Retina specialists explore anti-VEGF treatments for diabetic eye disease

As the rising prevalence of diabetes drives demand for care across medical specialties, ophthalmologists are increasingly focused on the causes, prevention and treatment of diabetic eye disease.

Overall, 4.2 million adults have diabetic retinopathy and 655,000 have vision-threatening diabetic retinopathy. Indeed, diabetic retinopathy affects almost one-third of adults with diabetes who are older than 40 years of age, according to the Centers for Disease Control and Prevention.

“Because the incidence of diabetes is increasing worldwide, the risk for developing central vision loss due to diabetic macular edema is also increasing. In my clinical practice, I frequently encounter diabetic patients who are referred for symptomatic vision loss and I diagnose them with center-involved diabetic macular edema,” Diana V. Do, MD, said.

Risk, management, prevention

Diabetic macular edema is the major cause of vision loss among people with type 2 diabetes and “is invariably present in patients with type 2 diabetes with [proliferative diabetic retinopathy],” according to one study in Eye and Vision. In another report, Varma and colleagues reported a greater burden of DME among non-Hispanic blacks, individuals with high blood glucose levels and those with a longer duration of diabetes.

Carl D. Regillo, MD, FACS, OSN Retina/Vitreous Board Member, described a correlation between severity of diabetic retinopathy and risk of DME, saying that as the level of diabetic retinopathy increases, from mild to moderate to severe, VEGF levels in the eye tend to increase, resulting in a greater likelihood that patients will have DME and its related vision problems.

Image: Do DV

“It is true that DME can occur at any stage of diabetic retinopathy, but it is more likely to be present and to affect vision as the level goes up,” Regillo said. “It is often stated that the correlation between drying and vision is not strong, that’s true, but the reason the correlation isn’t perfect is because you can reduce the swelling but not necessarily improve the vision. However, what is important to realize is that you can’t improve the vision without reducing the edema.”

There are also correlations between DME and systemic factors, such as lipid levels, hypertension and renal failure, although the degree of influence on DME is not well established, Regillo said.

Oral diabetes medications can influence DME severity as well. In particular, glitazones, which mitigate insulin resistance in patients with type 2 diabetes, are known precipitators of DME.

“That’s always something that we keep in mind when we see a patient with type 2 diabetes mellitus on oral medicines,” Regillo said. “The glitazone class of drugs is not used quite as often as it used to be, but it can cause edema or make edema worse.”

Important to prevention of diabetic eye disease is control of blood glucose levels. At 4 years after stopping intensive glycemic control in patients with type 2 diabetes, the risk for diabetic retinopathy development was reduced by half, Chew and colleagues reported in the National Eye Institute’s ACCORD Eye Study.

“The ACCORD follow-on study results underscore the importance of intense glycemic control by showing that the benefits of good control, even later in the course of disease, persist over a long time frame,” Regillo said.

In the study, diabetic retinopathy progressed in 5.8% of patients with intensive glycemic treatment and in 12.7% of those with standard treatment; the difference was statistically significant (P < .0001). Diabetic retinopathy progressed in 11.8% of patients who had received fenofibrate and in 10.2% of those who received a placebo. Fenofibrate did not show a lasting benefit, and intensive blood pressure control had no discernible effect, the authors wrote.

Treatment

According to some experts, the anti-VEGFs Eylea (aflibercept, Regeneron), Avastin (bevacizumab, Genentech) and Lucentis (ranibizumab, Genentech) have virtually revolutionized the treatment of diabetic retinopathy and DME.

“It’s very reassuring to know that the three commonly used anti-VEGF agents, aflibercept, bevacizumab and ranibizumab, are all very effective in the treatment of center-involved diabetic macular edema,” Do said.

The DRCR.net Protocol T study included 660 patients with DME who were randomized to receive intravitreal injections of 2 mg of aflibercept, 1.25 mg of compounded bevacizumab or 0.3 mg of ranibizumab.

At 2 years, mean visual acuity improved by 12.8 letters in the aflibercept group, 10 letters in the bevacizumab group and 12.3 letters in the ranibizumab group. Between-group differences were not statistically significant.

“These results suggest that aflibercept and ranibizumab are fairly equivalent in improving vision when used with this treatment regimen over a 2-year period,” Do said.

Among patients with baseline visual acuity of 20/50 or worse, mean visual acuity improved by 18.1 letters in the aflibercept group, 13.3 letters in the bevacizumab group and 16.1 letters in the ranibizumab group. The difference between aflibercept and bevacizumab was statistically significant (P = .02).

“When you look at both visual outcomes and anatomic outcomes, the results favor aflibercept and ranibizumab over bevacizumab, especially with regards to the proportion of eyes achieving a dry macula,” Regillo said. “That accounts for why [in] patients with more edema and particularly worse vision, 20/50 or worse at baseline, you start to see some differences between the drugs at both 1 and 2 years of follow-up in Protocol T.”

All three agents required about the same number of injections, 15 to 16 over a 2-year period.

“This similarity in the number of injections was a surprise because many retinal specialists initially thought that aflibercept may be more durable with a longer duration of action. But in this particular study, that did not appear to be the case,” Do said.

Antiplatelet Trialists’ Collaboration events occurred in 5% of participants in the aflibercept group, 8% of participants in the bevacizumab group and 12% of participants in the ranibizumab group. The difference between the aflibercept group and ranibizumab group was statistically significant (P = .047).

“There’s really not a great scientific rationale for why ranibizumab, which has a shorter half-life and is given at a lower dose than aflibercept, would actually have a higher systemic side effect profile. Also, the study is not powered to determine that these low rates of systemic side effects were significant. I think we can write that off and say that’s kind of a red herring,” John W. Kitchens, MD, said.

Aflibercept and ranibizumab were less cost-effective than bevacizumab in the treatment of DME, according to a post hoc analysis of the DRCR.net Protocol T trial that Ross and colleagues published in JAMA Ophthalmology.

Based on 2015 wholesale acquisition costs, the cost of aflibercept 2 mg was $1,850 per dose, ranibizumab 0.3 mg $1,170 per dose, and repackaged or compounded bevacizumab 1.25 mg about $60 per dose.

Investigators calculated incremental cost-effectiveness ratios (ICERs) for aflibercept, bevacizumab and ranibizumab for 1 year and 10 years. ICERs were calculated for participants and subgroups with baseline visual acuity of 20/32 to 20/40 or 20/50 or worse.

For all participants during year 1, the ICER of aflibercept compared with bevacizumab was $1.11 million per quality-adjusted life-year (QALY), and the ICER of ranibizumab compared with bevacizumab was $1.73 million per QALY.

“In eyes with decreased vision from DME, treatment costs of aflibercept and ranibizumab would need to decrease by 69% and 80%, respectively, to reach a cost-effectiveness threshold of $100,000 per QALY compared with bevacizumab during a 10-year horizon. For the subgroup with worse baseline vision, the costs would need to decrease by 62% and 84%, respectively,” the study authors wrote.

Other studies

Recent studies in the literature show various anti-VEGF treatment scenarios and outcomes.

A post hoc analysis of the VISTA and VIVID trials, published by Do and colleagues in Ophthalmology, showed that intravitreal injection every 4 weeks or 8 weeks was superior to laser treatment, with or without previous anti-VEGF therapy for DME.

In a report on the open-label extension of the RIDE and RISE trials published in Ophthalmology in 2016, Wykoff and colleagues showed that patients who received less frequent injections tended to have less advanced disease at baseline and responded better to initial ranibizumab treatment. The results suggested that earlier anti-VEGF treatment for DME may decrease the long-term treatment burden.

A meta-analysis published in Ophthalmology showed that rates of selected ocular and systemic adverse events with intravitreal aflibercept injection were similar to those of controls. Intravitreal aflibercept injection was generally well tolerated.

“Going back to the original anti-VEGF studies for oncology, we felt there were certain potential systemic side effects to high-dose intravenous anti-VEGF therapies, starting with bevacizumab,” Kitchens, the lead author, said.

The meta-analysis supported data from earlier phase 3 trials, showing that aflibercept had a safety profile similar to that of ranibizumab and bevacizumab, Kitchens said. However, he said that there is less data available regarding aflibercept than bevacizumab and ranibizumab.

“Aflibercept has not been looked at as extensively as Avastin and Lucentis from the systemic side effect standpoint just because it’s a newer drug,” Kitchens said.

Real-world study

At the Association for Research in Vision and Ophthalmology meeting in 2013, Szilárd Kiss, MD, presented real-world results of anti-VEGFs in treating DME, branch retinal vein occlusion and central retinal vein occlusion. Results were also published in Clinical Ophthalmology.

“The purpose was to see how we’re doing both in terms of how many times we’re injecting as well as outcomes. Really, it’s to see if the clinical trial results translate into the real-world results,” Kiss said.

Three separate cohorts of patients with DME, BRVO and CRVO underwent ranibizumab injections between 2008 and 2011. Investigators ensured that patients had not been diagnosed or treated in the previous 12 months and then recorded the number of injections patients received in the first year after being diagnosed.

The 2010 DME cohort included 1,009 patients.

Patients in the 2010 DME cohort received a mean of 3.6 injections over 12 months. Patients in the BOLT and DRCR.net Protocol I trials received a mean of nine injections, those in the RESTORE trial received seven injections, and those in the RIDE/RISE trials received 11 injections.

More than 60% of patients in the 2010 DME cohort received three or fewer injections, and fewer than 6% received 10 or more injections.

“This is really a significant undertreatment. One can argue that, in the real world, the patients may be a little bit different than in clinical trial settings. Though that may be true, it’s still a third of what the randomized pivotal trial showed, and it’s less than half of what the DRCR.net is showing,” Kiss said.

Kiss noted that patients in later cohorts received more injections than those in earlier cohorts.

In addition, real-world patients gained fewer letters of visual acuity than those in the clinical trials, Kiss said.

“That’s sort of sad and may be related to the number of injections,” he said.

Results show the importance of frequent follow-up early after diagnosis, Kiss said.

“The take-home message is that we’ve got to be more vigilant,” he said. “Anti-VEGF therapy seems like a home run, but that home run comes with eight injections in the first year and we’re only giving three. ... We’ve got to make sure that we uphold that treatment paradigm that has shown such benefit.”

In an article published in JAMA Ophthalmology, Marco A. Zarbin, MD, PhD, discussed the challenges of applying clinical trial results to clinical practice. Not all statistically significant results are reproducible or clinically important, he said.

“The P value refers to the probability of observing a given outcome under the conditions that you posit in a null hypothesis and given the statistical model that you use to predict the distribution of outcomes under the null hypothesis,” Zarbin said. “In frequentist statistical analysis, which typically is the type of analysis used in clinical trials, the statistics aren’t telling you whether the null hypothesis is true or false. They’re telling you the likelihood of observing this result, assuming the null hypothesis is valid and assuming that your model of the distribution of the outcomes is valid.”

Zarbin praised the DRCR.net’s research but urged clinicians to examine whether patients in clinical trials are representative of real-world patients in clinical settings.

“One of the enrollment criteria for Protocol T was that patients hadn’t had anti-VEGF therapy for a year or more. What if your patient has? Should you assume that your patient will get the same result as the patients in the Protocol T trial?” he asked.

Complement to anti-VEGF

In a study published in Ophthalmology, Peter A. Campochiaro, MD, and colleagues reported results obtained with AKB-9778 (Aerpio Therapeutics), a small-molecule vascular endothelial-protein tyrosine phosphatase (VE-PTP) inhibitor that stimulates Tie2 in vascular endothelial cells and suppresses VEGF-induced leakage. AKB-9778 was administered alone or in combination with ranibizumab in patients with DME.

Tie2, an endothelial cell selective receptor, modulates activity of the VEGF receptor. It has two binding proteins, angiopoietin 1 and angiopoietin 2. Angiopoietin 1 stimulates phosphorylation of Tie2 and activates it. Angiopoietin 2, a weak agonist, competes with angiopoietin 1 and reduces the activation of Tie2.

“In several retinal or choroidal vascular diseases, such as diabetic retinopathy, angiopoietin 2 levels are elevated, and another regulator, vascular endothelial-protein tyrosine phosphatase, is also elevated. VE-PTP dephosphorylates Tie2 and therefore, like Angiopoietin 2, is an inhibitor of Tie2. When VE-PTP and Angiopoietin 2 are elevated, Tie2 is deactivated and that causes stimulation by VEGF to be accentuated,” Campochiaro said. “We tested an inhibitor of VE-PTP, AKB-9778. By blocking VE-PTP, it prevents the inactivation of Tie2. Preclinical studies showed that subcutaneous injections of AKB-9778 made retinal vessels less responsive to VEGF. That’s what all the preclinical studies showed.”

The randomized, controlled study included 144 patients randomized to undergo subcutaneous injection of AKB-9778 15 mg twice daily and monthly sham intraocular injections, subcutaneous injection of AKB-9778 15 mg twice daily combined with monthly 0.3 mg ranibizumab, or monthly 0.3 mg ranibizumab injections combined with twice-daily subcutaneous placebo injections.

At 12 weeks, the mean change from baseline best corrected visual acuity was 1.5 letters in the AKB-9778 monotherapy group, 5.7 letters in the ranibizumab monotherapy group and 6.3 letters in the combination group.

Also at 12 weeks, central subfield thickness decreased 164.4 µm in the combination therapy group and 110.4 µm in the ranibizumab monotherapy group; the between-group difference was statistically significant (P = .008).

“Treatment of DME with ranibizumab provided a good reduction in edema,” Campochiaro said. “But treatment with the combination caused significantly more reduction in edema with many more patients having elimination of DME. About 50% of patients with DME don’t respond optimally to just blocking VEGF, and AKB-9778 is likely to improve outcomes in those patients and may reduce the number of anti-VEGF injections that are needed. In addition, patients treated with AKB-9778 alone had improvements in background diabetic retinopathy, and this will be pursued in future trials.” – by Matt Hasson

• References:
• Campochiaro PA, et al. Ophthalmology. 2016;doi:10.1016/j.ophtha.2016.04.025.
• Diabetic retinopathy. Centers for Disease Control and Prevention. http://www.cdc.gov/visionhealth/pdf/factsheet.pdf.
• Chew EY, et al. Diabetes Care. 2016;doi:10:2337/dc16-0024.
• Do, DV, et al. Ophthalmology. 2016;doi:10.1016/j.ophtha.2015.11.008.
• Evaluating how the treatments in the action to control cardiovascular risk in diabetes (ACCORD) study affect diabetic retinopathy (the ACCORD Eye Study). ClinicalTrials.gov. https://clinicaltrials.gov/ct2/show/NCT00542178.
• Kaidonis G, et al. Ophthalmology. 2015;doi:10.1016/j.ophtha.2015.05.004.
• Kiss S, et al. Clin Ophthalmol. 2014;doi:10.2147/OPTH.S60893.
• Kitchens JW, et al. Ophthalmology; 2016;doi:10.1016/j.ophtha.2016.02.046.
• Korobelnik J. Ophthalmology. 2014;doi:10.1016/j.ophtha.2014.05.006.
• Nguyen QD, et al. Ophthalmology. 2012;doi:10.1016/j.ophtha.2011.12.039.
• Lee R, et al. Eye Vis (Lond). 2015;doi:10.1186/s40662-015-0026-2.
• Ross EL, et al. JAMA Ophthalmol. 2016;doi:10.1001/jamaophthalmol.2016.1669.
• Varma R, et al. JAMA Ophthalmol. 2014;doi:10.1001/jamaophthalmol.2014.2854.
• Wells, JA, et al. Ophthalmology. 2016;doi:10.1016/j.ophtha.2016.02.022.
• Wykoff CC, et al. Ophthalmology. 2016;doi:10.1016/j.ophtha.2016.04.004.
• Wykoff CC, et al. Ophthalmic Surg Lasers Imaging Retina. 2016;doi:10.3928/23258160-20160324-01.
• Zarbin MA. JAMA Ophthalmol. 2016;doi:10.1001/jamaophthalmol.2016.1266.

• For more information:
• Peter A. Campochiaro, MD, can be reached at Johns Hopkins Hospital School of Medicine, Wilmer Ophthalmological Institute, Department of Ophthalmology, 719 Maumenee, 600 N. Wolfe St., Baltimore, MD 21287; email: pcampo@jhmi.edu.
• Diana V. Do, MD, can be reached at University of Nebraska Medical Center, 3902 Leavenworth St., Omaha, NE 68198; email: diana.do@unmc.edu.
• Szilárd Kiss, MD, can be reached at Weill Cornell Medical College, New York-Presbyterian Hospital, 1305 York Ave., 11th Floor, New York, NY 10021; email: szk7001@med.cornell.edu.
• John W. Kitchens, MD, can be reached at Retina Associates of Kentucky, 120 North Eagle Creek Drive, Suite 500, Lexington, KY 40509; email: jkitchens@gmail.com.
• Carl D. Regillo, MD, FACS, can be reached at Retina Service, Wills Eye Hospital, 840 Walnut St., Suite 1020, Philadelphia, PA 19107; email: cregillo@aol.com.
• Marco A. Zarbin, MD, PhD, can be reached at Department of Ophthalmology, New Jersey Medical School, 90 Bergen St., 6th Floor, Room 6155, Newark, NJ 07103; email: zarbin@earthlink.net.

Disclosures: Campochiaro reports he is a consultant for Aerpio Therapeutics, Alimera Sciences, Allergan and Genentech. Do reports she has served as a consultant to and received research funding from Alcon, Allergan, Genentech and Regeneron. Kiss reports he is a consultant for Allergan, Genentech and Regeneron. Kitchens reports he is a consultant for Allergan, Bayer and Regeneron. Regillo reports he is a clinical investigator for, has been a consultant for and has received research grants from Genentech and Regeneron. Zarbin reports he is a consultant for Frequency Therapeutics, Genentech/Roche, Healios KK, Makindus, Novartis and Ophthotech.

Does the incidence of APTC events in the DRCR.net Protocol T study influence your choice of anti-VEGF treatment?

Efficacy, not safety, was the primary outcome

Studies such as the DRCR.net Protocol T are designed primarily to look at efficacy. “The primary analysis consisted of three pairwise comparisons of mean VA change from baseline in the three treatment groups ... .” Thus, an efficacy outcome.

Of course, study investigators looked at safety as well, and they found a higher rate of APTC events with ranibizumab (12% with ranibizumab, 8% with bevacizumab and 5% with aflibercept). I doubt the sample size was determined to study safety differences, and I tend not to be overly influenced by secondary outcomes.

One can build arguments after the fact about why one drug or the other might be safer. If a consistent safety difference appears in future studies, the APTC findings will take on greater importance.

Andrew P. Schachat, MD, is vice chairman for clinical affairs, Cole Eye Institute, Cleveland Clinic. Disclosure: Schachat reports he is a consultant for the American Academy of Ophthalmology, is a member of SAB, Eastern Capital, and receives royalties from Elsevier.

Data support safety of aflibercept

The recent results of Protocol T of the DRCR.net reported that 5% of patients with diabetic macular edema treated with aflibercept suffered an APTC systemic event compared with 8% treated with bevacizumab and 12% treated with ranibizumab. The difference between aflibercept and ranibizumab was statistically significant.

These results do not change my approach to DME management as I primarily treat with the on-label drug that yielded the best visual results in the shortest period of time, with fewer injections: aflibercept. This study reassures clinicians and patients that safety is not being sacrificed for efficacy.

Jay S. Duker, MD, is an OSN Retina/Vitreous Board Member. Disclosure: Duker reports he is a consultant for and receives research funding from Carl Zeiss Meditec, Optovue and Topcon, consults for Allergan, Aura Biosciences, CoDa Therapeutics, Omeros and Santen, holds stock in Hemera Biosciences and Ophthotech, and serves on the board of directors of Eleven Biotherapeutics.