June 14, 2018
16 min read

Treatment of DME poses complex choices, strives to overcome limitations

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

Click Here to Manage Email Alerts

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

Diabetic macular edema remains a major concern despite the availability of technology for potentially earlier diagnosis and better options for treatment. The 2016 WHO Global Report on Diabetes showed that 422 million adults currently live with diabetes and that prevalence has almost quadrupled since 1980. This is mainly due to the dramatic rise in type 2 diabetes, which accounts for 90% to 95% of the cases. Inactivity, diet, overweight and obesity are the main causative factors.

As a consequence, the prevalence of diabetic retinopathy (DR) and DME is also expected to rise.

“In the U.S., about three-quarters of a million Americans have DME, and with the current trend, we can expect to see more and more cases in future years,” OSN Retina/Vitreous Board Member Carl D. Regillo, MD, FACS, said.

The good news is that the urgency is making research move faster. Current treatment options have been widely investigated, compared and combined, and new options are being tested in the search for increased efficacy and durability, and consequently lesser burden for the patient.

“Treatment options improved immensely over recent years, and that has been driven by the emergence of anti-VEGFs. In a dozen years, they have turned everything around. The AAO Preferred Practice Pattern recommends them as first-line treatment for center-involved DME, with or without adjunct focal laser treatment. While previously focal laser was the primary treatment, anti-VEGF alone can effectively manage about three-quarters or more of center-involving DME cases,” Regillo said.

If response is poor, steroids may be attempted as a second-line treatment, alone or in combination with an anti-VEGF.

According to Carl D. Regillo, MD, FACS, current options for DME have been widely investigated, compared and combined, and new options are being tested in the search for increased efficacy and durability, and consequently less burden for the patient.

Source: Carl D. Regillo, MD, FACS

“Focal laser is more of a potential backup treatment for most of our patients, and the majority of retinal specialists are using it less and less as time goes on,” he said.

Anti-VEGFs and their administration

The DRCR.net Protocol T study has been pivotal in providing useful guidelines on when and how to choose among anti-VEGF agents, providing comparisons of Avastin (bevacizumab, Genentech), Lucentis (ranibizumab, Genentech) and Eylea (aflibercept, Regeneron).

“The study showed that in patients with baseline visual acuity (VA) between 20/32 and 20/40, any of these agents can be chosen as they lead to a similar amount of letter gain of approximately 7 to 8 letters. However, for those with VA 20/50 or worse, Eylea did better than Avastin at 2 years with gain of 18 letters vs. 13 letters. Lucentis at 1 year was inferior to Eylea, but by 2 years it had caught up and there was no statistically significant difference. In terms of drying effect measured on OCT at 2 years, both Eylea and Lucentis were better than Avastin for the moderate vision loss group, while only Eylea was statistically significantly better than Avastin for those with 20/50 or worse VA at baseline. However, all three drugs were able to reduce the amount of macular edema,” OSN Retina/Vitreous Board Member and OSLI Retina Board Member Judy E. Kim, MD, said.


In summary, the two approved drugs are preferable, but off-label bevacizumab is still a good option if necessary. If all options are available, aflibercept is better for eyes with poor vision and a thicker retina, as it seems to improve visual acuity and dry the macula faster, she said.

Judy E. Kim

Anti-VEGFs for DME are typically administered monthly until the macula is dry, and either as-needed or treat-and-extend treatment can be used thereafter.

“Although I use the treat-and-extend approach for neovascular AMD, I favor the PRN in this setting mainly because recurrence of DME is often unpredictable and many patients eventually are able to come off treatments for extended periods of time, if not indefinitely in some cases. Another difference between neovascular AMD and DME treatment is that the monthly anti-VEGF injection induction phase to get the macula dry is typically much longer for DME. It may take 6 to 12 or more months of monthly treatment to get the maximal drying effect with DME as opposed to 3 to 4 for neovascular AMD,” Regillo, who is also an OSLI Retina Board Member, said.

DRCR.net Protocol I and Protocol T also suggested as-needed with quarterly monitoring as a treatment schedule during the maintenance phase. However, recent studies such as RETAIN and TREX-DME showed that treat-and-extend can be equally effective.

Anti-VEGFs work well if used early in the course of disease, aggressively and possibly long term, Kim said.

“I stress with my patients that if we work together at fighting DME aggressively early on with systemic blood sugar control and monthly injections of anti-VEGF even for 6 months to 1 year, Protocol I and T have shown that we are likely to need fewer injections in the second year, even less in the third year and maybe no injection in year 4 or 5. While intravitreal injections for management of age-related macular degeneration may be lifelong, with DME I try to motivate and encourage my patients that, for many, there is light at the end of the tunnel,” she said.

However, because the response to anti-VEGF is variable among patients, it might be necessary to switch to another anti-VEGF, which might be beneficial in some cases, or to steroids.

“If after about four to six or so monthly injections I am not satisfied with the results, and I am using either bevacizumab or ranibizumab, I might first switch to aflibercept. If I have used aflibercept from the start, I switch to intravitreal steroids,” Regillo said.


Steroids as a second-line option

In more severe, chronic DME, intravitreal steroids are an alternative second-line option when patients show a poor response to anti-VEGFs. They can be used at any stage of DME, but caution is advised in patients with glaucoma, especially those with a cup-to-disc ratio greater than 0.8, or hard-to-control ocular hypertension.

“I tend to use steroids for more severe cases of DME and in patients with chronic DME,” Sophie J. Bakri, MD, said. Bakri is an OSLI Retina Board Member.

The advantage of steroids is that they control the inflammatory component of DME, which is prevalent in the most severe and chronic forms of the disease.

“Sometimes in very severe cases I have used steroids in combination with anti-VEGF to control the edema and quieten the inflammation. This helps to obtain a more dramatic DME reduction initially,” Bakri said.

Sophie J. Bakri

Both Ozurdex (dexamethasone intravitreal implant, Allergan) and Iluvien (fluocinolone acetonide intravitreal implant 0.19 mg, Alimera Sciences) are approved to treat DME.

“Those may be patients in which I wish to use other treatment modalities as well. For example, to a newly presenting poorly controlled diabetic patient with extensive DME, I would recommend medical control of glucose and any associated hypertension or medical comorbidities. I may then treat with dexamethasone and anti-VEGF. Since the presentation to me is acute, I would want to see how the patient does before committing to a 2+ year implant,” Bakri said.

For patients with chronic DME who require multiple anti-VEGF injections or steroid implants on a continued basis, she would consider the long-term Iluvien implant that was shown in studies to work best in chronic cases with a sustained effect of 1 to 3 years.

The MEAD trial, published in 2014, was the main study comparing Ozurdex with sham over a 3-year period. The implant met the primary efficacy endpoint for best corrected visual acuity improvement of more than 15 letters with three to four injections over 3 years.

While studies said that the effect of Ozurdex is sustained over 6 months, real-life experience showed that the effect wears off after 4 months in most patients.

“I reinject at 4 months, monitoring the patient regularly. If I see that the edema is still there after two or three injections, I might consider Iluvien,” Bakri said.

“With Ozurdex, I see patients 6 weeks after implantation, checking for IOP spikes, and then 6 to 8 weeks later to assess by OCT if the effect is still there. With Iluvien, I see them 6 to 8 weeks after implantation and then every 3 or 6 months depending on the case,” she said.


Timing of switch

“In the U.S., anti-VEGF is first line; then we have steroids as a second option if patients don’t respond. In other parts of the world, including Europe, the switch to steroids is done a lot more quickly, possibly because of the durability of steroids compared to anti-VEGF among other reasons. Here in the U.S., we seem to be more concerned about the side effects of steroids,” Kim said.

According to Anat Loewenstein, MD, switching early to steroids, even after three injections, is beneficial and might save vision if patients do not respond to anti-VEGF.

“We should be able to predict as early as possible if a patient’s response to treatment is going to be good and switching to an alternative treatment in case the answer is no. We need a lot more data, but what comes out loud and clear is that if a patient does not respond to a specific therapy, then we definitely need to start thinking about switching him or her to another treatment,” she said.

Most physicians who start with anti-VEGF would refrain from the fifth and sixth injection only if there was no response after the fourth, but she sees no harm in switching to dexamethasone after three injections if the patient does not respond.

Anat Loewenstein

For Loewenstein, the fear of development of cataract is not an absolute contraindication to using dexamethasone, and she uses it also in phakic patients.

“In pseudophakic patients it is easier to make this decision, but I don’t think that being phakic is a contraindication. If there are signs of incipient cataract, which is not uncommon in this population, surgery would have to be done anyway not too far in the future, so it is completely fine to start with Ozurdex, even as first-line therapy. However, in phakic patients with a completely clear lens, anti-VEGF should be first line whenever possible,” Loewenstein said.

Steroids as first-line therapy

Both Loewenstein and Kim agreed upon specific situations in which steroids could be considered as first-line therapy.

“There are patients who, for a number of different reasons, may not be able to come for frequent follow-up, monitoring and injections. According to Protocol I and T, anti-VEGF requires five to six monthly injections, and if the patient cannot comply with this schedule, there is no chance of success. If I predict that compliance is going to be low, I would rather use Ozurdex because it lasts for at least 3 months and IOP can be monitored by the local ophthalmologist,” Loewenstein said.


“Patients come from all over, may have to drive hours, may have problems with transportation or may have health-related issues. Not all patients require injections on a monthly basis in the long term but some do, and compliance may become a problem. If these patients are pseudophakic, I may propose Ozurdex to start, and if it shows a good response without IOP issues, then move on to Iluvien,” Kim said.

A second category in which both specialists agree involves patients who have had stroke or myocardial infarction. Although there is no evidence that anti-VEGF causes thrombotic events, caution should be preferred.

“There will never be a study big enough to prove it, but meta-analysis of existing studies suggests that there might be an association. I would suggest not taking the risk and give steroids as first line,” Loewenstein said.

“If I have a patient with a recent history of stroke or myocardial infarction, I am careful with anti-VEGFs. Since DME can be tolerated to some degree, I prefer to skip an injection or consider steroid injection at that visit,” Kim said.

Another population could be women who are pregnant or breastfeeding, as there is a small chance that anti-VEGF might affect development of the fetus or infant.

“None of these clinical trials would be done in women who are pregnant or nursing so, for the sake of safety, this would be a set of population where I might consider using intravitreal steroids if needed rather than anti-VEGF,” Kim said.

Patient selection is key

Choices should ultimately be made on a case-by-case basis and through careful patient selection, Bakri said.

“For example, I would be cautious in using steroids in a patient with uncontrolled glaucoma. I would exhaust other options before using steroids, and if I have to, I would consult with the glaucoma specialist to make sure all glaucoma options are maximized,” she said.

For patients with chronic DME, in which it is clear multiple anti-VEGF injections or steroid implants would be required on a continued basis, she would consider the long-term fluocinolone acetonide implant.

“Iluvien has been associated with more complications, but the alternative would be to inject multiple anti-VEGF and then Ozurdex every 3 or 4 months and that would cause the same complications. I balance risks and benefits and treat accordingly,” Bakri said.

“If needed, we sometimes have to choose the lesser of two evils. Even if steroids may cause IOP issues, controlling IOP with pressure-lowering medications or even surgery would be better than letting the retina permanently deteriorate with chronic DME. Likewise with steroid-induced cataract. We develop cataract anyway as we age. We don’t want to cause it in a young person, but if it is a choice between needing to operate on a steroid-induced cataract vs. having permanent damage to the macula from DME, then we have to choose the steroid option,” Kim said.


Biomarkers of response to therapy

There is evidence that approximately 50% of patients maybe be poor responders to anti-VEGF, but the definition of what might be a poor response or nonresponse to the therapy is still unclear.

“Is it based on VA not improving or on OCT thickness not improving? In Protocol T, after at least six monthly injections, the aflibercept group had 16% of eyes with vision 20/32 or worse and 32% with persistent DME on OCT. The ranibizumab group had 27% with vision reduction and 41% with persistent DME. Finally, bevacizumab had 39% with persistent vision reduction and 66% with persistent DME. The ranibizumab group in Protocol I was similar to what was found in Protocol T; there were 32% with reduced vision and 40% with persistent DME. However, remember that anti-VEGFs showed benefit overall in both studies,” Kim said. What this suggests, she said, is that “anti-VEGFs are great in improving VA and drying up the retina in most, but not for everyone.”

“On the other hand, in the ‘real world’ outside of clinical trials, studies are showing that we may not be treating long enough and frequently enough with anti-VEGF agents, leading to an undertreatment. Fewer treatments with anti-VEGF may result in poorer visual and anatomic outcomes compared to clinical trials. It is possible that these patients might have responded to treatments if given more injections. We have to be careful in defining poor responders,” Kim said.

Biomarkers of response to therapy are an important and lively area of research currently, although nothing has yet been proven in large multicenter trials.

“The best biomarker we have so far is baseline VA. The better VA is at baseline, the lesser the improvement because of the ceiling effect but the better the final VA. Studies are also looking into potential correlations between specific phenotypes and the better response to one agent over another,” Loewenstein said.

Other biomarkers that have been proposed are the presence of subretinal fluid and the presence of hyperreflective dots, which seem to be implicated in a better response to dexamethasone as compared with anti-VEGF.

“It is an interesting hypothesis, but it has never been proven in large multicenter trials,” Loewenstein said.

Another important factor is chronicity. If the patient has chronic disease and a history of several treatment attempts, the chances to improve vision are much lower.

“Chronic disease is often correlated with disorganization of the retinal inner layers (DRIL), which is strongly associated with low visual acuity improvement. If after treatment DRIL decreases, the chances for visual acuity improvement tend to increase,” Loewenstein said.


Combination therapy

In Protocol U, patients with persistent edema and loss of vision after 6 months to 1 year of monthly anti-VEGF treatment were randomly assigned to receive dexamethasone as an adjunct to continued ranibizumab. At 6 months, the ranibizumab-only group had an increase of 3 letters and the combination group had an increase of 2.7 letters.

“If we look at the overall results, the combination had no additional effect. However, the subgroup of pseudophakic patients did better with the combination than with ranibizumab monotherapy. In addition, a significant number of eyes in the entire cohort showed flat retinas on OCT. This might suggest that some of the patients did not improve VA because of cataract,” Loewenstein said.

Another possible caveat of Protocol U was that patients received the steroid after a fairly long course of ranibizumab injections, which might depotentiate the efficacy of steroids.

“The problem is that apart from Protocol U, we don’t have a large multicenter trial. An interesting experience comes from Germany, where combination is used by some specialists as first line with good results,” Loewenstein said.

“Personally, I do not use combination upfront, with steroid and anti-VEGF at the same time. But if steroids are first line and don’t work, after a month I might add an anti-VEGF, and this can be looked at as a combination because the steroid is still onboard,” she said.

Laser still an option

Lasers still play an important role in the treatment of DME and DR, according to Elias Reichel, MD.

“Typically we use laser in non-center-involving DME and in patients who don’t respond to anti-VEGF therapy after six monthly injections. There is still some question on what the best timing is, but if the patient is not sufficiently responding to anti-VEGF, it is reasonable to combine it with laser,” he said.

In his opinion, combination with laser should have priority over a switch to steroids.

“With corticosteroids, there are issues with IOP. It is yet another injection, and then if you look at the results of Protocol U, there appears to be no significant advantage in adding corticosteroids to anti-VEGF,” he said.

Protocol T used conventional thermal laser, but new systems such as the Iridex IQ 577 with MicroPulse mode and subthreshold effect offer the potential to treat in the fovea without causing damage.

Laser treatment has the advantage of lasting for many months but the disadvantage of a rather slow response.

“It can be frustrating for physicians who are used to anti-VEGFs and see a response within a week or two. However, it is a noninvasive procedure and a good option for those patients who get terribly anxious about injections. They may be few, but there are some who don’t want injections,” Reichel said.


Reichel uses laser in all of his patients with non-center-involving DME, which represents only 3% to 5% of his cases, and as a combination therapy, which is one out of three patients with center-involving DME.

“Not many young specialists use laser nowadays. They are more used to injections. But the older ones, over the age of 45, are familiar with laser, and I would say that probably about half of us use laser for DME,” Reichel said.

Unmet needs and new options

There are unmet needs for DME, mainly related to the high rate of poor responders and the burden of frequent injections.

“There are still patients that have suboptimal improvement, so better efficacy is still an unmet need. Durability is definitely an issue. Anti-VEGF therapy works very well on average, but it takes long-term treatment and frequent administration. We want a treatment that lasts longer and works better and faster. We want to prevent recurrences, too,” Regillo said.

Current options have inherent limitations and are not entirely sustainable, according to Pravin U. Dugel, MD, who is an OSN Retina/Vitreous and OSLI Retina Board Member.

“Very few patients can bear the burden of being treated every month for years and years, so sustainability is a big problem. The other big problem are those patients, approximately 50%, who initially are poor responders and go through several attempts of continuing, changing or combining therapies because we don’t know what may work,” he said.

However, research is moving fast, and there are many exciting new potential treatments in the pipeline, including a new generation of anti-VEGF drugs, the most prominent of which is brolucizumab (RTH258, Novartis). It is a single-strand antibody fragment, the first of its kind, with a much greater concentration than any of the existing anti-VEGFs and therefore potentially greater efficacy and durability.

“Brolucizumab has met its primary endpoint in neovascular AMD, and I am sure that it will be studied in DME and DR as well,” Dugel said.

“The other one is abicipar pegol (DARPin, Allergan), which according to the manufacturers could potentially offer dosing every 3 months. Again, studies on DME have not started yet but will most likely be the next step, depending on the results of AMD,” he said.

Among combination drugs, RG7716 from Genentech/Roche seems to be the most promising, according to Dugel.

“It is a bispecific drug, combining anti-vascular endothelial growth factor A (anti-VEGF-A) and anti-angiopoietin-2 (anti-Ang-2). Results are very encouraging,” he said.

Nesvacumab, an Ang-2 antibody co-developed in combination with aflibercept by Regeneron and Bayer, showed a biological signal, but not large enough to move into phase 3 testing. Another interesting new option awaiting results is OPT-302 (Opthea), a VEGF-C and VEGF-D inhibitor used in combination with anti-VEGF-A.


Most intriguing is AKB-9778 by Aerpio Therapeutics, a Tie2 activator that may open a new pathway into alternative ways of administration.

“It is a very exciting new option because it is administered subcutaneously specifically for DR. By targeting DR, we may be able to prevent complications of DR very early, in what is hopefully a new, less invasive and more sustainable way, treating both eyes simultaneously,” Dugel said.

Treating DME in the early stages, or rather treating DR to prevent progression to DME, is an important and attractive field for drug developers, he said.

There are also a number of sustained delivery systems currently under investigation. Furthest along in development is the Ranibizumab Port Delivery System (RPDS, Genentech). It is currently being tested in neovascular AMD in the phase 2 LADDER trial. There is hope that the same technologies might be beneficial for managing DME.

“Long-acting delivery could potentially be a winning strategy also to treat DR with or without DME, preventing progression to more severe levels of DR,” Regillo said.

“We are heading in the direction of treating DR before DME or PDR develops, and anti-VEGFs with long-acting delivery look to be a promising strategy to achieve that goal,” he said. – by Michela Cimberle

Disclosures: Bakri reports she is a consultant for Allergan, Genentech and Novartis. Dugel reports he is a consultant for Alcon, Allergan, Genentech, Regeneron, Roche, Alimera Sciences, Aerpio Pharmaceuticals and Opthea. Kim reports she is a consultant for Genentech, Alimera Sciences and Notal Vision. Loewenstein reports she is a consultant for Allergan, Novartis, Bayer and Notal Vision. Regillo reports he is a consultant for Genentech, Regeneron, Allergan, Alcon and Novartis. Reichel reports he is a consultant for Iridex, Regeneron and Genentech.

Click here to read the POINTCOUNTER, "Should an early switch to intravitreal steroids be considered in DME patients with a suboptimal response to anti-VEGF therapy?"