RPS Recap: Progress in the Treatment of Diabetic Eye Disease

First in a series recapping research and analysis presented at our annual meeting


Progress in the Treatment of Diabetic Eye Disease

The first in a series of articles recapping research and analysis presented at our annual meeting.


Last month's Retinal Physician Symposium, held April 21 to 24 at the Atlantis Paradise Island Resort in the Bahamas, brought together many innovators at the forefront of retina research, who shared their insights and expertise with those in attendance — and, now, our readers as well. This month we begin a series of articles detailing the clinical pearls from their lectures, and additional findings are available at

The morning of the third day of the Retinal Physician symposium was dedicated to discussing research into retinal complications of diabetes. We offer summaries of the presentations here.


Harry W. Flynn, Jr., MD, who is is professor and J. Donald M. Gass Distinguished Chair of Ophthalmology at the Bascom Palmer Eye Institute in Miami, opened the session with a talk on the use of spectral-domain optical coherence tomography in the management of diabetic retinopathy, particularly diabetic macular edema. Dr. Flynn began by giving an overview of how SD-OCT aids retinal physicians in both evaluating the pathology of diabetic macular disease and in monitoring patient response to laser treatment, intravitreal pharmacotherapies, and vitrectomy.

Dr. Flynn then turned to slides of a patient with DME, indicating how SD-OCT enabled detection of significant edema involving the fovea, as well as how the retinal thickness map allowed him to determine the extent of the edema. Dr. Flynn compared the current diagnostic abilities provided by SD-OCT with the earlier ETDRS standards for clinically significant macular edema, which used contact-lens biomicroscopy (Figure 1). “Now we have technology to supplant our subjective analysis,” he said.

Figure 1. Clinically significant macular edema used to be measured with contact-lens biomicroscopy.

Pointing to a recent study on patient response to focal/grid photocoagulation by the Diabetic Retinopathy Clinical Research Network (DRCR) published last year in Retina, Dr. Flynn pointed out that, thanks to OCT, we are now more aware of the course of response to laser in patients than was previously possible. This is in sharp contrast to response rates measured by the ETDRS and reported in Archives of Ophthalmology 25 years ago.

Furthermore, Dr. Flynn pointed out that, in the case of a 68-year-old male patient with DME, SD-OCT not only assisted with efficacious treatment with grid laser, but it also provided him with the map for retreatment, should retreatment be needed.

Turning to the DRCR Network's Protocol B — monotherapy with hydrogel-based intravitreal triamcinolone acetonide (IVTA, 1 mg or 4 mg) compared to laser for DME. Using OCT, central subfield thickness was measured, and Dr. Flynn pointed out how 4 mg IVTA effected the greatest mean reduction in central thickness at four months, compared to the lower dose of IVTA and laser.

However, these patients tended to stabilize, while “good old-fashioned laser” continued to improve out to 24 months. Comparing visual acuity, Dr. Flynn showed how laser was similarly more effective than 4 mg IVTA over the long term.

“What about patients like this?” Dr. Flynn asked, turning to slides of an 82-year-old woman with DME and a visual acuity of 20/400 (Figure 2). “This is the reality of our clinical practice every day. Do you think laser would help this patient?” The patient had undergone panretinal photocoagulation, and had a macular thickness of over 500 μm. Treated with IVTA, this patient had impressive improvement of edema but no improvement in visual acuity.

Figure 2. Diabetic macular edema in an 82-year-old female patient. Her visual acuity was 20/400.

Looking at the DRCR Network's results for Protocol B, Dr. Flynn focused on ≥10-letter worsening in visual acuity at two years. In eyes with 20/200 or worse visual acuity at baseline, 17% of patients in the laser group and 15% of patients in the 1 mg IVTA group got worse, while none of the patients in the 4 mg IVTA arm worsened. Conversely, 77% of patients in the 4 mg IVTA arm experienced ≥10 letters of improvement in visual acuity, compared to 42% of laser-treated patients and 46% of patients treated with 1 mg IVTA. Conceding that the data lacked statistical significance, Dr. Flynn nevertheless stated his preference for pharmacotherapy in patients with this severe a level of DME and 20/200 or worse visual acuity.

There are alternative therapies, Dr. Flynn said, mentioning alternative preparations of IVTA, as well as intravitreal implants containing corticosteroids such as fluocinolone or dexamethasone. These agents have either recently completed randomized clinical trials or soon will complete them. Dr. Flynn mentioned in particular Alimera's Iluvien fluocinolone acetonide implant, which appears promising despite the downside of glaucoma and cataracts. Dr. Flynn also discussed anti-VEGF agents under investigation for DME (DRCR Protocol I).

He then turned to discussing proliferative diabetic retinopathy, which, he said, can also be tracked using SD-OCT. He began with an SD-OCT image from a patient who presented with marked vision loss in the right eye.

Dr. Flynn believed the patient might have been suffering from epiretinal membrane, but, as he said, “OCT told the story.” Instead, the patient was diagnosed with a tractional retinal detachment (Figure 3). With SD-OCT, it is also possible to image the optic nerve head, Dr. Flynn said. Postoperatively, retinal reattachment was tracked with SD-OCT, and foveal contour returned.

Figure 3. SD-OCT images of a patient with a tractional retinal detachment.

Dr. Flynn pointed out that imaging with SD-OCT is not hampered by retained silicone oil, and he emphasized this point with more slides from patients. With blood present, Dr. Flynn explained that imaging is possible “around the blood,” and he showed slides of retinal reattachment visible beneath a preretinal hemorrhage.

Discussing a study on vitrectomy presented at AAO Subspecialty Day 2009 by Julia A. Haller, MD, which examined DRCR Network Protocol D, comparing the procedure for DME with vitreomacular traction vs DME alone, Dr. Flynn discussed how SD-OCT tracking allowed for the monitoring of impressive reductions in central retinal thickness (Figure 4), though, once again, visual acuity remained about the same (Figure 5). “All of us have had that frustration,” Dr. Flynn said, “of a beautiful result when it comes to the anatomy, but the function remains poor.” He showed slides from his “best case,” in which edema largely resolved but visual acuity did not return to normal. In DRCR Protocol D, 37% of patients improved by ≥10 ETDRS letters, but 23% declined by ≥10 letters.

Figure 4. Pars plana vitrectomy for diabetic macular edema showed improvement over a six-month period.

Figure 5. Pars plana vitrectomy for diabetic macular edema did not affect visual acuity to any great extent.

Dr. Flynn then discussed PRP and showed how the laser burns caused by PRP can be imaged with SD-OCT (Figure 6). He laid out the treatments given in the DRCR Network's Protocol I, which consisted of combination therapies for clinically significant macular edema (CSME), and Protocol J, which tested combination therapies in PDR with CSME. Dr. Flynn concluded his presentation by reiterating how much SD-OCT aids retinal physicians in day-to-day treatment of patients with diabetic macular edema.

Figure 6. SD-OCT shows laser burns following panretinal photocoagulation.


The next presentation was given by Rajendra S. Apte, MD, assistant professor in the Department of Ophthalmology & Visual Sciences at Washington University.

Dr. Apte's lecture picked up where Dr. Flynn's left off, looking at the development of a new treatment algorithm for the management of patients with DME by stressing an evidence-based approach.

Dr. Apte began by stating that all macular edemas “are not created equal.” He compared focal vs diffuse DME, explaining that focal DME has less foveal thickening, better visual acuity, less severe retinopathy, and better response to treatment, while diffuse DME tends to be associated with fibrosis and atrophy after laser treatment.

Returning to the DRCR Networks comparison protocol of IVTA vs focal/grid laser photocoagulation (ie, Protocol B), Dr. Apte said, “The laser is like a workhorse,” indicating that it tends to have a greater benefit over the long term. Furthermore, he stressed that 1 mg IVTA is less effective than either 4 mg IVTA or laser in both the short and long terms. That being said, Dr. Apte expressed hope that the DRCR Network's Protocol I (a combination therapy protocol) would bear out use of IVTA in the longer term.

Dr. Apte also discussed the problems of rising intraocular pressure and development of cataracts in patients treated with IVTA. In patients receiving 4 mg IVTA, the rate of cataract approaches 100% over several years (Figure 7).

Figure 7. In patients receiving intravitreal triamcinolone, the rate of cataract approaches 100% by 36 months.

He then looked at the DRCR Network's trial of peribulbar triamcinolone with and without focal photocoagulation for mild DME. Going out to 34 weeks, comparing five arms of the protocol, combination therapy of anterior sub-Tenon's injection of triamcinolone plus laser was the most effective in reducing retinal thickness, although the results were not statistically significant.

Dr. Apte then briefly discussed dexamethasone and fluocinolone trials for DME, including the Fluocinolone Acetonide for Macular Edema (FAME), as well as anti-VEGF agents such as pegaptanib sodium (Macugen, Eyetech), ranibizumab and bevacizumab (Avastin, Genentech). First, Dr. Apte reviewed the phase 2 dose-finding study of pegaptanib, which compared three doses of the drug to a sham injection. It was the first trial that showed the possibility that pegaptanib could be effective in treating DME, Dr. Apte said, and he displayed data demonstrating a 4.7-letter increase in visual acuity at 36 weeks and a −68 μm change in central retinal thickness. However, at the 52-week point, improvements had disappeared (Figure 8).

Figure 8. Bevacizumab improves center point thickness initially, but that improvement disappears.

Turning to the DRCR Network's phase 2 trial of bevacizumab for DME, which randomized patients to five arms combining two doses (1.25 mg and 2 mg) of bevacizumab with and without a reinjection at six weeks and with and without laser, Dr. Apte considered the question of whether 2 mg bevacizumab had a positive effect on DME.

First, when laser was compared with bevacizumab 2.5 mg given at baseline and at the six-week mark, there was initial rapid thinning of the fovea peaking at three weeks, but then laser “catches up” by the 12-week time point. Differences in therapies were only statistically significant at week 3. However, differences in study arms were statistically significant out to 12 weeks when comparing visual acuity, and patients receiving the two injections of 2.5 mg bevacizumab had better visual acuity (as high as eight-letter improvement at nine weeks).

The next question considered was whether 1.25 mg bevacizumab combined with laser was a better therapy than bevacizumab alone. In this protocol, the drug-only arm outperformed the combination arm, although, again, the results were not statistically significant. The conclusions from this protocol are that there is apparently some benefit of bevacizumab when combined with focal grid photocoagulation and that 2.5 mg bevacizumab does not have an appreciably greater effect on DME than 1.5 mg bevacizumab.

Dr. Apte then briefly discussed the READ-2 trial, which compared ranibizumab, laser and combination of the two, and which found ranibizumab monotherapy superior in terms of improvement in visual acuity. He also discussed briefly the vitrectomy protocol of the DRCR Network, noting that, although it is true that improvements were limited in this protocol, it is also true that most patients were not treatment-naïve and perhaps if primary vitrectomy were given, they would have improved more.

Noting now that there is a shifting paradigm in treating DME, Dr. Apte pointed out that treatment of DME continues to evolve, that pharmacotherapy is here to stay, and that changing practice patterns are not always based on sound clinical data. “Before Protocol B data were published,” he noted, “I don't think people would have predicted that laser would continue to have progressive effects on OCT out three years — and maybe even longer.”

Here, Dr. Apte turned to the treatment algorithm (Figure 9). First, he stressed the importance of glycosylated hemoglobin control. Then, based on the type of edema (focal, diffuse, and those edemas that coexist with some kind of vitreoretinal surface abnormality), treatment options follow. In the case of focal CSME (Figure 10), laser photocoagulation should be performed. Resolution of the edema should be followed by monitoring. Partial resolution, Dr. Apte suggested, should be followed by retreatment with laser. In cases of no response or progression of retinopathy, Dr. Apte suggested the use of corticosteriods, but he stressed the importance of monitoring side effects, including possible complications from intravitreal injections.

Figure 9. The first step in the treatment algorithm is hemoglobin control.

Figure 10. In focal macular edema, laser photocoagulation is the first line of treatment.

With diffuse DME (Figure 11), two subtypes exist: cystic and noncystic. Though laser is unlikely to help on its own in cases of cystic diffuse DME, the first line of treatment is still laser, as is the case with noncystic diffuse DME. However, for patients with cystic diffuse DME, Dr. Apte recommends laser in combination therapy with either anti-VEGF agents or steroids, again noting the possible side effects of either class of drugs. In the case of partial or no resolution of DME, Dr. Apte then recommends reinjection, injection with other agents, or, in intractable cases, surgical management with vitrectomy.

Figure 11. Diffuse macular edema entails two types — cystic and noncystic — and treatments differ accordingly.

Dr. Apte closed by touching on the role of inflammation in the process of DME: “I think as we go down the road, we will see one or more agents that will address the role of inflammation in treating DME patients.”

(Note: Results of the DRCR Network's Protocol I demonstrate significant efficacy of combining laser with either ranibizumab or steroid, but these results were not public at the time of the Retinal Physician Symposium. As such, the algorithm will be modified to reflect anti-VEGF as a primary agent that would be recommended. This will be addressed in a future issue.)


David Boyer, MD, who is clinical professor of ophthalmology at the Keck School of Medicine at the University of Southern California and practices with Retina Vitreous Associates Medical Group in Beverly Hills, gave two presentations during this session. The first was on Alimera's (Alpharetta, GA) Iluvien fluocinolone acetonide intravitreal insert, with Dr. Boyer presenting the results of the Fluocinolone Acetonide for Diabetic Macula Edema (FAME) trial.

Dr. Boyer began by describing the Iluvien insert, indicating its size at 3.5 mm by 0.37 mm and stating that it is inserted intravitreally in an interoffice setting with a 25-g injector (Figure 12). The two doses at which the insert was tested are 0.23 μg per day (low dose) and 0.45 μg per day (high dose). Unlike the dexamethasone implant, the Iluvien insert is not biodegradable. Both doses disperse fluocinolone at clinically significant levels out to 400 days.

Figure 12. The size of the Iluvien fluocinolone acetonide insert is demonstrated.

Dr. Boyer then laid out the study details of the FAME trial, discussing how laser treatment was available at six weeks at the investigator's discretion. Retreatment with either low dose, high dose or sham occurred at 12 months. Nearly 80% of patients were retained in the trial, which Dr. Boyer indicated was high for a trial of diabetic patients.

Notably, the number of patients who were treated with laser was much higher (58.9%) in the sham group than in either the low-dose (36.7%) or high-dose (35.2%) group. This was also the case for patients receiving off-protocol treatments (eg, bevacizumab or ranibizumab) in cases of poor response, with 28.6% of patients in the sham group receiving off-protocol treatment vs 12.5% and 13.9% in the low- and high-dose groups, respectively.

Discussing the anatomical and functional efficiency data, Dr. Boyer indicated both fluocinolone cohorts showed greater improvements over sham, although patients who received laser treatments improved comparably over the long term. Three-line improvers were nearly twice as high in both fluocinolone groups over the sham group. Even more significantly, improvements of three lines or more were over 40% in both dosage groups and over 16% for the sham group when only patients who had not received laser treatment or off-protocol treatments were considered.

Noting that the tendency for improvements in patients to dissipate over time was linked to the development of cataracts, Dr. Boyer presented data on visual acuity in pseudophakic patients, showing that, once cataracts were removed, visual acuity began to improve again (Figure 13).

Figure 13. Once cataracts are removed, visual acuity rates improve once again.

“At what cost are we getting this improvement?” Dr. Boyer now asked. He started by stating there were no larger occurrences of vitreous hemorrhage, vitrectomy, or retinal detachment in either treatment group. Knowing that cataracts and a rise in IOP are inherent risks with any steroid treatment, Dr. Boyer submitted that cataracts were almost ubiquitous, compared to around one-quarter of patients in the sham group. Rises in IOP were dose-dependent in the drug groups, showing that the low dose was safer.

Dr. Boyer concluded his presentation with a slide showing that not all inserts fall into the vitreous, with an insert visible in a patient's iris (Figure 14). He indicated that the Iluvien insert will go before the FDA for approval in the very near future.

Figure 14. A misplaced Iluvien insert can be visible in the iris.

Dr. Boyer's other presentation was on iCo-007, which is a second-generation antisense drug developed by Isis Pharmaceuticals (Carlsbad, CA). He began by saying, “Most of the time, we're talking about treating outside the cell. Now we're going to talk about something that has to get inside the cell in order to function.” He gave a brief history of antisense drugs, referring to their use during the early years of the HIV/AIDS epidemic. Then he stated that the second generation of these drugs are more stable, more potent and less inflammatory than the first.

The drug has shown efficacy in animal models of retinal vein occlusion and choroidal neovascularization. The half-life of the drug is six to eight weeks in monkeys, indicating a halflife of perhaps six months in humans. It clears quickly from the vitreous while being retained in the retina and choroid.

Dr. Boyer then turned to results of a phase 1 dose-escalating trial of iCo-007 in DME. Patients were given a single intravitreal injection and then followed up for six months. Most of these patients were not treatment-naïve and had severely compromised visual acuity.

Turning to the safety data first, Dr. Boyer explained that there were no drug-related severe adverse events and no IOP issues or cases of ocular inflammation. The drug was below detectable levels (2 ng/mL) in the blood plasma of treated patients.

Dr. Boyer then showed several case studies. In the first two cases, biological effect was delayed for several weeks. At a much higher dose, shown in a third case, there was more of what Dr. Boyer termed a “wowie” effect, with a patient who had a central retinal thickness of 867 μm improving to 124 μm at 24 weeks (Figures 15 and 16).

Figure 15. In this pretreatment image, a patient has a central retinal thickness of 867 μm.

Figure 16. In the post-treatment images, the patient's central retinal thickness has improved to 124 μm.

Dr. Boyer closed by indicating that a phase 2 trial for iCo-007 is in the planning stage and that additional data will be presented at future meetings as they become available.


Keep an eye out in coming issues of Retinal Physician for more recaps of presentations from our annual symposium. Sessions on retinal vein occlusions, combination therapies, and surgical innovations will follow. RP