistical significance was considered significant if p-value < 0.05. Statistical analysis was carried out using IBM SPSS Statistics ver. 20.0 (IBM Corp., Armonk, NY, USA).

An interventional case series study recruited 66 consecutive participants. Fourteen participants were excluded because they had received fewer than 3 consecutive intra-vitreal injections of a single anti-VEGF agent (10 patients) or 3 non-consecutive intra-vitreal injections (4 patients). Ten additional participants were excluded because they received only 2 intra-vitreal aflibercept injections and refused to have the third injection. A total of 42 patients (42 eyes) met the inclusion criteria and were included in the study. Mean age of the included patients was 60.04 ± 6.89 years (range, 49 to 71 years). Twenty-six males and 16 females were included in the study. Table 1 demonstrates patient characteristics and baseline assessment.

Mean pre-aflibercept injection CMT was 451.57 ± 107.09 µm, which decreased significantly at 1, 2, and 3 months after switching to aflibercept (346.52 ± 79.03, 328.24 ± 81.98, and 313.71 ± 85.79, respectively, p < 0.001) (Table 2). The patients were was subdivided into two groups based on baseline CMT: less than or equal to 450 µm and more than 450 µm. At the 3-month follow-up, mean reduction of CMT was 126.55 ± 34.71 and 150.30 ± 118.06 µm, respectively, in the two groups (p = 0.44, Mann-Whitney test).

At baseline, mean logMAR visual acuity was 0.87 ± 0.23, which improved significantly to 0.62 ± 0.29 after a single aflibercept 2.0 injection (p = 0.04). Visual acuity continued to improve significantly after the second (0.56 ± 0.34, p = 0.02) and the third aflibercept injections (0.46 ± 0.35, p = 0.03) (Table 3). Furthermore, mean logMAR visual acuity at 3-month follow-up was 0.25 ± 0.17 (equivalent to 6 / 12 on the Snellen chart) in patients with CMT less than o r e qual to 450 µm (22 e yes) a nd 0.68 ± 0.36 (equivalent to 6 / 30 on the Snellen chart) in patients with CMT more than 450 µm (20 eyes) (p = 0.03, Mann-Whitney test). On the other hand, patients who presented with logMAR visual acuity greater than or equal to 1.0 (14 eyes) had a final logMAR visual acuity of 0.8 ± 0.33 (equivalent to 6 / 38 on the Snellen chart), while patients who presented with logMAR visual acuity less than 1.0 (28 eyes) had a final logMAR visual acuity of 0.28 ± 0.21 (equivalent to 6 / 12 on the Snellen chart).

Both baseline visual acuity and retinal thickness had a strong and significant relationship to final logMAR visual acuity and retinal thickness at 3 months (Pearson correlation coefficient 0.64, p < 0.001 and 0.66, p < 0.001, respectively).

In SD-OCT follow-up scans, most of the eyes that presented with spongiform macular edema, intra-retinal cysts, and subretinal fluid at the time of shifting to aflibercept showed improvement by the third month (Fig. 1A, 1B). Progression of VMI abnormality was not found in any participant, and 1 eye developed ERM by the third month.

No serious systemic adverse events (e.g., cerebro-vascular stroke, myocardial infarction) were recorded during the study. Only four cases of subconjunctival hemorrhage were reported, with no other serious ocular adverse events (e.g., endophthalmitis, vitreous hemorrhage, retinal detachment).