Retinal surgeons and clinical investigators have long been impressed with the results of using anti-vascular endothelial growth factor in various retinal diseases; now they’re analyzing it more closely to see how it works over the long term and when it’s combined with other therapies. At this year’s ARVO, research groups will share their results from these anti-VEGF studies, as well as provide more insight into retinal surgery outcomes.

AMD Investigations

Investigators from the Regeneron/Bayer-sponsored VIEW studies have performed a long-term analysis of retreatment with aflibercept and ranibizumab for neovascular age-related macular degeneration.

In the studies, the researchers randomized 2,412 patients with wet AMD to monthly ranibizumab 0.5 mg (Rq4), monthly intravitreal aflibercept injection 2 mg (2q4), monthly IAI 0.5 mg (0.5q4), or IAI 2 mg every other month (2q8) following three initial monthly doses. They evaluated the primary outcomes at week 52. Between weeks 52 and 96, the physicians administered injections at 12-week intervals, but they could be given as frequently as every four weeks for any of the following: increased central retinal thickness ≥100 μm compared to the lowest previous value; loss of five or more ETDRS letters from best previous score with recurrent fluid; new onset classic neovascularization or hemorrhage or new or persistent fluid on optical coherence tomography; or leak on fluorescein.

Between weeks 52 and 96, all of the patients received an average of 4.1 to 4.7 injections, and there was a small overall trend of acuity loss at week 96. The researchers say that subgroup analysis found that the slight loss was mainly caused by approximately a fifth of patients in all groups who lost at least five letters between weeks 52 and 96 in spite of p.r.n. retreatment. The physicians report that this group was stable from baseline until week 52, but lost on average more than 11 letters with either drug when switched to reactive treatment, and that CRT didn’t change. The subgroup received a number of injections similar to the full study population.

The investigators performed a separate subgroup analysis on patients who lost five or more letters between two consecutive treatments and subsequently received reactive treatment for weeks 52 through 64. At 96 weeks, acuity in these patients decreased from gains at Week 52 (+8.5 to 10.3 letters), to a VA of -2.5 to -3.8 letters below baseline, with no obvious changes in CRT. They add that frequent ocular adverse events were conjunctival hemorrhage, eye pain, retinal hemorrhage and reduced VA.

The researchers say that there seems to be a subset of patients in whom reactive treatment won’t recover vision that’s lost, and that proactive treatment results in better vision.3171

The researchers from the Comparison of Age-related Macular Degeneration Treatment Trials have done an analysis of the results to determine risk factors for developing geographic atrophy. Two of the researchers receive financial support from, or are consultants to, makers of anti-vascular endothelial growth factor drugs.

Selected Risk Factors for Geographic Atrophy from CATT3658

Number of subjects
Percent with geographic atrophy in two years
Adjusted hazard ratio (95 percent CI)
Baseline VA in study eye
1.7 (1.2, 2.5)
1.7 (1.1, 2.6)
2.7 (1.4, 4.9)
 RAP Lesion
1.7 (1.2, 2.4)
Intraretinal Fluid
Fluid, not in foveal center
1.8 (1.1, 3.0)
Fluid, in foveal center
2.1 (1.3, 3.3)
The researchers looked at the 1,020 CATT patients who had no GA visible on digital color photos or FA at the time of enrollment. The patients had been randomly assigned to treatment with ranibizumab 0.5 mg or bevacizumab 1.25 mg and to a two-year-long regimen of either monthly or p.r.n. dosing, or monthly injections for a year and p.r.n. the following year.

The investigators report that GA developed in 187 (18.3 percent) of the patients by two years. In multivariate analysis, poor baseline visual acuity, presence of retinal angiomatous proliferation, absence of blocked fluorescence on FA, presence of GA in the fellow eye, ranibizumab treatment and a monthly treatment regimen, thinner subretinal fluid, decreased sub-RPE height and intraretinal fluid in the foveal center were independently associated with increased risk of GA. The researchers also looked at possible genetic connections in 770 patients who took part in the CATT genetic study. They found that the ARMS2 and HTRA1 risk alleles were associated with an increased incidence of GA.3658

Researchers from the European-based Inhibit VEGF in Age-related choroidal Neovascularization study have performed an analysis of the relationship between hemorrhage and OCT signs and the likelihood of neovascular AMD being active during follow-up. Some of the doctors receive financial support from, or are consultants to, Bayer and Novartis.

In the analysis, the Network of Reading Centres UK prospectively graded image sets from the IVAN study for presence or absence of hemorrhage and intraretinal fluid cysts, neuroretinal foveal thickness and height of subretinal fluid at the fovea. FA was graded for the presence or absence of leakage, which was the reference standard for activity. Feature discrimination between presence or absence of activity was quantified by the area under the receiver operating characteristic curve.

The researchers say that at 12 months there was leakage in 41 percent of the patients (183/449 angiograms), and at 24 months 38 percent (168/436 angiograms) had leakage activity. At 12 months, hemorrhages were present in 11 percent of the patients and intraretinal fluid cysts were present in 36 percent. At two years, 8 percent showed hemorrhage and 35 percent had evidence of cysts. The median neuroretinal foveal thickness height was 150 µm at 12 months and 140 µm at two years. The investigators say that 14 percent of OCTs at 12 months and 15 percent at two years showed subretinal fluid, and that the median height of SRF, when present, was 70 µm at both follow-up points. The positive/negative likelihood ratios for the features (the number of times more likely a positive test comes from someone with the disease vs. someone without it/the number of times more likely a negative test comes from someone with the disease vs. someone without it) were as follows:

 • hemorrhage: 4.38/0.84 at 12 months, 5.51/0.87 at 24 months;

 • fluid cysts: 1.72/0.72 at 12 months, 1.25/0.88 at 24 months; and

 • any SRF: 5.47/0.76 at 12 months, 8.1/0.71 at 24 months.

The areas under the receiver operating curve were ≤ 0.7 for neuroretinal foveal thickness and SRF, or a combination of features.

The investigators say that none of the features they studied does a good job of diagnosing FA-classified active disease well. They say that though hemorrhage and any SRF have high specificity, and can help rule in the presence of activity, most of the FAs classified with active disease didn’t have those features. As a result, they think that physicians should consider adopting OCT-guided treatment for neovascular AMD.3660

Researchers have completed a proof-of-concept study of Acucela’s orally administered emixustat HCl for the treatment of geographic atrophy associated with non-exudative AMD. Most of the researchers are either consultants for Acucela or receive financial support from the company.

Emixustat's Suppression of ERG b-wave Rod Recover4506
2 mg a.m. dose
5 mg a.m. dose
5 mg p.m. dose
7 mg a.m. dose
10 mg a.m. dose
Rate of recovery (slope)
Percent suppression

By demonstrating the ability to suppress ERG b-wave rod recover, researchers say emixustat's mechanism of action works as intended.

Emixustat HCl is a rod visual cycle modulator that inhibits isomerase (RPE65) activity and reduces retinal toxins that damage the retinal pigment epithelium and the overlying photoreceptors. In the prospective, double-masked, Phase II study, researchers randomized 72 patients to receive 2-, 5-, 7-, or 10-mg emixustat HCl to be given before noon (a.m. dosing) or 5 mg to be given after noon (p.m. dosing); or oral placebo daily for up to 90 days in a 3:1 ratio. To assess the drug, after 30 minutes of dark adaptation, researchers recorded electroretinograms. They then photo-bleached the eyes and recorded the ERGs immediately, then again at 10, 20 and 30 minutes. Rod b-wave amplitudes were expressed as a percentage of the prebleach dark-adapted rod amplitude from baseline. The investigators compared the rate of rod recovery (slope) for each emixustat HCl group to the placebo group.

The researchers say that the drug suppressed ERG b-wave rod recovery after light exposure in a dose-related fashion. Suppression plateaued by day 14 of dosing, and reversed within seven to 14 days after cessation. There were no systemic adverse events, and chromatopsia and delayed dark adaptation were the most common ocular AEs. Two subjects had a treatment-related serious AE, moderate chromatopsia at the 5-mg dose level. The study doctors say that all ocular AEs resolved upon drug cessation, and that the common events were explainable based on the drug’s mechanism. A long-term, Phase II/III study is under way.4506

The AMD Gene Exome Chip Consortium, sponsored by the National Institutes of Health, is gradually putting together some of the pieces of the genetic puzzle behind AMD, and will share results of their first round of analysis at this year’s meeting.

Researchers from the United States and Germany designed a custom genotyping array consisting of 250,000 common and rare coding variants discovered in large-scale sequencing experiments (and enriched for variants discovered in sequencing experiments targeting patients with AMD), and an additional set of 250,000 common variants distributed evenly across the genome. Then, working with the NIH Health Center for Inherited Disease Research, they began to genotype all available samples using this custom array. They ultimately curated and organized DNA samples for more than 48,000 people, consisting of 48.25 percent AMD cases with the rest acting as non-AMD controls.

The researchers say that, among the AMD cases, 46.25 percent had CNV, 14.7 percent had GA and 9 percent had GA in one eye and neovascular disease in the other (resulting in 70.2 percent of cases having advanced disease). In the remaining cases, 16.5 percent had large drusen and 13.3 percent had earlier signs of disease. The researchers say that they expect an 80-percent power level in detecting variants with a frequency of >0.1 percent that lead to greater than a 2.45-fold increase in disease risk.4977

Researchers in the Age-Related Eye Disease Study 2 say that visual acuities in AMD patients improve significantly after cataract surgery, despite the retinal issues.

AREDS2 is a five-year, multicenter, randomized, controlled trial of nutritional supplements for AMD that has enrolled 4,203 patients with different degrees of AMD. In assessing secondary outcomes, researchers analyzed pre- and postoperative characteristics of participants who underwent cataract extraction during the five-year trial, using both clinical data and standardized lens and fundus photographs obtained at baseline and yearly thereafter. A centralized reading center graded photographs for lens opacities and severity of AMD.

     Researchers say that a new approach to OCT called swept-source may enable enhanced vitreal imaging, and reveal more detail than current methods. One of the investigators holds a patent related to the technology.
     In the study of 22 normal eyes, EVI enabled researchers to see features such as vitreous separation from the retina and hyaloid detachment near the optic nerve head. They say that SS-OCT's main advantages are its high sensitivity over a long imaging range, and that its high speed combined with its motion correction and image merging enables wide-field volumetric imaging. Appearing above are a wide-field OCT (A), a standard OCT image (B) and EVI (C). EVI shows a posterior precortical vitreous pocket (blue asterisk), hyaloid adhesion at the optic nerve (red arrow), space of Martegiani (yellow asterisk) and vitreous separation (green arrow).3167

In this particular subset of the study, the researchers analyzed visual acuity results for 1,232 eyes (793 patients) that had cataract surgery during AREDS2. After adjustment for age at surgery, gender, and type and severity of cataract, the mean changes in visual acuity were as follows: Eyes with mild AMD (AREDS AMD Scale 1 to 3, n: 31) gained 10.7 letters (p<0.0001) compared to preoperative acuity; eyes with moderate AMD (AAS 4 to 6, n: 346) gained 11.2 letters (p<0.0001); eyes with moderate AMD (AAS 7 to 8, n: 464) gained 8.8 letters (p<0.0001); eyes with non-central geographic atrophy (AAS 9, n: 70) gained nine letters (p<0.0001); and eyes with advanced AMD (AAS 10 to 11, n: 321) gained seven letters (p<0.0001). The mean time from measurement of preoperative acuity to cataract surgery was 5.9 ±3.6 months, and the mean time from surgery to measurement of postoperative acuity was 7.0 ±3.6 months.243

Ophthotech has released the Phase IIb data from its study of its anti-platelet derived growth factor drug Fovista, in combination with ranibizumab, compared to ranibizumab monotherapy in neovascular AMD.

In the prospective, controlled study, a physician randomized 449 patients with wet AMD to receive one of the following every four weeks for 24 weeks: Fovista 0.3 mg in combination with ranibizumab 0.5 mg; Fovista 1.5 mg in combination with ranibizumab 0.5 mg; or sham in combination with ranibizumab.

The researcher reports that the combination using Fovista 1.5 mg met the primary endpoint of superiority in mean visual acuity gain compared to ranibizumab alone (10.6 ETDRS letters gained at week 24 vs. 6.5 letters, p=0.019). There was an additional 62 percent benefit from baseline in the Fovista 1.5 mg combination therapy arm over ranibizumab monotherapy, and there was a classic dose-response curve. The researcher notes that the relative magnitude of visual benefit increased over time, and that the superiority of Fovista 1.5 mg combination therapy over ranibizumab monotherapy was consistent across all subgroups, including those based on baseline vision, lesion size and central retinal thickness. Fovista 1.5 mg combination was superior to ranibizumab monotherapy across multiple treatment endpoints including three, four and five lines of vision gain. OCT and fluorescein angiography analysis showed patients receiving Fovista 1.5 mg combination therapy had greater reduction in neovascular size compared to those receiving ranibizumab monotherapy. No significant safety issues were observed for either treatment group in the trial.2175

Diabetic Retinopathy

Researchers from the Genentech-sponsored RISE and RIDE trials analyzed the trial data to determine the effect of intravitreal ranibizumab on diabetic retinopathy severity.

To recap the trials’ structure, 759 patients with diabetic macular edema were randomized in a 1:1:1 ratio to receive monthly 0.3-mg or 0.5-mg ranibizumab or sham injections. The sham patients were able to receive 0.5-mg ranibizumab during the third year. Everyone was eligible for macular laser starting at month three, and panretinal photocoagulation was also available.

Delaying ranibizumab reinjections may stunt their effect in diabetic retinopathy, say researchers.4028
The researchers report that a higher percentage of eyes in the ranibizumab arms had a regression of their DR of at least two or three steps on the
ETDRS severity scale compared to the sham/0.5-mg crossover group at three years. Fifteen percent of the 0.3-mg eyes and 13.2 percent of the 0.5-mg patients experienced three steps or more of improvement, compared to just 3.3 percent of the sham/0.5-mg crossover patients. Through 36 months, 33.9 percent of eyes originally randomized to sham developed proliferative DR as measured by a composite outcome that included photographic changes and clinically important events such as occurrence of vitreous hemorrhage or application of PRP. Only 12.8 percent of the 0.3-mg group and 15.1 percent of the 0.5-mg group progressed to PDR. However, in the third year, the investigators say that the proportions of patients exhibiting DR progression were similar in all three treatment arms.

The physicians say that the data provides strong evidence that ranibizumab is effective in reducing DR severity and inhibiting its progression to PDR, but add that delays in starting ranibizumab therapy may stunt this therapeutic effect.4028

Researchers from a large-scale, randomized study supported by Novartis Pharmaceuticals Canada share their results from treating DME with ranibizumab alone and with laser.

The researchers randomized 241 patients in a 1:1:1 ratio; 78 received a combination of ranibizumab and laser, 81 received ranibizumab alone and 82 received only laser. The protocol consisted of administering three consecutive monthly injections to patients, then following them for 10 months during which further injections could be given based on retreatment criteria. For patients receiving it, laser was administered according to ETDRS guidelines at intervals no shorter than three months. The mean best-corrected acuity and central retinal thickness for the intention-to-treat group (n: 212) were 63.7 ±10 letters and 437 ±141 µm, respectively.

The investigators say the mean change in BCVA and CRT from baseline to a year was +8 letters (95-percent CI: 5.5, 10.5) and -144 µm (95-percent CI: -182, -106), respectively, for the combination arm (n: 60); +8.3 letters (95-percent CI: 6.5, 10.1) and -135 µm (95-percent CI: -170, -100), respectively, for the ranibizumab arm (n: 62) and +1.1 letters (95-percent CI: -2.5, 4.6; n: 46) and -112 µm (95-percent CI: -157, -68; n: 45), respectively, for the laser arm. During the trial, the mean number of injections received was 8.5 ±3 for the combination arm (n: 71) and 8.6 for the ranibizumab arm (n: 74). The average number of laser sessions was 1.6 ±1 for the combination arm (n: 69) and 2.5 ±1 sessions for laser arm (n: 69). The most common reason for discontinuation was an unsatisfactory effect in 11 percent of the laser arm (n: 82), 3 percent of the combination arm (n: 78) and 1 percent of the ranibizumab arm (n: 81).4025

A group of researchers from France, one of whom receives financial support from Allergan, has performed a retrospective, multicenter study of the intravitreal dexamethasone implant dubbed the Multicenter Ozurdex Assessment foR diabeTic macular edema, or MOZART.

The researchers placed the implant in 69 eyes of 59 patients with DME and followed them for at least six months (mean: 9.8 months). The patients’ mean age was 65 years. The mean systolic blood pressure was 138 mmHg and the mean HbA1c was 7.2 percent. Seventeen patients (24 percent) were naive to any macular treatment.

At baseline, the patients’ mean central retinal thickness was 540 µm. Post­­­­op, the average CRT decrease was: 188 µm at month one (M1);
235 µm at month two (M2); 117 µm at month four (M4); and 77 µm at month six (M6). The initial BCVA letter score was 54.4 letters, and the mean BCVA improvement postop was: 2.1 letters at M1; 5.4 at M2; 2.4 at M4; and 1.6 at M6. For treatment-naive patients this gain was higher: 5.8 letters at M1; 6.7 at M2; 8.7 at M4 and 6.7 at M6. During the follow-up, 28 percent of patients had a BCVA better than 20/40 (73 letters), vs. only 6 percent at baseline.

Researchers noted a gain in BCVA greater than 15 letters in 28 percent of patients, while a loss greater than 15 letters occurred in 6 percent. The mean number of implant injections was 1.2 at six months, with an average of 4.9 months before reinjection. The mean initial intraocular pressure was 15 mmHg. Ocular hypertension greater than 25 mmHg, managed by topical treatment, was observed in 7 percent of patients. Cataract progressed in 4 percent of patients, vitreous hemorrhage occurred in 1.5 percent and there were no cases of endophthalmitis.

Overall, the investigators say that the implant showed anatomical and functional effectiveness, but note that patient follow-up must be adapted to the implant’s duration of action, with a visit before the second month to detect pressure spikes and one before the fourth month to detect DME recurrence.2387

Surgeons in the RELATION study of ranibizumab and/or laser for DME report good results from combining the treatments.

RELATION is a multicenter, 12-month, two-armed, double-masked, parallel-group, active-controlled clinical trial in which patients with visual impairment due to DME were randomized 2:1 to ranibizumab in combination with focal/grid laser photocoagulation (combined group) or focal/grid laser photocoagulation combined with sham injections (laser group). After initial treatment starting at baseline with laser and four monthly ranibizumab/sham injections, treatment was given as needed. Also, a subgroup of patients with concomitant PDR was included, and this group received additional panretinal laser photocoagulation at baseline, then treatment as randomized.

Out of 128 patients, the investigators randomized 85 to the combined group and 43 to the laser group. They note that BCVA in the combined group was significantly better than BCVA in the laser group at final follow-up (mean change from baseline +6.5 letters for combined vs. +1.4 letters for laser, p=0.001). Eighty percent of centers used SD-OCT instruments, and reported that the reduction of total retinal volume (within a 6-mm3 ETDRS grid) on OCT was significantly higher in the combination group than in the laser group (mean change from baseline was -1.174 mm3 for the combination therapy vs. -0.501 mm3 for the laser group).

For the subgroup analysis, 27 patients had PDR at baseline (20 in the combined and seven in the laser group). Presence of PDR had no significant effect on BCVA outcome. Eight patients (40 percent) with PDR in the combined group, but none of the PDR patients in the laser group, showed regression of PDR during follow-up. The researchers say that the adverse event profile was similar to previous studies in NPDR and PDR patients.1239

Surgical Studies

In a study sponsored by ThromboGenics, researchers say that the company’s new drug ocriplasmin was effective in patients with symptomatic vitreomacular adhesion who would usually be candidates for vitrectomy.

The study included symptomatic patients with OCT-confirmed VMA who were randomized to receive a single intravitreal injection of 125 µg ocriplasmin (n: 464) or placebo (n: 188). Clinical criteria for consideration for vitrectomy was visual acuity of 65 ETDRS letters (20/50) or less (n: 301) or full-thickness macular hole (equivalent to stage II) at baseline (n: 153). A total of 127 patients met both criteria. In patients without baseline FTMH, the investigators evaluated the rate of VMA resolution at 28 days post-injection. In patients with baseline FTMH, they evaluated VMA resolution and FTMH closure at day 28.

The researchers report that pharmacologic VMA resolution at day 28 was observed in a significantly larger proportion of eyes in the ocriplasmin group compared to placebo. Among patients with a baseline VA of 65 letters or fewer, 33.2 percent of eyes treated with ocriplasmin achieved VMA resolution compared with 11.5 percent in the placebo group (p<0.001). Half of the patients with a baseline FTMH had VMA resolution vs. 25.5 percent in the placebo group (p=0.006). This correlated with a FTMH closure rate of 40.6 percent in the ocriplasmin group and 10.1 percent in the placebo group (p<0.001).

In terms of vision, the researchers say they saw greater changes in mean BCVA in both the VMA and FTMH ocriplasmin groups, +6.6 and +6.8 ETDRS letters, respectively, compared to placebo. At six months, 43 percent of the VMA patients treated with ocriplasmin had a gain of two or more lines of best-corrected acuity, as did 44.3 percent of the FTMH ocriplasmin group. In the placebo groups, only 28.7 percent of the VMA patients (p=0.018) and 30.4 percent of the FTMH (p=0.104) gained two or more lines. Also, 25.2 percent of the VMA patients treated with ocriplasmin and 27.4 percent of the FTMH treatment group had an improvement of three or more lines. For placebo patients, this gain was 10.3 percent in VMA cases and 13 percent for macular hole patients (p=0.003 and p=0.063, respectively). The study doctors say that most suspected treatment-related adverse events were mild, non-serious and occurred within seven days post-injection.1942

Researchers from Los Angeles’ Doheny Eye Institute say a little creative procrastination may help improve outcomes with reoperations for macular pucker and macular hole.

The researchers evaluated 10 pseudophakic eyes undergoing reoperation for MP or MH using visual acuity as an outcome measure. They excluded eyes with a history of prior vitreoretinal surgery, retinal detachment, vein occlusion, wet AMD or diabetic retinopathy. The surgeons used 25-ga. vitrectomy in all cases, and performed chromodissection with doubly diluted indocyanine green dye in all primary MH surgeries and in all reoperations for MH and MP. VA was correlated with immunohistochemistry for neurofilament and transmission electron microscopy of excised tissue in six out of 10 cases.

The investigators report that VA improved by more than three lines in six cases and worsened by more than three lines in three cases. All cases with a six-month or larger interval before reoperation (six cases) showed VA improvement (>3 lines), while three out of four with less than a six-month interval had VA worsening (>3 lines; p=0.03). The average postoperative logMAR VA was 1.59 ±1.07 (20/800) for patients with an interoperation interval shorter than six-months, with positive neurofilament staining and retinal cell debris present on the peeled membrane in 2/2 eyes. The surgeons say that waiting six months or more before reoperating resulted in logMAR VA of 0.42 ±0.25 (20/50) (p=0.03) and resulted in no evidence of neurofilament staining or retinal elements on the peeled membranes (0/4 eyes).

The researchers theorize that if repeat vitrectomy with membrane peeling is performed too early, there may not be adequate time for Müller cells to reform a layer of endplates over the denuded retinal nerve fiber layer, exposing it to damage during the second operation with resultant poor vision. Waiting longer than six months before reoperating, they say, may allow enough time to reform normal tissue planes, enabling a better surgical plane of dissection, which seems to be associated with less inner retinal damage and superior final vision.2147  REVIEW

Dr. Regillo is director of the Retina Service of Wills Eye Institute, and a professor of ophthalmology at Thomas Jefferson University School of Medicine.