This article is more than 5 years old. Information may no longer be current.

Protocol T comparison of anti-VEGFs gives clinicians guidance for treating DME

Issue: April 25, 2015

Add topic to email alerts

Receive an email when new articles are posted on

Please provide your email address to receive an email when new articles are posted on .

Subscribe

Added to email alerts

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

Back to Healio

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Back to Healio

Eagerly awaited findings of the DRCR.net Protocol T study did not disappoint, delivering “exciting” results for retina specialists looking for an answer to the question of which anti-VEGF to use in cases of center-involved diabetic macular edema.

One-year results of the head-to-head comparison of Eylea (aflibercept, Regeneron), Avastin (bevacizumab, Genentech) and Lucentis (ranibizumab, Genentech) showed all three drugs similarly improved vision, on average, in patients with baseline visual acuity of 20/40 to 20/32, but for patients with baseline visual acuity of 20/50 or worse, statistically significantly greater mean improvement in visual acuity was achieved using aflibercept.

Before 2014, ranibizumab was the only drug approved by the FDA to manage DME, and the choice was simple.

“If circumstances allowed use of an FDA-approved product, Lucentis 0.3 mg was the perfect choice prior to fall 2014,” OSN Retina/Vitreous Board Member Seenu M. Hariprasad, MD, said. But 2014 was a milestone year for the management of DME, he said, with FDA approvals given for Ozurdex (dexamethasone intravitreal implant 0.7 mg, Allergan), Iluvien (fluocinolone acetonide intravitreal implant, Alimera Sciences) and aflibercept, which made the choice more complicated.

Although the DRCR.net Protocol T results answer many questions about differences between the anti-VEGFs, other questions remain, according to Seenu M. Hariprasad, MD.

Image: Linder H

“We did not have data to guide our decision-making about the use of the three available VEGF inhibitors,” Hariprasad said, but now there is “clear-cut guidance to make treatment decisions.”

The study results will have an impact on clinical practice, according to Diana V. Do, MD, who said that she would likely recommend aflibercept for eyes with center-involved DME and a visual acuity of 20/50 or worse. But for eyes with a visual acuity of 20/40 or better, she said she is comfortable choosing any of the three agents.

“Many ophthalmologists and retina specialists were eagerly awaiting the results of the study in order to better guide their treatment recommendations,” she said.

The randomized controlled trial of 660 adults at 89 clinical sites throughout the U.S. was conducted by the Diabetic Retinopathy Clinical Research Network (DRCR.net) and sponsored by the National Institutes of Health. Findings were published in The New England Journal of Medicine.

Diana V. Do

Do said that the study proves, again, that DME can be treated with VEGF inhibitors on an as-needed regimen with monthly evaluations and still result in excellent visual acuity gains.

“These results are very exciting; many retina specialists are also looking forward to the 2-year results to see if the vision gains or the number of injections change over a longer period of time,” she said.

Getting underway

“Fortunately, the study was easy to perform,” John A. Wells III, MD, FACS, protocol chair for the trial and one of the three vice chairs of DRCR.net, said. “The biggest challenge, though, was how to pay for it.”

The study was designed in the aftermath of the CATT, which compared ranibizumab and bevacizumab in patients with wet age-related macular degeneration.

“The original thought was to design a trial to compare ranibizumab and bevacizumab for DME,” Wells said. “However, before we got the study off the ground, aflibercept became approved for AMD and became available in the marketplace, so we thought we should really compare all three drugs.”

Hence, the design of the trial was changed to a three-drug comparison. But the high cost of ranibizumab and aflibercept might have made the study too expensive.

“Although we were funded by government grants, collaboration with industry is allowed under DRCR Network guidelines,” Wells said.

Regeneron and Genentech stepped up to provide aflibercept and ranibizumab for free. Government funding absorbed the cost of bevacizumab due to its relatively low price.

There were also discussions as to whether the study should be an inferiority, non-inferiority or superiority trial. “We decided to make it a superiority study,” Wells said. In addition, patient recruitment was expected to be difficult, but all 660 patients were recruited in 12 months.

The study’s treatment algorithm is not as complicated as commonly believed, but to ensure compliance with re-treatment requirements, investigator discretion was eliminated.

“It would have been a challenge to ask investigators to figure it out themselves,” Wells said. As a result, everything became Web-based, so that the data from each of the patient’s visits could be entered into a computer. “The computer then told the doctor whether to treat the patient or not,” Wells said.

More than 99% of required study injections were given, and more than 95% of patients (excluding deaths) completed the 1-year follow-up.

“It was previously unknown whether the three drugs were of comparable efficacy. That was the gap we were trying to fill,” Wells said. “We had two FDA-approved drugs that we knew were effective, but bevacizumab, which is off label yet much cheaper, had not been widely studied in a randomized controlled fashion. There is no FDA-approved ocular indication for bevacizumab.”

Letter gain and controversy

“When all 660 patients in the trial are considered, aflibercept was significantly better in achieving letter improvement, a two- to three-letter gain over bevacizumab or ranibizumab,” Daniel F. Martin, MD, chair of the Cole Eye Institute at Cleveland Clinic, said. “However, this overall result has limited clinical application because the pre-specified subgroups behaved so differently. The key to understanding and applying the results of this trial are found in these subgroups.”

Daniel F. Martin

Wells said the subgroup analysis of eyes with better baseline vision (between 20/32 and 20/40) and eyes with worse baseline vision (20/50 or worse) was not arbitrary.

“We pre-specified, before the study started, that we were going to look at these two different subgroups,” Wells said.

“There was no difference between drugs if you were 20/40 or better whereas there was a large advantage for aflibercept if you were 20/50 or worse,” Martin, the lead author of an associated editorial in The New England Journal of Medicine, said. “When there is an interaction that is highly significant between baseline visual acuity and outcome, the overall results do not apply to patients in either subgroup.”

For those patients who started at better baseline vision, all three drugs resulted in a gain of about eight letters on average, Wells said.

“So you can use any of the drugs in these eyes and feel confident that you will get improvement in vision,” he said. “But in eyes with 20/50 or worse vision, you will get more improvement on average with aflibercept.”

Some controversy, though, swirls around the visual acuity findings of the study. For example, in a statement from Genentech issued on Feb. 18, the company said, “A 4.7 letter difference was observed with a baseline visual acuity of 20/50 or worse,” between ranibizumab (14.2 letters) and aflibercept (18.9 letters). Genentech questioned “whether differences in baseline characteristics influenced study results particularly in patients with worse vision.”

Cost conundrum

In the CATT, which Martin chaired, no difference in visual acuity was found between ranibizumab and bevacizumab in patients with wet AMD.

“That head-to-head experiment was repeated five times around the globe, for a total of six studies, all of which showed no difference between ranibizumab and bevacizumab in VA outcomes at 1 year. And now we have this DRCR Protocol T study that shows no difference between ranibizumab and bevacizumab in visual acuity for DME,” he said. “I do not know how many more times we need to repeat this experiment for people to accept there is no meaningful difference in visual acuity efficacy between these drugs. At some point, if there is no difference, cost has to become one of the major variables that influences how we treat patients.”

Hariprasad said that he “wholeheartedly” agrees with Martin’s message. For patients with worse than 20/40 baseline vision, aflibercept should be considered as the first-line agent and bevacizumab as the alternative, Hariprasad said. “If vision is 20/40 or better, the choice comes down to either aflibercept or bevacizumab, given lack of significant visual differences between bevacizumab and ranibizumab and the large cost difference between these two drugs,” he said.

According to Charles C. Wykoff, MD, PhD, a single treatment of bevacizumab can be obtained for $60 compared with ranibizumab 0.3 mg for roughly $1,100 and aflibercept 2 mg for approximately $1,900.

“We live in a world where cost matters,” Wykoff said. “Retina specialists are acutely aware of being tracked for what medications they use and the cost of those medications. Therefore, in many circumstances, retina specialists choose to use the less expensive alternative first, even though that medication may be significantly worse at achieving retinal drying.”

Charles C. Wykoff

From an anatomic perspective, in this study, bevacizumab “performed significantly worse than both aflibercept and ranibizumab,” Wykoff said.

Wykoff was surprised by how inferior bevacizumab was anatomically to the other two agents.

“The affinity of bevacizumab for VEGF may be less than for either aflibercept or ranibizumab,” he said.

Hariprasad noted that aflibercept was most effective at quickly drying the retina, as seen on OCT, numerically better than ranibizumab and statistically significantly better than bevacizumab.

Wykoff said it is important that patients be informed about cost differences. The study did not change medication for any patient. However, “you could create a system where you start with the cheaper medication; then, if you do not see the anatomic and visual benefit that you like, switch to a more expensive agent,” he said.

Wykoff himself uses all three medications and does not hesitate to switch from one to another in a patient with an incomplete response.

“Physicians are surprisingly immune to how much things cost,” Martin said. “The health care industry is changing rapidly, and with increased transparency and emphasis on value-based care, cost is going to become as important of a variable as is efficacy. You can be certain that if we as a specialty do not practice in a cost-conscious way, those who are actually paying the bill will try to limit our choices. That is what we are seeing today, and I submit it is in part a result of our not ‘owning’ the total cost of care as a vitally important component of how we practice.”

For the 75% of patients with DME who present at 20/40 or better, “it is perfectly reasonable to start with bevacizumab, a highly effective and the least expensive option,” Martin said.

“Because there is a cost differential between the three drugs, cost concerns can certainly play a role,” Do said. “Both aflibercept and ranibizumab are much more expensive than bevacizumab. Patients who do not have insurance cannot afford to undergo treatment with the more costly therapies. Fortunately, the study results demonstrate that bevacizumab is an effective and safe treatment for DME, so I would feel very comfortable offering this medication to patients with financial restrictions.”

Study limitations

In an editorial in OSLI Retina, Wykoff commented on the study and its limitations: “The bevacizumab used was not the same quality of bevacizumab that we as retina practitioners use. The bevacizumab that was provided for the trial was compounded at a single compounding pharmacy and delivered in glass vials, whereas the bevacizumab that we use clinically is delivered in plastic syringes.”

The editorial referenced a 2015 article in JAMA Ophthalmology that showed a wide variability in the integrity of bevacizumab when compounded in plastic syringes.

“We believe that the potency and the biologic activity of the bevacizumab we use clinically may be different from the bevacizumab of the study,” Wykoff said.

A second limitation of the study was that the protocol was not strictly monthly or as-needed dosing.

“For the first 6 months of the trial, patients were given an injection every month unless the patient was 20/20 and had no significant fluid on an OCT scan,” Wykoff said. “But after 6 months, once an eye was stable, neither losing nor gaining vision and neither losing nor gaining central retinal thickness on OCT, for two injections in a row, the next injection was deferred until the patient lost vision or gained more retina thickness.”

Wykoff said this protocol is different from what most clinicians use in daily practice.

“You may achieve different results if you use the medications in a different manner,” he said.

“The results of this study provide retina specialists with more evidence-based medicine in which to guide their treatment recommendations for DME,” Do said. “Ophthalmologists should carefully read the manuscript and determine how they can apply the conclusions of this research to their patient population. The enrollment criteria and re-treatment criteria may be different from what we see and do in our daily clinical practice, so the ultimate decision on which VEGF inhibitor to use may be more complicated and is best discussed between the patient and physician.”

Looking ahead to year 2

“DME is a chronic disease,” Wykoff said. “We need more than 1-year data. I look forward to the second-year results of the trial.”

Two-year study results will likely be shared in early 2016, according to Wells.

“There will be no other studies directly from this protocol after that,” he said.

There are other medications for DME in development, Wells said.

“One thing this study and other studies have shown is that while these drugs are effective, about 25% to 33% of eyes, despite a year or more of anti- VEGF therapy, continue to have chronic persistent edema,” he said. “I know there is an effort to try to develop drugs that might target those eyes that do not respond completely to the anti-VEGFs.”

“Although the DRCR T answered many questions about the differences between the anti-VEGFs, other questions remain,” Hariprasad said. “For instance, we need to understand better the role of implantable corticosteroids and laser in the management of DME, perhaps in combination with the anti-VEGFs.”

One area of research the DRCR.net is interested in pursuing is the role of steroids in treating DME. – by Bob Kronemyer

What are the real-world therapeutic strategies for treating DME in practice today?

Start with anti-VEGF

When a patient presents with center-involving symptomatic diabetic macular edema, most retina specialists will recommend starting treatment. Most of these patients will be phakic patients, and I think the majority of retina specialists will start with an anti-VEGF agent. The modern literature, regardless of the agent, overwhelmingly supports the use of at least multiple sequential months of regular, routine injections of ranibizumab, bevacizumab or aflibercept.

David A. Eichenbaum

The majority of patients will do well with any of the three commonly employed commercially available agents from a disease-control standpoint. However, there are certain specific clinical situations in which using one agent may be more preferable to another. For example, ranibizumab is an antibody fragment, which provides some theoretic reduction of systemic risk; therefore, a patient who may have recently had a thromboembolic event, an amputation, a peripheral infection or a fracture may be better served by ranibizumab, as opposed to bevacizumab or aflibercept.

On the other hand, new literature from DRCR.net Protocol T suggests that patients with poorer presenting vision, assuming they are healthy and stable systemically, may be better served by aflibercept.

Still, we do not have a preponderance of evidence supporting clear superiority of any one of the three agents in any specific clinical setting. All three are very good drugs and reasonable initial therapeutic choices.

Many patients, especially after the first 12 to 24 months of treatment, whether monthly or slightly less often, require substantially less treatment due to modification of diabetic retinopathy severity. Back steps in diabetic retinopathy severity is demonstrated most profoundly with ranibizumab, but has also been shown with aflibercept and bevacizumab. The baseline populations in the trials showing regression of retinopathy have differing levels of severity, which may account for the minimally dissimilar effects between the agents. Anti-VEGF agents probably sort of “reset” the clock for diabetes, if the patient is dosed in a fashion similar to the pivotal trials with any of the agents.

I believe that the published demonstrations of efficacy and safety trump the divergent cost of the FDA-approved drugs. However, in specific situations, it is important to discuss cost with the patient, and it sometimes has a basis on agent selection.

David A. Eichenbaum, MD, is a partner at Retina Vitreous Associates of Florida, Tampa. Disclosure: Eichenbaum reports he is a paid speaker, consultant and/or has supported research from Genentech, Allergan, Regeneron, USRetina and Hemera Biosciences.

Steroid may be introduced in patients with poorer vision

DME treatment has truly undergone a revolution in the past couple of years, and even in the last year. The old treatment paradigm was laser as the first-line therapy. However, the new treatment paradigm is pharmacotherapy, with typically an anti-VEGF injection. I usually start with aflibercept, in part because aflibercept came out on top in the DRCR.net Protocol T, especially in patients who have moderate to severe vision loss (20/50 and worse).

Szilárd Kiss

After the first injection, the patient returns in about 1 month, at which time we re-evaluate the visual acuity, as well as the extent of the edema (assessed mostly with OCT). Typically, patients tend to respond very well to the anti-VEGF therapy. Then patients normally have a second or third injection, spaced 4 to 6 weeks apart. After the second or third injection, I also re-evaluate both the level of response with OCT and visual acuity.

If I do not observe the response that I think I should be seeing for drying out the retina and improving visual acuity, I begin contemplating other therapy, usually an intraocular steroid. I typically turn to an Ozurdex (dexamethasone intravitreal implant 0.7 mg, Allergan) injection, especially for a younger phakic patient who may not require a long-term steroid effect, but rather 4 to 6 months. I usually see the patient again 2 months after the Ozurdex injection to evaluate the response. I may also augment the steroid injection with an anti-VEGF injection to try to better dry out the retina and improve vision.

The paradigm sort of shifts, though, in a patient with severe edema and perhaps a vision of 20/200 or worse, for which I introduce a steroid earlier — after the second anti-VEGF injection or even initially. This is a protocol especially for someone who perhaps has had diffuse macular edema for many years. In these scenarios, I often inject the Iluvien implant (fluocinolone acetonide, Alimera Sciences), which lasts 2 to 3 years. More than 85% of my patients can be successfully treated with anti-VEGF and/or steroid injections.

I rarely use laser these days. I find that augmenting with laser is sometimes necessary, but often you can manage with just pharmacotherapy. In fact, there is some evidence from the DRCR.net Protocol I that if you do laser, you may be leaving some vision on the table.

Szilárd Kiss, MD, is director of clinical research at Weill Cornell Medical College, New York. Disclosure: Kiss reports he is a paid consultant to Genentech, Regeneron, Allergan and Alimera Sciences.