Clinical Trial Updates

Putting Research Into Practice for Retinal Diseases

Introduction

The clinical management of retinal diseases such as age-related macular degeneration (AMD), diabetic macular edema (DME), and retinal vein occlusion (RVO) continues to evolve rapidly as novel therapies and new clinical trial outcomes expand and refine our practice patterns. In this educational activity, our primary goal is to clarify optimal management of these common retinal diseases in the context of data from recent clinical trials. We will be emphasizing salient points and take-home messages for our readers.

—Peter K. Kaiser, MD

Age-related Macular Degeneration

Dr Kaiser: Dr Martin, you designed CATT^*1 and have extensive knowledge of the other key AMD treatment comparison studies. Please give us an overview of the results of these trials.

Dr Martin: There have been 6 randomized clinical studies comparing bevacizumab and ranibizumab for the treatment of AMD: CATT, IVAN, MANTA, GEFAL, BRAMD, and LUCAS.^1-6 All 6 have reported their results, with 4 studies published^1-4 and 2 studies presented.^5,6 The data from these studies show no differences with respect to visual acuity outcomes between bevacizumab and ranibizumab. There are, however, small differences with respect to anatomic outcomes, such as amount of residual fluid and retinal thickness. Although both agents are remarkably effective in decreasing fluid, ranibizumab was shown to decrease fluid more. The absolute differences were small and there did not appear to be a visual acuity correlate.

There are no differentiating safety signals that emerge when comparing bevacizumab and ranibizumab, and when you consider the major events—death, myocardial infarction, stroke—the 6 trials do not show any statistically significant differences. In CATT, there was some concern about the difference in total serious adverse events, with a higher rate among bevacizumab users than among those using ranibizumab (34.5% vs 28.4%, P=.02)⁷; however, these findings were not reproduced in the other 5 trials.

Dr Kaiser: A key finding of the comparison studies was the conclusion concerning treatment regimens, namely discontinuous, or as-needed, protocols vs monthly/fixed treatment. Some of the studies reported that discontinuous, or as-needed, treatment was “not non-inferior” to monthly treatment; and clinicians have interpreted the findings related to these terms differently or inconsistently. What is your assessment of these results?

Dr Martin: CATT and IVAN were the only trials from the group of 6 major trials to compare monthly treatment with as-needed, or discontinuous, treatment. In analyzing the 2-year data from CATT,⁷ and examining all 15 pair-wise comparisons, we find no statistically significant difference in any of the treatment arms. We took advantage of the power that we had and collapsed the data with respect to drug and dosing regimen. This revealed a difference of 2.4 fewer letters gained with the as-needed regimen than with the monthly treatment regimen after 2 years (P=.046). This may lead people to conclude that we should not use an as-needed regimen; however, this is where the particular differences of applying population statistics to an individual level arise. I still usually begin with a less-than-monthly regimen because the data from CATT showed that 10% to 20% of patients needed only 3 injections, and although this need for low treatment frequency occurs in a minority of patients, the monthly treatment approach does not allow for identification of these patients. With the CATT data, the treatment curves do not begin to diverge until 9 months of therapy (Figure 1).⁷ Applying this information to my clinical practice, I will usually start with as-needed or, sometimes, treat-and-extend. If I find that the patient is losing ground or having multiple relapses, I will not hesitate to change to more frequent therapy. I am trying to balance the treatment burden and the risk that comes with every injection against a very small difference in outcomes with a less-than-monthly injection regimen.

Figure 1: CATT 2-Year Data: Mean changes in visual acuity from enrollment over time.⁷ Reprinted with permission from Elsevier.

Dr Ho: I find it interesting that you have redirected your treatment paradigm to begin with a less-than-monthly approach. In my opinion, individualizing treatment is important, but I am not willing to “let the reins out” right away with an as-needed regimen when I am starting treatment for a patient with newly diagnosed wet AMD. I will do 3 or 4 monthly injections and then begin to consider a treat-and-extend regimen, because my bias in AMD, based on the totality of all the trials, is that more treatment is better for my patient. Although the treatment burden is greater, I start off with several monthly injections.

Dr Duker: I think that true as-needed protocols are diminishing in popularity. In the 2013 American Society of Retina Specialists Preferred Practice Patterns Survey, 16.4% of US respondents reported that they would use the wet AMD treatment strategy of seeing the patient monthly and treating only if there were signs of neovascular activity.⁸ To do an as-needed regimen properly, you still need to see the patient monthly, at least at the beginning, and that is still a burden to the patient. Treat-and-extend has really emerged as the way the majority of us handle patients with wet AMD.⁸

Dr Kaiser: I think it is important to clarify exactly what “as-needed” means. Most physicians who use the term “as-needed” (or “prn”) to describe their treatment plan are not actually doing a true monthly as-needed regimen like the ones performed in the comparison studies they are relying on for guidance; instead, they are doing a treat-and-extend type of as-needed regimen that extends the time between visits, but still results in as-needed injections. Neither CATT nor IVAN evaluated that type of regimen. The CATT and IVAN as-needed regimens still necessitated assessment with optical coherence tomography (OCT) every 28 days, with treatment based on signs of active neovascularization.^1,2

Dr Duker, Dr Martin brought up an interesting point earlier when he mentioned that patients in CATT and IVAN receiving ranibizumab had less residual fluid than those treated with bevacizumab.^1,2,7,9 Then the VIEW studies came along and indicated that patients receiving aflibercept had less fluid than those treated with ranibizumab.¹⁰ What is the relevance of these facts to you as a clinician?

Dr Duker: I think that there may be a real difference among the anti-vascular endothelial growth factor (VEGF) agents with respect to the “drying effect”, which may reflect possible differences in anti-VEGF activity for each medication. Some theoretical studies with pharmacokinetic modeling have looked at concept, but it has not been established in clinical studies.¹¹ We all see patients with AMD who have small amounts of persistent subretinal fluid that does not resolve with monthly treatment, and yet their vision does not suffer. Unlike in DME, in which signs of anatomic improvement continue to be seen many months into treatment, such a pattern is not the case with most patients who have AMD, in which condition anatomic stability typically is achieved in the first 3 to 6 months of treatment.

For patients with good vision and a small amount of subretinal fluid, I will regard their small amount of fluid as a new baseline, and employ a treat-and-extend protocol relative to that baseline. When the fluid increases or the visual acuity drops, I shorten the interval between treatments. I think that intraretinal fluid is a better indicator of disease activity than is subretinal fluid.

Dr Martin: I agree with Dr Duker–in the CATT macular morphology paper, subretinal and subretinal pigment epithelium fluid did not have any adverse effects on visual acuity outcomes; in fact, the mean visual acuity at 1 year was actually 3 letters better for patients with subretinal fluid.¹² The story with intraretinal fluid was quite different; patients who had small amounts of intraretinal fluid in the center of the fovea had a mean visual acuity that was 9 letters worse.¹²

Dr Duker: I think that we tend to overlook this structure/function relationship in neovascular AMD because we use OCT as a gauge of disease activity. We have this mantra that we need to dry up all fluid, even though some eyes just will not respond no matter how much treatment we give them.

Dr Martin: Knowing this information regarding which fluid categories tend to be more problematic with respect to visual consequences, and knowing the concerns about retinal thinning with monthly treatment and geographic atrophy seen in CATT and IVAN,^9,13 I wonder about our drive toward the goal of a completely dry OCT. Is that the best goal for our patients? It is certainly a simpler approach, but I continue to encounter data that make it a little easier to tolerate tiny amounts of residual fluid that will not go away over time.

Dr Ho: I think that a binary approach of “dry” or “not dry” may be limiting. In the HARBOR study—the only study to have all spectral-domain (SD)-OCT, the reading center had an algorithm graded on a binary approach,¹⁴ whereby many of the variables such as cysts in the retina and fluid under the retina could be evaluated on a continuum. Ultimately, I still think that in AMD we want to dry the macula as much as possible. Practically, when I begin to see atrophy in my patients with AMD who are receiving continuous anti-VEGF therapy, I begin to back off a bit because I fear the potential for exacerbation of retinal pigment epithelial atrophy and choroidal atrophy.¹⁵

Dr Duker: Much of what we are discussing here pertains to the “maintenance” phase of treatment, during which we are trying to maintain visual gains and when a little bit of fluid may be tolerable. What about initial treatment? One of the analyses that came out of the VIEW study looked at patients who were still wet after 3 monthly injections of anti-VEGF therapy. They were either switched to injections every 2 months or maintained on monthly therapy: those who were treated with anti-VEGF therapy every 2 months had worse visual outcomes than those who were maintained on monthly therapy.¹⁶ What should our initial approach be? Should we treat monthly to start? Should we monitor patients monthly?

Dr Martin: I suspect that virtually all clinicians begin treatment on a monthly basis with their drug of choice until they have achieved a desired end point. For most of us, that will be a dry OCT with as much normalization of the foveal anatomy as possible. That goal will never be achieved in some patients, and they may wind up with monthly injections. In other patients, normal foveal anatomy may be restored with a single injection. And this is where the debate begins. Are 3 monthly injections at the beginning of treatment necessary in every case if you intend to do as-needed or treat-and-extend? Specifically, if you achieve normal foveal anatomy after the second injection, is the third injection necessary? Many physicians give it because it has become part of clinical practice. But where are the data that shows it is necessary—those data do not exist. If the purpose is to add extra insurance that the lesion will remain dry after the third injection when you do switch to as-needed or treat-and-extend, again, there are no data to support that. In CATT, 12 monthly injections did not add any insurance to lesion stability when patients switched to as-needed in the second year of the study (Figure 2).⁷ So why do we think that 1 more monthly injection would? It is perfectly fine to give that third injection with normal foveal anatomy. You just have to understand it is not a data-driven decision. It is the vestige of an old protocol that has been perpetuated in the absence of data.

Figure 2. Mean change in visual acuity from enrollment, over time, by dosing regimen within drug group.⁷ Reprinted with permission from Elsevier.

Dr Kaiser: I think all of us here use a treat-and-extend regimen, which ultimately necessitates more injections than does as-needed, but certainly less than monthly treatment. Given that, what are your thoughts on the capacity of treat-and-extend protocols to individualize therapy? We have the results of the LUCAS study,⁶ which used an “inject and extend” protocol that provided treatment at every visit, extending treatment intervals when disease activity was not evident, and shortening intervals when there were signs of recurrent activity. This protocol is a little bit different from the approach most of us employ in treat-and-extend, but still gives us Level 1 evidence supporting such an approach, especially with respect to visual outcomes.

Dr Ho: The whole point of individualized therapy was brought out elegantly in the second year of the HARBOR study,¹⁷ when patients were treated on an as-needed basis. During the second year, there was a surprisingly flat distribution of the number of injections needed to maintain the visual gains achieved in the first year. Treat-and-extend allows you to individualize AMD therapy.

Dr Kaiser: According to the visual acuity outcomes from the VIEW extension study,¹⁸ in which patients were required to continue a modified as-needed regimen (treatment at least every 3 months at the beginning and every 2 months after aflibercept approval), the mean change in the visual acuity curve was almost a straight line. This stands in contrast to other long-term extension studies such as HORIZON and SECURE,^19,20 in which patients were switched to as-needed protocols from monthly treatments, or to the as-needed treatments according to predefined retreatment criteria. In HORIZON and SECURE, there was a drop-off in visual acuity over time in the extension phase. From my perspective, that validates the use of a treat-and-extend protocol to preserve the early visual gains.

After considering these data pertaining to frequency of treatment, what might be an initial agent of choice for treatment of AMD, and for what period of time might treatment intervals be extended ?

Dr Ho: I think we have 3 excellent treatments for neovascular AMD. I prefer to use on-label therapy, such as aflibercept or ranibizumab. If a patient has an associated retinal pigment epithelial detachment, I lean toward the use of aflibercept. For patients who are refractory or who develop tachyphylaxis to 1 of those 2 anti-VEGF agents, I will switch to the other agent.²¹ I also have used bevacizumab in situations in which available resources were insufficient to meet the financial demands of the first 2 options. My outer treatment extension boundary for patients with wet AMD is approximately 12 weeks, although I tend to get nervous after 8 weeks.

Dr Duker: I, too, prefer to use an on-label drug, and I usually begin treatment with aflibercept, on the basis of the results of the phase 3 VIEW 1 and VIEW 2 studies.¹⁰ I will try to get patients out to 3 months, and I rarely will go longer. If I do extend the time between treatments to longer than 3 months, I still want to see the patient regularly, and not allow 4 or 5 months to elapse between visits. When the visits get that far apart, then the regimen approach becomes as-needed, rather than treat-and-extend.

Dr Kaiser: I also prefer to use on-label therapy if the patient’s financial considerations do not limit treatment options. In my practice, we are leaning toward using aflibercept. For patients who have insurance limitations or who are concerned about the co-payment amount, I will start with bevacizumab. If they do well with bevacizumab, then I continue that therapy. If they do not, I will then consider switching to a different agent.

Dr Martin: I discuss all 3 anti-VEGF agents with my patients. If the choice is left entirely to me, I begin with bevacizumab. I may start with an as-needed or treat-and-extend regimen, again after discussing the options thoroughly with the patient. In terms of how far I would extend treatment intervals, I agree with my colleagues, that beyond 8 or 10 weeks, I start to get nervous, at least in the first year of treatment.

Dr Kaiser: Dr Ho, do you use different treat-and-extend regimens according to which anti-VEGF agent you choose?

Dr Ho: I do not; I get very cautious with any of the 3 anti-VEGF agents that we have been discussing when it gets to the 3-month point. My experience thus far is that aflibercept may last a little longer for some patients, but I have seen the opposite as well. Each patient may respond differently to aflibercept, to bevacizumab, or to ranibizumab.

Dr Duker: I also think that AMD has different activity levels at different times. Some patients who have had their treatment intervals extended out to 3 months may experience an increase in disease activity and suddenly require a greater frequency of injections. I also think that it is appropriate to periodically, say every 6 to 12 months, attempt to extend treatment intervals for patients who have been receiving monthly injections, to see if the disease activity level has changed and if it is possible to space out the treatment intervals.

Dr Kaiser: In how many of your patients with AMD have you been able to stop anti-VEGF therapy? Even when I have stopped therapy in a patient, it seems that I have to restart it 6 months later.

Dr Duker: Only in a very few patients have I been able to stop anti-VEGF therapy. Once I am able to get a patient out to 3 months and he or she is doing well, I still want to evaluate the patient on a regular basis, effectively adopting a monthly as-needed approach. My patients will frequently choose the additional time afforded by getting an injection every 3 months over spending the time to come into Boston every month.

Dr Kaiser: My experience has been similar. I rarely can stop my treat-and-extend regimens. For patients with AMD who do well for several months without treatment, I am always concerned about the possibility of hemorrhage, which can be devastating, so I continue to treat even when they are dry, and extend the time intervals.

Let us now discuss safety comparisons for the different anti-VEGF agents for a moment. Several randomized, published comparison studies have shown similar safety profiles for most known anti-VEGF adverse events for bevacizumab and ranibizumab as well as for ranibizumab and aflibercept.^2-6,9,10 We do not, however, have any head-to-head safety analyses between aflibercept and bevacziumab. Regarding aflibercept and ranibizumab, the bottom line from the phase 3 VIEW studies was that the 2 drugs were equal in safety.¹⁰ A European Public Assessment Report (EPAR), however, issued by the European Medicines Agency, contained a safety analysis that some have used to suggest that there may be a difference between aflibercept and ranibizumab.²² What are your thoughts on this information?

Dr Duker: To me, there really is no safety difference, if you look at the totality of the EPAR data, as well as the data presented in the VIEW studies.^10,22 Most retina specialists feel that aflibercept, bevacizumab, and ranibizumab have similar systemic safety profiles.⁸ There are naturally going to be differences if we examine a subgroup of a subgroup, and thus are looking at small numbers. In the case of the EPAR dataset, it showed a higher rate of cerebrovascular events for older patients (≥85 years) with the use of aflibercept (20/283 [7.1%]) than with the use of ranibizumab (1/87 [1.1%]) after the first year of treatment.²² In the second year of treatment, the rate of cerebrovascular events among patients aged 85 years and older was 9.5% (27/283) in the aflibercept-treated group and 3.4% (3/87) in the ranibizumab-treated group.²² If you look at that same subset of elderly patients, they reported a higher rate of cardiovascular events with the use of ranibizumab than with aflibercept at year 1 (8/87 [9.2%] vs 8/283 [2.8%]) and at year 2 (9/87 (10.3%) vs 21/283 [7.4%]). With this subgroup of a subgroup analysis, I do not believe that aflibercept is more likely to cause a cerebrovascular accident than is ranibizumab, or that ranibizumab is more likely to cause myocardial infarction (MI) than is aflibercept.

Dr Kaiser: We recently conducted a safety assessment of aflibercept across all the phase 2 and phase 3 trials in AMD, DME, and central retinal vein occlusion (CRVO). The data have been presented, but not published. There were no statistically significant differences with respect to systemic safety events such as death and Antiplatelet Trialists’ Collaboration-adjudicated arterial thromboembolic events among aflibercept-treated groups and their comparators in the 8 trials that we assessed (CLEAR-IT 2, VIEW 1, VIEW 2, COPERNICUS, GALILEO, DA VINCI, VIVID-DME and VISTA-DME).²³

Dr Martin, on the basis of the large patient numbers in this safety assessment study, what is your opinion regarding how we are affecting the stroke rate of our patients by using these drugs? If there is an increased risk associated with the use of anti-VEGF therapy, what is the magnitude of that increase?

Dr Martin: I think that it is fair to say that any potential increase in stroke rate with the use of anti-VEGF therapy is low. I do not think we can ever say that there is no increased risk at all, because we are never going to do a randomized trial that is large enough to really capture that data. And so we are left with reviewing administrative databases. There are analyses, such as the one conducted by Campbell and colleagues,²⁴ that showed no changes in the rate of hospitalization for stroke among patients in Canada treated with bevacizumab or ranibizumab for AMD. There always will be some degree of uncertainty, but from my perspective, there are no differences in safety among these 3 drugs, according to the evidence presented in major trials. With respect to the use of anti-VEGF therapies as a class, I think it unlikely that there is a doubling of risk of major events such as MI and stroke. If there is any increase in risk, it must be very, very small.

Diabetic Macular Edema

Dr Kaiser: Diabetic macular edema has had some interesting developments as well. A lower dose of ranibizumab (0.3 mg) has been approved by the US Food and Drug Administration (FDA) for treatment of DME, accompanied by the 3-year results of the RISE and RIDE studies.²⁵ We now also have phase 3 results of VIVID-DME and VISTA-DME pertaining to the use of standard-dose aflibercept in the treatment of DME²⁶; this medication is currently being evaluated by the FDA for potential approval for DME. Bevacizumab and even pegaptanib continue to demonstrate efficacy in the treatment of DME in follow-up assessment studies.^27,28

One of the interesting things that many clinicians may not realize is that the 0.3-mg dose, not the 0.5-mg dose, of ranibizumab was approved for the treatment of DME. What are your thoughts regarding the selection of that particular strength for approval?

Dr Ho: The basic concept was to achieve the lowest dose that was still safe and efficacious. There were some numerical differences when the investigators used a Benefit Risk Action Team framework assessment of the different ranibizumab doses in this vasculopathic population of typically younger patients who may require longer courses of anti-VEGF therapy to preserve vision in the setting of persistent diabetes mellitus.²⁵

Dr Duker: In the pooled 3-year data for RISE and RIDE, the efficacy results were equivalent for the 0.3- and the 0.5-mg doses.²⁵ Given that patients with diabetes have an elevated risk for vascular complications, including stroke and MI, it was thought that the lower dose would reduce any risks potentially associated with systemic VEGF suppression. Interestingly, the 0.5-mg dose is the one approved in Europe for the treatment of DME.²⁹

Dr Kaiser: The data in RISE and RIDE also showed that patients who initially received sham treatments and crossed over to receive ranibizumab therapy were not able to match the visual acuity gains seen in those patients who were initially started on ranibizumab.²⁵

Unlike the situation in AMD, there are no direct comparison trials that have reported their results among the different anti-VEGF agents for DME. RISE and RIDE compared ranibizumab with sham injection,²⁵ and the DRCR.net (Diabetic Retinopathy Clinical Research Network) Protocol I compared ranibizumab in varying combinations with triamcinolone, laser therapy, and sham injection.³⁰ There have been no major comparison trials pertaining to the use of bevacizumab in DME, and the VIVID-DME and VISTA-DME studies compared aflibercept with laser.²⁶ DRCR.net is conducting a CATT-like treatment comparison trial, Protocol T, in which all the drugs except pegaptanib are being tested head-to-head, and those study results will soon be available.³¹ In current clinical practice, pegaptanib, bevacizumab, and ranibizumab are available. Dr Ho, do you find that there are any differences among these options for your patients with DME? Which one do you begin with, and is the treatment paradigm the same as it is for AMD?

Dr Ho: Because bevacizumab and ranibizumab are more commonly used, I will focus my comments on those treatments. Although bevacizumab and ranibizumab have similar efficacy in the treatment of AMD, I believe that ranibizumab has greater efficacy in the treatment of DME, according to my personal experience. I am looking forward to the results of the aforementioned DRCRnet Protocol T study, expected later this year, comparing bevacizumab, ranibizumab, and aflibercept for the treatment of DME.³¹ Ultimately, DRCR.net, RISE/RIDE, VIVID-DME, VISTA-DME, RESTORE, and RESOLVE all point toward a major paradigm shift to first-line anti-VEGF therapy for DME.^{25,26,30,32,33}

Regarding the issue of visual acuity and response to treatment, I want to draw your attention to an important point pertaining to anti-VEGF treatment gleaned from comparing the visual acuity curves between AMD and DME. The slope of the curves is usually quite steep over the first 4 months in the AMD trials, such as MARINA, ANCHOR (Figure 3), and VIEW^34–36; the slope is much more gradual for the DME studies, including RISE/RIDE (Figure 4), DRCRnet Protocol I, and VIVID-DME/VISTA-DME.^25,26,30 This difference in curve slopes highlights the point that with DME, one frequently needs to be persistent and consistent with therapy over many months in order to gauge the true level of functional response to anti-VEGF treatment. I believe that this persistent approach applies to RVO, as well.

Figure 3. BCVA curves for ANCHOR.³⁵ Reprinted with permission from The New England Journal of Medicine.

Figure 4. BCVA curves for RISE and RIDE.²⁵ Reprinted with permission from Elsevier.

Dr Duker: There are major differences between the treatment of AMD and DME. First, with DME, a little bit of fluid on the patient’s OCT does not necessarily equate with an elevated risk for adverse visual outcomes. Second, patients with DME require multiple injections in order to dry the macula, and it may take several injections to predict response to therapy. In patients with AMD, you can often predict response after 1 or 2 injections.

There are many good treatment alternatives for DME, including laser, steroids, and surgery. I do not have a set approach to the treatment of DME because of these other options.

Dr Kaiser: Dr Duker, has your sequence of treatment progression changed? We used to begin treatment for clinically significant macular edema with laser therapy. Has your DME treatment paradigm shifted at all?

Dr Duker: Yes. First of all, I am not treating non-foveal-involved macular edema with anti-VEGF. I am treating with focal laser photocoagulation; most retina specialists regard this as an effective, initial treatment modality for clinically significant macular edema that does not have foveal involvement.⁸ For all other patients, unless they have a mechanical issue such as an epiretinal membrane or taut posterior hyaloid, I start with an anti-VEGF agent like ranibizumab. For my patients who have a circinate or other leakage pattern that I believe is amenable to laser therapy, I will add laser.

I have an up-front discussion with patients about the need for multiple treatments, which may necessitate the use of different agents. If a patient is pseudophakic and does not respond to an initial course of anti-VEGF therapy, I will add steroid therapy. For patients who require monthly anti-VEGF injections, in whom I am not able to extend the treatment interval, I will use an off-label dexamethasone implant. This approach has allowed me to provide some of my patients 3 and 4 months between intravitreal injections. Singer and colleagues have examined the combination of anti-VEGF therapy and a dexamethasone implant in the treatment of recalcitrant edema secondary to RVO, and noted a reduced need for subsequent anti-VEGF therapy.³⁷ This same strategy has worked well in my patients with DME.

Dr Kaiser: Let us explore the idea of combination treatment with anti-VEGF and laser therapy. Several studies, such as READ-2 and RESTORE, have looked at the combination of laser and anti-VEGF therapy for DME in which no additional improvement in visual outcomes was demonstrated in the combination groups relative to the outcomes in the anti-VEGF monotherapy groups.^32,38 In these studies, the combination approach may have decreased the number of injections, but there was no additional improvement in visual acuity. How do you respond to these results?

Dr Duker: In terms of an overall population, I agree with the results, but I think there are individual eyes with individual patterns of leakage that still benefit from laser and so I offer it. For example, I have patients with 15-year follow-up after a single focal laser to circinate exudate who have required no further therapy.

Dr Kaiser: Let us continue exploring other combination approaches; Dr Martin, what are your thoughts on the incorporation of steroids into the treatment for DME?

Dr Martin: Corticosteroids are particularly effective at eliminating fluid.³⁹ I begin with an anti-VEGF agent, keeping in mind that it takes multiple anti-VEGF injections to determine therapeutic effect. If I am unable to eliminate the fluid and if I think that fluid is clinically important, I will probably select steroid therapy for a patient who is pseudophakic. If the patient is phakic, I will probably switch anti-VEGF agents before moving to a steroid.

Dr Ho: For pseudophakic patients, I view corticosteroids as a way to reboot the system. DME is a complex, multifactorial disease, involving VEGF, cytokines, and other inflammatory components. The steroids can provide a different mechanism to dry the macula, and a corticosteroid trial may lead to an entirely different pattern of requirements for anti-VEGF injections.

Dr Kaiser: The steroid studies demonstrated the greatest effect with the initial injection. Subsequent injections have a progressively diminished effect, so intermittent use of this therapy may be helpful in improving outcomes.⁴⁰

Dr Ho: A corticosteroid approved for DME would help patients. Some new phase 3 safety and efficacy data have been released, but are as yet unpublished, pertaining to the use of a dexamethasone implant for the treatment of DME.⁴¹

Dr Kaiser: How does surgery fit into the treatment paradigm, particularly for patients who do not have signs of traction or epiretinal membrane?

Dr Duker: Surgery is a fourth-line option for patients who do not respond to the other therapies that we have been discussing. The data indicate that anatomical outcomes are better than functional outcomes with surgical intervention, and I have found that to be true in my clinical experience. Although you can achieve a dry macula at least 50% of the time after removing the posterior hyaloid and/or epiretinal membrane, the visual results do not necessarily mirror these anatomic results.⁴²

But many of these patients have been treated for years without improvement in their macular edema. Had the surgery been done earlier, perhaps the macular edema would have been resolved earlier and the visual results would be better. I still reserve surgery as fourth-line therapy at this point in my practice.

Dr Ho: If there is a patient with a mechanical issue, I think that surgery is a reasonable option. I think that “the jury is still out” on the optimal timing of surgical intervention.

Retinal Vein Occlusion

Dr Kaiser: Let us discuss treatment options for CRVO. Dr Martin, what is your initial approach for these patients?

Dr Martin: I usually start with bevacizumab for the treatment of CRVO, as I have generally had excellent results with this option in my clinical experience. For those patients who are phakic and do not respond adequately, I will consider changing to a different anti-VEGF agent. If a patient is pseudophakic and does not respond adequately, I may still switch to a different anti-VEGF agent, but I will certainly give more serious consideration to the use of an intravitreal steroid.

Dr Ho: I tend to start with either ranibizumab or aflibercept, both of which are on-label for the treatment of CRVO. I think that there is an advantage in overall safety relative to bevacizumab, because the processes that are necessary for an on-label medication to be developed and commercially sold are more tightly regulated. We have limited prospective comparative data on these 3 agents in the treatment of retinal venous occlusive disease, but in my clinical experience, all can be highly effective. Although bevacizumab is widely used, there may be safety concerns related to the use of this compounded product because it is not subjected to the same level of scrutiny in the preparation process as is an on-label agent; and, there has been documented noncompliance with face-mask use hypothesized to be the source of contamination of compounded intravitreal bevacizumab resulting in several cases of endophthalmitis.⁴³

Dr Duker: Although there are no direct head-to-head comparisons among the 3 principal anti-VEGF agents—aflibercept, bevacizumab, and ranibizumab (pegaptanib, the fourth anti-VEGF agent, is available though not widely used)—I usually start with aflibercept therapy, which is on-label, for my patients with CRVO. In the absence of a direct study comparison, my evaluation of results among clinical trials looking at the anti-VEGF agents suggest to me that aflibercept is superior to ranibizumab for the treatment of CRVO.^44-46

Dr Ho: There may be an element of “apples to oranges” when you do cross-trial comparisons, though. For example, in the CRUISE study, which examined the safety and efficacy of ranibizumab,⁴⁴ the patients may have received prior treatment, and therefore had a longer duration of disease (4 months on average) at the beginning of the study than patients enrolled in COPERNICUS and GALILEO (which examined the safety and efficacy of aflibercept), in which the average disease duration was 2 months.^45,46 It is a subtle difference, but it might have an effect on data application to clinical practice.

Dr Duker: I agree that there may be differences, but when I look at cross-trial comparisons, there are some intriguing data. In CRUISE, 16.9% of patients treated with sham injection experienced ≥15 letters of improvement after 6 months, and 46.2% and 47.7% of those treated with ranibizumab, 0.3 mg and 0.5 mg, respectively, experienced that level of improvement.⁴⁴ In COPERNICUS, 12.3% of patients treated with sham injection experienced ≥15 letters of improvement after 6 months, and 56.1% of patients treated with aflibercept experienced that level of improvement.⁴⁵ In GALILEO, 22.1% of patients treated with sham injection had gained ≥15 letters after 6 months, compared with 60.2% in the aflibercept-treated group (Table 1).⁴⁶ If aflibercept and ranibizumab were equal in terms of efficacy, then the magnitude of improvement relative to sham treatment should have been similar. Again, this is crosstrial comparison, but I would rather rely on the use of crosstrial comparison to help shape my treatment approach for patients with CRVO than on my own clinical experience, because of the greater number of patients encompassed by clinical trials.

Table 1. Percent of CRVO Patients Experiencing ≥15 Letters Improvement After 6 Months⁴⁶

	Sham Injections	Anti-VEGF Injections	Magnitude of Improvement Relative to Sham
CRUISE (ranibizumab)	16.9%	~46.9%	~2.7 times
COPERNICUS (aflibercept)	12.3%	56.1%	~4.5 times
GALILEO (aflibercept)	22.1%	60.2%	~2.7 times

Dr Kaiser: In your clinical practice with ranibizumab, do you follow the protocol used in CRUISE, that is, the use of 6 injections before considering any change in therapy?

Dr Duker: I treat with the same therapy until the macula is dry or until there are 2 consecutive visits with no further improvement. At that point I have a discussion with my patient, in which we talk about subsequent treatment options, such as continuing monthly dosing, beginning a period of observation, or starting a treat-and-extend protocol.

Dr Kaiser: Dr Martin, you mentioned earlier that you begin with bevacizumab. What would prompt you to switch therapy? What are your criteria for making a switch?

Dr Martin: I will consider switching therapy if I do not achieve the degree of drying or visual acuity results that I want.

Dr Kaiser: When do you switch or add steroids?

Dr Martin: I will switch after an average of 4 to 6 injections, according to the visual acuity response curves seen in phase 3 trials such as CRUISE, COPERNICUS, and GALILEO.^44-46 Changes in visual acuity tend to be slow and to evolve over the course of the first year; I am not too quick to change therapy. I tend to switch or change treatment options more often for patients who require every-4-weeks injections for this predominantly VEGF-driven disease. As seen in SCORE and GENEVA, corticosteroid injection or the use of a dexamethasone implant can be quite effective in drying out the macula and resetting the responsiveness to pharmacologic treatment, and use of these procedures has allowed me to extend my treatment intervals.^47,48

What are the panel’s thoughts on the utility of widefield imaging and peripheral photocoagulation therapy in situations in which you have difficulty completely drying out the macula or extending treatment intervals?

Dr Duker: I do not object to the concept of widefield imaging and peripheral photocoagulation, but there are not a lot of data out there to support its use. One small prospective study conducted by Spaide and colleagues failed to show any influence of peripheral panretinal photocoagulation (PRP) on injection frequency or improved visual acuity in patients with CRVO who had been treated with ranibizumab.⁴⁹ We have all been in situations in which we have “thrown the kitchen sink at” recalcitrant edema, and still do not achieve resolution. In that situation, using targeted peripheral PRP would be low-risk and possibly worth trying.

Dr Kaiser: At the Cleveland Clinic Cole Eye Institute, we have done quite a bit of targeted peripheral PRP based on widefield angiography images, and I have been underwhelmed by the results. I have looked at fluorescein angiograms and seen massive amounts of peripheral nonperfusion, which I treated with PRP. Despite the use of laser, my number of anti-VEGF injections in these patients has not dropped appreciably. Even in situations in which there is early, peripheral neovascularization, in my personal experience, the results have been underwhelming.

Dr Martin: My experience with patients who have nonperfusion on fluorescein imaging has been similar. I have not been able to reduce the anti-VEGF injection frequency.

Dr Duker: I think there is a reason for this lack of efficacy. Patients may have areas of peripheral nonperfusion on fluorescein angiography, but I think that VEGF is being released diffusely through the retina. Even though you are targeting areas that are nonperfused, those may not be the areas from which all the VEGF is coming. I am more impressed with steroids than with targeted peripheral PRP.

Dr Ho: I do not use targeted PRP because we do not have the widefield imaging technology; the empiric results have been underwhelming for me, and steroids can be very effective in treating patients with macular edema due to CRVO who do not respond to anti-VEGF therapy.⁵⁰

Dr Kaiser: Lastly, let us discuss branch retinal vein occlusion (BRVO). Is your treatment approach for BRVO the same as it is for CRVO? Does anyone here follow the Branch Vein Occlusion Study recommendations and begin with an observation period,⁵¹ or do you tend to start anti-VEGF therapy right away?

Dr Duker: If the patient has macular edema and decreased visual acuity, I begin anti-VEGF therapy; a recent study by Tan and colleagues showed that, compared with standard-of-care grid laser, anti-VEGF therapy provides sustained improvements in visual gain and anatomic outcomes.⁵² I also consider adding laser therapy once any existent hemorrhages resolve; there have been some prospective studies that have looked at the combination of laser and anti-VEGF therapy for the treatment of BRVO, in which patients who received this combination had a lower number of repeat anti-VEGF injections in follow-up.^53,54 Targeted laser therapy for BRVO can be quite effective.

Dr Martin: I agree with Dr Duker’s approach.

Dr Ho: I also tend to start therapy right away for patients who have macular edema and decreased visual acuity.

Dr Kaiser: I believe that we all can concur on this matter. The treatment paradigm for BRVO continues to evolve, and results of the recently presented VIBRANT study, comparing the use of laser therapy and monthly aflibercept for the treatment of macular edema following BRVO,⁵⁵ may further shape our treatment approaches. All the retinal diseases that we have discussed herein are topics of active, ongoing investigation, and we must be vigilant as new evidence comes to light, in order to create optimal treatment protocols for our patients. Fortunately, we do have an increasing armamentarium to aid in the management of patients with AMD, DME, and RVO, as evidenced by findings of some of the recent clinical investigations pertaining to retinal diseases, summarized in Table 2 following.

Table 2. 2013 Studies in the Management of Patients With AMD, DME, and RVO and Corresponding Top-Line Results

Study	Comparison	Top-Line Results
Age-related Macular Degeneration
LUCAS6	Multicenter randomized trial comparing bevacizumab vs ranibizumab for treatment of wet AMD, using inject-and-extend protocol	• 1-year outcomes showed that mean number of treatments was 8.8 for bevacizumab, 8.0 for ranibizumab (P=.002), with equivalent effects on visual acuity • 6 patients (2.8%) in ranibizumab group had nonfatal MI vs 0 in bevacizumab group • Nonfatal strokes occurred in 1.4% of patients receiving ranibizumab vs 0.9% of patients receiving bevacizumab • 1 vascular-related death in bevacizumab group
SECURE20	Multicenter open-label phase IV extension study to assess long-term safety of 0.5-mg ranibizumab in treatment of wet AMD	• Ranibizumab given according to flexible, visual acuity-guided protocol, was well tolerated over 2 years • On average, patients lost BCVA relative to baseline
VIEW Extension Study18	Analysis of effect of different medication dosing regimens on persistent macular fluid in patients with wet AMD	• Patients who had persistent fluid following 3 monthly loading doses had better visual acuity (P<.05) with aflibercept dosed every 4 weeks than with aflibercept dosed every 8 weeks or ranibizumab dosed every 4 weeks
Diabetic Macular Edema
RISE/RIDE 3 year results25	Multicenter randomized, double-masked, sham injection-controlled (for 2 years) studies comparing monthly 0.3-mg to 0.5-mg ranibizumab for treatment of DME	• VA outcomes seen at 2 years with ranibizumab-treated groups were maintained through year 3 • Patients who made transition from sham treatment to 0.5-mg ranibizumab in third year did not see same amount of improvement as patients who were initially treated with ranibizumab • Incidence of SAEs potentially related to systemic VEGF suppression was 19.7% in 0.5-mg group, 16.8% in 0.3-mg group
VIVID-DME and VISTA-DME26	Multicenter randomized double-masked studies comparing monthly aflibercept vs bimonthly aflibercept (after 5 monthly treatments) vs laser for treatment of DME with central involvement	• 31.1% to 41.6% of aflibercept-treated eyes gained ≥15 letters at 1 year vs 7.8% to 9.1% of laser-treated eyes • Overall incidence of ocular, nonocular, and serious adverse events was similar for both laser- and aflibercept-treated groups
Dexamethasone implant study41	3-year multicenter masked randomized controlled trial to assess safety and efficacy of dexamethasone in treatment of DME	• Significantly more patients treated with dexamethasone had ≥15 letters improvement than those treated with sham • Mean number of treatments over 3 years: 4.1 for 0.7 mg, 4.4 for 0.35 mg, and 3.3 for sham • Rates of cataract-related AEs: 67.9%, 64.1%, and 20.4% for 0.7 mg, 0.35 mg, and sham, respectively • IOP increases generally controlled with medication or no therapy
Retinal Vein Occlusion
VIBRANT55	Multicenter double-masked trial to compare monthly aflibercept (2 mg) vs laser for macular edema secondary to BRVO	• 53% of patients receiving monthly aflibercept gained ≥15 letters from baseline at week 24 vs 27% who received laser • Mean BCVA improvement was 17.0 letters in aflibercept group vs 6.9 letters in laser group (P<.0001) • SAE incidence was 9.9% in aflibercept group vs 9.8% in laser group • 1 death and 1 APTC-defined event occurred, both in laser group
AEs=adverse events; AMD=age-related macular degeneration; APTC=Antiplatelet Trialists’ Collaboration; BCVA=best corrected visual acuity; BRVO=branch retinal vein occlusion; DME=diabetic macular edema; IOP=intraocular pressure; SAEs=serious adverse events; VA=visual acuity; VEGF=vascular endothelial growth factor.

Trial Abbreviations Used

ANCHOR	Anti-VEGF Antibody for the Treatment of Predominantly Classic Choroidal Neovascularization in Age-Related Macular Degeneration
BRAMD	Netherlands study to compare the effectiveness and costs of bevacizumab and ranibizumab in patients with exudative age-related macular degeneration.
CATT	Comparison of Age-related Macular Degeneration Treatments Trials
CLEAR-IT	CLinical Evaluation of Anti-angiogenesis in the Retina Intravitreal Trial
COPERNICUS and GALILEO	Vascular Endothelial Growth Factor [VEGF] Trap-Eye: Investigation of Efficacy and Safety in Central Retinal Vein Occlusion [CRVO]
CRUISE	Ranibizumab for the Treatment of Macular Edema after Central Retinal Vein OcclUsIon Study: Evaluation of Efficacy and Safety
DA VINCI	DME And VEGF Trap-Eye: INvestigation of Clinical Impact
GEFAL	French Evaluation Group Avastin Versus Lucentis
GENEVA	Global Evaluation of Implantable Dexamethasone in Retinal Vein Occlusion With Macular Edema
HARBOR	pHase III, double-masked, multicenter, randomized, Active treatment-controlled study of the efficacy and safety of 0.5 mg and 2.0 mg Ranibizumab administered monthly or on an as-needed Basis (PRN) in patients with subfoveal neOvasculaR age-related macular degeneration
HORIZON	Extension study to evaluate the safety and tolerability of ranibizumab in subjects with choroidal neovascularization secondary to AMD or macular edema secondary to RVO
IVAN	Inhibit VEGF in Age-related Choroidal Neovascularization
LUCAS	Lucentis Compared to Avastin Study
MANTA	Multicenter Anti-VEGF Trial in Austria
MARINA	Minimally Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular Age-Related Macular Degeneration
READ-2	Ranibizumab for Edema of the mAcula in Diabetes
RESOLVE	Phase 2 study evaluating safety and efficacy of ranibizumab in diabetic macular edema
RESTORE	Phase 3 study evaluating ranibizumab monotherapy or combined with laser vs laser monotherapy for diabetic macular edema
RISE and RIDE	Phase 3 studies evaluating efficacy and safety of intravitreal ranibizumab in diabetic macular edema
SCORE	Standard Care vs. Corticosteroid for Retinal Vein Occlusion
SECURE	Long-Term Safety of Ranibizumab 0.5 mg in Neovascular Age-related Macular Degeneration
VIBRANT	Phase 3 study of aflibercept in treating macular edema due to branch retinal vein occlusion
VIEW 1 and 2	VEGF Trap-Eye Investigation of Efficacy and Safety in Wet AMD
VISTA-DME and VIVID-DME	Double-masked, randomized, active-controlled, phase 3 studies of the efficacy and safety of intravitreal administration of VEGF Trap-Eye in patients with diabetic macular edema

References

Post Test

To obtain AMA PRA Category 1 Credit™ for this activity, complete the CME Post Test online at http://bit.ly/retinal2014 and receive an instant certificate of credit upon successful completion of the post test and evaluation.