Retinal Disease: Advances in Treatment

Association for Research in Vision and Ophthalmology (ARVO) 2012

Roger F. Steinert, MD; Baruch D. Kuppermann, MD, PhD


July 11, 2012

This feature requires the newest version of Flash. You can download it here.

Editor's Note:
The following material is a transcript of a Skype interview. However, after the interview, Dr. Kuppermann provided the editor with written clarification of some of this material, with the purpose of giving the reader the most accurate information.

The Retina for Non-Retina Specialists

Roger F. Steinert, MD: Hello. I am Dr. Roger Steinert, Chair of Ophthalmology and Director of the Gavin Herbert Eye Institute at the University of California, Irvine. I have the distinct pleasure of introducing Dr. Barry Kuppermann, who is Chief of the Retina Service at the Gavin Herbert Eye Institute.

Our task today is to discuss the Association for Research in Vision and Ophthalmology (ARVO) meeting, particularly with reference to the topics on retina and the advances that we have been seeing in treating some very serious conditions in the retina. Barry is an internationally renowned authority and researcher in this area, so there is no one better suited for this discussion. Welcome, Barry.

Baruch D. Kuppermann, MD, PhD: Thank you.

Dr. Steinert: I would like to gear this toward non-retina specialists, which is still most of us, despite your best efforts to train everybody as a retina person. I would like to have you start by telling us a little bit about the latest results of CATT [Comparison of Age-related Macular Degeneration Treatment Trial],[1] explain what it is, and what the implications are for the treatment of retinal disease.

CATT at Year 2

Dr. Kuppermann: Thanks, Roger. It's a pleasure to be here, and I appreciate the opportunity.

CATT is a landmark National Institutes of Health (NIH)-funded trial that compares bevacizumab with ranibizumab for the treatment of wet macular degeneration. Bevacizumab is the monoclonal antibody of vascular endothelial growth factor (VEGF), which is approved by the US Food and Drug Administration (FDA) for systemic use in certain cancers, whereas ranibizumab is the antigen-binding fragment of that monoclonal antibody that was developed specifically for the eye. Ranibizumab (Lucentis®) and bevacizumab (Avastin®) are both proprietary to Genentech (South San Francisco, California), which was purchased by Roche (South San Francisco, California).

Bevacizumab was developed for the whole body and was discovered, with credit to Phil Rosenfeld at Bascom Palmer Eye Institute, to be useful in small doses intravitreally to treat wet macular degeneration with choroidal neovascularization. The retina community used intravitreal bevacizumab for about a year before ranibizumab was approved by the FDA for wet age-related macular degeneration (AMD); this is going back 5-7 years. Since that time, both drugs have been used extensively for wet AMD, with seemingly similar benefits to patients.

A question arose, however, as to how similar these 2 compounds were in treating wet macular degeneration -- was bevacizumab really as good as ranibizumab? This led to the development of CATT in an attempt to compare the safety and efficacy of these 2 widely used drugs for the treatment of wet AMD.

To go back in time a bit, when we first started using bevacizumab -- before ranibizumab was approved by the FDA -- the retina community was nervous about the possible side effects of bevacizumab, a drug that we were using off-label. Therefore, rather than give intravitreal bevacizumab on a strict regular interval, Phil Rosenfeld, through the PrONTO study that he designed and led, showed that we could achieve good results with as-needed injections guided by optical coherence tomography (OCT) imaging. In contrast, the pivotal trials (ANCHOR and MARINA) showing the benefits of ranibizumab had fixed regimens of dosing every 4 weeks.

Because we had already started using bevacizumab before the approval of ranibizumab and we weren't sure about how to use bevacizumab, we were using it in an aggressive "as-needed" fashion. We would bring patients back every 4-6 weeks and examine them; if there was fluid on OCT, blood visible in the retina, or a decrease in vision, we would inject them with bevacizumab.

When ranibizumab was approved, rather than use it as a fixed regimen every 4 weeks per the package insert, we started using it on an as-needed basis every 4-6 weeks, on the basis of OCT leakage and other symptoms -- much like we had already been doing with bevacizumab. Therefore, CATT was designed to not only compare the safety and efficacy of bevacizumab with ranibizumab but also to determine whether aggressive, as-needed injections provided similar efficacy as fixed regimen injections every 4 weeks.

It was a noninferiority study, which is a little bit complicated statistically, and it had 4 arms. The 4 arms in CATT were:

  • Ranibizumab, fixed dosing every 4 weeks.

  • Bevacizumab, fixed dosing every 4 weeks;

  • Ranibizumab, aggressive "as-needed" dosing; and

  • Bevacizumab, aggressive "as-needed" dosing.

A potentially controversial aspect of this study design is the absence of a loading dose. Many people would administer bevacizumab or ranibizumab in 3 monthly doses before switching to the as-needed mode, whereas other people simply used a single dose at the beginning and continued with as-needed dosing. CATT incorporated the as-needed dosing from the beginning into the trial design.

A year ago at ARVO, we heard the first 12-month results of this 2-year study.[2] What we saw last year was that bevacizumab every 4 weeks was noninferior to ranibizumab every 4 weeks. Ranibizumab on an as-needed basis was also noninferior to ranibizumab and bevacizumab every 4 weeks. Although results for bevacizumab were a little bit worse on an as-needed basis, these findings were deemed indeterminate -- it was not possible to determine whether the results that were nominally worse were actually inferior, owing to the statistical design of the study. In other words, as-needed bevacizumab was a little bit worse than the other regimens, but not so much as to be deemed inferior.

An important concern was also raised about macular thickness on OCT. The findings for ranibizumab were better, particularly the 4-weekly dosing regimen.

The question that was raised last year was, what would the findings show at year 2? The visual acuity outcome curves were just beginning to separate at year 1, so how would they manifest at year 2? Would the OCT thickness and the lack of retinal drying in the as-needed and fixed regimen bevacizumab groups manifest as decreases in visual acuity that were significant by year 2?

In addition, the year 1 results showed an increased incidence of systemic safety findings in the bevacizumab treated groups, although the findings were not necessarily the type of side effects associated with anti-VEGF therapy, such as strokes and other arterial thromboembolic events. These findings, however, were observed and statistically significant, even though the study was presumed to be underpowered statistically to show any safety differences between groups or drugs. Would these safety differences be worse at year 2?

Year 2 had one other twist. Patients on ranibizumab stayed on ranibizumab, and patients on bevacizumab stayed on bevacizumab, but they were randomly assigned to their dosing regimen. The 4-weekly-dosing patients were randomly assigned to either as-needed or to 4-weekly dosing, to see whether as-needed dosing would alter the outcomes. There was a complicated heuristic algorithm, making the results harder to present.

Dan Martin, the chair at Cole Eye Institute at the Cleveland Clinic, was the lead investigator, and he presented the data a year ago and again at ARVO this year. The bottom line was that the changes that we anticipated might be apparent at year 2 did not manifest. Once again, noninferiority was evident throughout the protocol, although there was a spread of results. Ranibizumab 4-weekly dosing did best, and bevacizumab as-needed dosing did the worst of the 4 groups in terms of nominal measures of visual acuity, but the differences were still statistically noninferior.

Beauty is in the eye of the beholder, and there was something for everyone in CATT. Basically, the results of year 1 were corroborated at year 2. Whichever treatment regimen was used (bevacizumab or ranibizumab, every 4 weeks or as-needed), very few people changed their practice patterns. Some people may be shifting toward bevacizumab, given that even after 2 years, visual acuity was found to be noninferior to visual acuity after treatment with ranibizumab. Even though bevacizumab was numerically worse, statistically it was noninferior. The discrepancy in OCT thickness remains.

Many of us expected that the year 2 findings would show a difference between the 2 drugs, but this expectation was not borne out. Similarly, the differences in safety findings did not show any greater spread at year 2 than they did at year 1.

IVAN at Year 1

Dr. Steinert: What is the IVAN [Inhibit VEGF in Age-related Choroidal Neovascularization] study?[3]

Dr. Kuppermann: IVAN was a similar study conducted in the United Kingdom, with different parameters. They compared bevacizumab vs ranibizumab, administered in "continuous" and "discontinuous" (to use their terminology) regimens. However, before the discontinuous ("as-needed") treatment, they administered a loading dose of 3 monthly injections. No further injections were given to patients in the discontinuous regimen group unless the eyes being treated met retreatment criteria -- primarily, the presence of any fluid on OCT.

Again, different from CATT, which used a single loading dose injection at the beginning of the trial, IVAN used 3 monthly injections. When retreatment criteria were met in CATT, a single injection was given, whereas when retreatment criteria were met in IVAN, 3 monthly injections were repeated.

Of interest, a controversial aspect of this study was that the noninferiority index differed. The threshold of what would be considered noninferior in CATT was 5 letters, but in the IVAN trial, it was only 3.5 letters. Defining noninferiority is as much art as science. There are statistical mechanisms for defining noninferiority, but at the end of the day, IVAN adjusted the noninferiority margin from its statistical calculations to what was considered more clinically relevant.

If you were to ask most investigators, or most retinal specialists, how much difference in visual acuity would be considered an appropriate noninferior margin, we would all agree that a difference of 1 letter or 2 letters would be in the bounds of noninferiority. A difference of 3 letters is where the debate begins. Some people would say that a difference of 3 letters is beginning to suggest inferiority. Many people would say that 4 or more letters of difference definitely suggests inferiority. The IVAN investigators split the difference between the 3 and 4 letter mark, defining noninferiority as 3.5 letters.

Although none of the groups in IVAN was found to be inferior (corroborating the CATT findings), it was very close, missing inferiority by a hair. Nonetheless, the IVAN investigators used a tighter noninferiority margin, and bevacizumab was still found to be noninferior to ranibizumab. So, IVAN corroborated the CATT results, in a different population in the United Kingdom, with a slightly different protocol -- a series of 3 monthly injections instead of 1 injection for "as-needed" treatment whenever retreatment criteria were met. However, the results were similar.

The Cost Consideration: Bevacizumab vs Ranibizumab

Dr. Steinert: Am I correct in deducing that these trials are not going to create any changes in typical practice patterns?

Dr. Kuppermann: I would agree with that. If there is a change, it might be a greater comfort with using bevacizumab.

In the United Kingdom, their institute for cost-effectiveness analysis is the National Institute for Health and Clinical Excellence (NICE). NICE might encourage the use of bevacizumab because it is so much cheaper than ranibizumab. Bevacizumab costs $20-$50 per dose. Systemic bevacizumab costs about $4000, but to treat the eye, all you need is about a $20 dose out of the vial. By comparison, a ranibizumab dose costs about $1900.

Not only did CATT (year 2) and IVAN (year 1) results came out this year, but other trials looking at these drugs are ongoing in Europe. So, countries can make individual decisions for their own regulatory agencies as to whether they want to allow use of bevacizumab.

Many countries in Europe do not allow bevacizumab use because it is off-label, and you are not allowed to use an off-label drug when an approved drug is available. Now that ranibizumab is approved, bevacizumab use is banned in many of those countries. In the United States, we are allowed to use off-label drugs if they are cost-effective. Because bevacizumab was so much cheaper than ranibizumab, Medicare and the insurance companies allowed us to use bevacizumab off-label initially, without any formal process, but eventually the process became more clarified and codified.

Europe is going to go through a change, and they might start using more bevacizumab. On the other hand, there is pushback about this, because the OCT thickness looks much better with ranibizumab treatment, and the frequency of injections was different.

Side effects could also be issues with bevacizumab, which was originally a systemic drug and is still used primarily as a systemic drug. The nonocular systemic complications of bevacizumab (the side effects) are significantly more frequent. Some of the reported adverse effects don't make a lot of sense, such as gastrointestinal bleeding seemingly unrelated to the traditional arterial thrombolic events typically associated with anti-VEGF therapy. No differences in the adverse effect profile were found between bevacizumab and ranibizumab in the eye studies, but cumulatively, a statistically significantly higher number of systemic complications are seen with bevacizumab. This has led to some pushback about using bevacizumab.

I am not sure what the US solution will be. We are hoping that physicians here in the United States will continue to be allowed to make those decisions. But in Europe, they may be looking carefully at the cost-effectiveness, especially as their own countries complete comparison trials. One by one, European countries are going to make those decisions.

Anti-VEGF Therapy for Diabetic Macular Edema

Dr. Steinert: The other interesting application has been this surge in use of, or at least interest in the use of, anti-VEGF therapy for treatment of diabetic macular edema. What was the buzz at ARVO about that, and what were the trial results?

Dr. Kuppermann: There has some evolution of thought about that topic. It is interesting to us as retina specialists, on the basis of both our clinical experience and our understanding of the disease state, that bevacizumab and ranibizumab seem to be quite similar in the treatment of wet macular degeneration. We knew that if there was going to be a difference, it would be small, and this was confirmed by CATT and IVAN.

The other indication that we use these drugs for, before we get into diabetes, is retinal vein occlusion. That is a label indication for ranibizumab. We used bevacizumab off-label before we got access to ranibizumab, and results seem to be similar, much as we see with wet AMD.

Diabetes is the other big potential indication for these drugs, along with macular degeneration and retinal vein occlusion. Our off-label use of bevacizumab didn't seem to be quite as effective during the first clinical trial, which was a Diabetic Retinopathy Clinical Research Network ( protocol. It is a NIH-funded trial controlled by clinicians. We designed the trials together as a group. I am always happy to wave that flag, because it's a great organization and well run. There are more than 100 investigators across the country. The clinical trials are all designed by the investigators without direct pharmaceutical company input.

In the protocol, I saw that ranibizumab, on a monthly basis, had dramatically better results than laser -- and, in fact, better results than steroids injected every 4 months. It is important to note, however, that the patients in the steroid group who were pseudophakic at baseline, who received injections of triamcinolone acetonide every 4 months plus laser at baseline, had visual acuity results similar to results with ranibizumab. It was a pretty impressive result: The eyes treated with ranibizumab monthly and the pseudophakic eyes treated with triamcinolone overall had very good visual acuity outcomes.

Genentech/Roche, the owners of both ranibizumab and bevacizumab, conducted pivotal trials evaluating the safety and efficacy of monthly ranibizumab for the treatment of diabetic macular edema. One of the dilemmas, and why diabetic indications take longer, is that the FDA mandates a 3-year trial for diabetic macular edema. For retinal vein occlusion, the FDA wants a 1-year trial -- 6 months of efficacy data, and another 6 months of safety data. For AMD, they want a 2-year trial -- 1 year of efficacy data, and another year of safety data. For diabetes, they want 2 years of efficacy data, and a third year of safety data.

So, diabetes is a harder indication to achieve. RISE and RIDE are the 2 pivotal trials by Genentech, assessing ranibizumab for the treatment of diabetic macular edema, and impressively, they show continued benefit.

With respect to ranibizumab for wet macular denegation, after the first 3 doses -- and this is why there was a loading dose used in the IVAN trial -- if you look at the curves, visual acuity tends to plateau after the third ranibizumab injection. You see a rise for the first 3 months and a steady state thereafter. With ranibizumab for diabetic macular edema, you get 50%-60% of the "bang" in the first 3 months. So the curve rises steepest initially, but it doesn't plateau. It continues to rise gradually over the full 24 months of the trial as patients receive monthly injections. That was very exciting.

We don't yet have the year 3 results on ranibizumab; we heard the 2-year results last June, so we should be hearing the year 3 results soon. Hopefully, it will be approved by the FDA for diabetic macular edema, because it is very different.

For example, there was a study out of the United Kingdom (BOLT) that looked at bevacizumab.[4] The bevacizumab results were similar to what our experience has shown. For example, if you look at 3-line gainers (the acid test for FDA approval), about 11% of eyes were 3-line gainers at 1 year; this improves to 32% at 2 years vs 42% of eyes with ranibizumab at 2 years.

The bottom line is there was no news yet on that trial, but RESTORE,[5] one of the European trials looking at ranibizumab in diabetic macular edema, showed that when as-needed treatment was administered in the second year, they maintained the gains achieved in the first year. They used a mean of 3.7 or 3.8 injections in the second year. In the third year, they used only 2.7 injections to maintain the gains.

It will be interesting to see, in RISE and RIDE, where they continue to give injections every month, if there are continuing gains from the monthly injections. The as-needed regimen in diabetes may fall by the wayside, because even in a dry retina, in theory if you keep injecting, there may be a benefit.

We are very eager to hear the year 3 results from RISE and RIDE. We won't have those until later this summer, and certainly by the next ARVO. We should hear about them by this fall.

The Human and Nonhuman Genome

Dr. Steinert: We will look forward to that. We only have a couple of minutes left, and I don't want to miss the opportunity to hear a brief summary of some of the highlights of the talk given by J. Craig Venter,[6] one of the preeminent authorities in genetics and genomics and a featured speaker at ARVO.

Dr. Kuppermann: Dr. Venter has had a very impressive career, with what he has achieved with analysis of the human genome. He gave a fascinating lecture. He has been looking at the impact of the nonhuman genome on humans in our close environment.

For example, Dr. Venter sampled the air we breathe in an urban environment, looking for nonhuman DNA in the air. Apparently animals, including humans, are aerosolizing (I'm not sure that's the right term) -- releasing epithelial cells all the time, which are lifted into the air that we then breathe in. These cells are in the air all the time. He sampled air in midtown Manhattan, for example, 20 stories up, and found that a huge percentage (more than 50%) of the air contained the DNA of rodents. It's kind of chilling; I hate to think about what the air contains at ground level. It was an "oh my God" moment. We all looked at each other -- we are consuming that all the time.

Dr. Steinert: Don't go in the subway.

Dr. Kuppermann: After all the work done on the human genome by Dr. Venter's group and others, there is now a large effort on looking at all the things in the human body that are nonhuman, and looking at this nonhuman DNA genome repository in our bodies. It turns out that something like only 10% of the genes in the human body are human. The rest are bacterial, from our gut and various organs.

There is some evidence (and this was also featured recently in Scientific American[7]) that if you look at the gut alone, for example, and consider the field of epigenomics -- where the environment can influence which genes are expressed -- it could be that the bacterial genes in our body are altering the environment and therefore modifying our own genetic expression. Humans are 99.9% equal from a genetic analysis standpoint, but if you look at the expression of those genes, it might be the gut bacteria and other parts of the body that influence the expression of those genes.

One example was people (myself included) who have had Helicobacter pylori infection. I was placed years ago on a regimen of antibiotics for H pylori. It turns out that H pylori may regulate levels of leptin and ghrelin to control satiety. Maybe that's why I have gained my luxurious extra 10-20 pounds (although it might just be aging, of course, because it has been 5 years since my treatment).

But it's interesting how the bacteria in our body play such a vital role in our health. They are not enemies by any means. It's very much a symbiotic relationship. There are pathogens that are our enemies, but the overwhelming majority of the bacteria in our body are our friends. In fact, there is a huge effort under way called Metagenomics of the Human Intestinal Tract (MetaHIT), in which many organizations have taken part to try and perform genetic analysis of all these additional genes in the human body that are not human genes. I thought those were the key features of interest in Dr. Venter's presentation.

Dr. Steinert: Excellent. It has been a terrific discussion, Barry. My many thanks to Dr. Barry Kuppermann from the University of California, Irvine, Gavin Herbert Eye Institute, and thanks also to our audience for tuning in.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.