The Looming Unmet Need in Diabetic Eye Care

Part 1: A growing epidemic necessitates new approaches in therapy.

PRAVIN U. DUGEL, MD

Diabetes is a growing global health burden. Its prevalence is expected to rise from 382 million cases in 2013 to 592 million cases in 2035, a 55% increase.¹ In North America and the Caribbean, the growth in prevalence is forecast to be 37%. In other areas of the world, such as the Middle East/North Africa and Africa, the growth rate approaches or exceeds 100%.

It may be hard to imagine such immense growth in the prevalence of this disease, but it is clear that physicians will need the treatments necessary to care for patients effectively. Diabetes is a leading cause of kidney failure and a major risk factor for heart disease, stroke, and neuropathy.

As the diabetes epidemic progresses, retina specialists will see the accompanying surge in diabetic retinopathy and diabetic macular edema. DR affects approximately 28.5% of diabetic patients who are 40 years old or older,² and DME affects approximately 10% of all diabetic patients.³

IDENTIFYING GAPS IN CURRENT TREATMENT STRATEGIES

In the search for new therapies to help us safeguard our patients’ sight as the diabetes epidemic grows, it is instructive to consider what we have learned about our current treatment options, as well as diabetic eye disease in general.

Based on multiple randomized clinical trials, anti-VEGF-A monotherapy has supplanted laser photocoagulation as standard of care for patients with DME. For example, in the RISE/RIDE, VIVID/VISTA, and DRCRnet Protocol I and T trials, patients did very well. Favorable anatomic and visual acuity outcomes were achieved and maintained over the course of the studies.^4-7 The best results were obtained with frequent injections, i.e., ranibizumab (Lucentis, Genentech, South San Francisco, CA) monthly and aflibercept (Eylea, Regeneron, Tarrytown, NY) every eight weeks after five monthly injections.

However, it has been shown that, outside of clinical trials, in the “real world,” patients are receiving far fewer injections. Utilization of anti-VEGF-A injections in clinical practice is lower than in randomized, clinical trials.

Four recent analyses of insurer databases reported mean number of injections over 12 months of 3.6, 2.4, 3.1, and 3.9.^8-11 These data compare with, for example, in year 1, 11 in RISE/RIDE,⁴ nine in Protocol I,⁶ and seven in RESTORE.¹² These differences are likely due to the heavy burden of frequent office visits for a population of patients that spends an inordinate amount of time in doctors’ offices, but it means it is doubtful whether we are achieving outcomes as good as in the clinical trials.

It is also important to note that number of injections is not a proxy for treatment burden. While the need for injections tends to decrease over the course of a clinical trial, there are no data to suggest that the favorable results obtained would be possible if the patients were not seen on a very regular basis.

Furthermore, patients’ response to anti-VEGF-A monotherapy for DME varies, independent of the number of injections they receive. Evidence of this is seen in randomized clinical trials, as well as in a recent analysis of Medicare data that I conducted.¹³

Not surprisingly, the analysis showed that the number of injections given was significantly less than one would expect in a clinical trials. In addition, two distinct groups emerged. For one group, the diagnosis of DME dropped off, indicating that their disease was modified. For the other group, the disease persisted regardless of how many injections they received. Approximately 50% of patients had a great result, and approximately 50% did not.

The DRCRnet Protocol I trial⁶ produced a similar finding. Approximately 50% of patients responded well to anti-VEGF-A treatment and approximately 50% of patients did not. Again, number of injections was not responsible for the differences in response. In some cases, the patients who received the highest number of injections also had the poorest result.

OTHER HELPFUL CONCEPTS DRAWN FROM CLINICAL TRIALS

The wealth of clinical data we now have from studies of anti-VEGF-A treatment for DME has led to other insights we can use to help guide the development of future therapies. For example, it has confirmed the theory that diabetic eye disease is mediated by factors in addition to VEGF.

It appears that the vascular permeability that is present throughout the natural course of DR/DME — as severity progresses from early focal leakage to diffuse leakage, fibrosis, pigmentary changes and loss of photoreceptor cells — is mediated initially by VEGF and later by inflammation.

There seems to be a point — we don’t know exactly when — at which a multifactorial “switch is thrown,” changing the disease from primarily VEGF-mediated to primarily inflammation-mediated. We see evidence of this in the RISE and RIDE extension studies¹⁴ in which patients who were switched to ranibizumab treatment after having received sham for two years never caught up with the vision gains of patients who were initially treated with ranibizumab.

In contrast, in the RESTORE trial,¹² patients were switched from sham to ranibizumab after one year and they did catch up. The differing result in this aspect of the two trials suggests that the disease itself changes at some point.

Also supporting the multifactorial switch theory is the body of evidence showing that inflammatory cytokines are elevated in the eyes of patients with diabetes¹⁵ and they are expressed at a higher level in eyes with DME.¹⁶ And, in a subgroup analysis from the FAME trial of the fluocinolone acetonide implant (Iluvien, Alimera, Alpharetta, GA), vision outcomes were better in patients with DME duration of three or more years at baseline (34% gained 15 or more letters of vision).¹⁷

The difference in efficacy between chronic and nonchronic DME patients wasn’t based on baseline vision, OCT findings or treatments received during the trial. This, too, suggests a transition in the pathophysiology of the disease.

If the theory that diabetic eye disease changes at some point from primarily anti-VEGF-driven to primarily inflammation-driven is correct, it would obviously be helpful to know when this occurs. Currently, we can assess disease status only phenotypically, but it has recently been determined that the response to three monthly injections (in 12 weeks) of an anti-VEGF-A agent can predict long-term treatment outcomes in patients with DME.

In what’s referred to as the EARLY Analysis,¹⁸ raw data from the anti-VEGF plus deferred laser arm and the anti-VEGF plus prompt laser arm of the DRCR.net Protocol I trial were pooled, and patients were divided into three groups: those whose improvement from baseline was fewer than five letters of vision at 12 weeks, those who improved by five to nine letters at 12 weeks, and those who improved 10 or more letters at 12 weeks. At 12 weeks, one year, and the end of the study (three years), the vision changes in the three groups were as follows:

After controlling for potential confounders, the strong association remained between vision gains from baseline to 12 and 52 weeks, making the EARLY Analysis the first study to demonstrate that long-term response to anti-VEGF-A monotherapy in DME can be predicted after three injections. Knowing this, we can be ready to initiate additional therapies with alternate modes of action, including steroids or future treatments, when necessary.

Another useful finding from recent randomized clinical trials is that some patients not only do well with anti-VEGF-A monotherapy, but the treatment can actually modify their disease, making additional injections unnecessary. (Figures 1 and 2).

Figures 1 and 2. It has been shown that anti-VEGF-A therapy can modify diabetic eye disease, ie, improving not only DME but also DR, making further treatment unnecessary.

In RISE/RIDE, for instance, approximately one-third of patients, with monthly injections, experienced a two-step or greater regression of DR severity, and approximately 5% experienced a three-step regression. This knowledge led to the FDA approving ranibizumab and aflibercept for the treatment of DR as well as DME, and future therapies will likely be designed to work earlier in the disease process, prior to the development of DME.

WORKING TOWARD THE IDEAL TREATMENT

Given what we know about DR, DME, and our current treatment options, it follows that newly developed treatments ideally would have:

• an excellent safety profile

• the effect of vascular stabilization

• a bilateral effect

• a drug delivery device/system for a sustainable treatment strategy.

In an effort to develop tomorrow’s treatment strategies, several avenues are being explored, including a refillable, implantable ocular drug pump (Replenish, Inc., Pasadena, CA) and a refillable drug port delivery system (Genentech and ForSight VISION4).

Based on the idea that an impactful strategy for combatting diabetic eye disease is to target DR rather than DME, which occurs later in the disease process, Aerpio Therapeutics (Blue Ash, OH) is evaluating a drug candidate, known as AKB-9778, that has a novel mechanism of action.

By inhibiting the essential protein vascular endothelial protein tyrosine phosphatase (VE-PTP), AKB-9778 activates the TIE-2 transmembrane tyrosine kinase receptor, which is located almost exclusively on endothelial cells and essential for vascular stability.

TIE-2 inhibits permeability, blood retinal barrier breakdown and inflammation. The drug activates TIE-2 regardless of the levels of TIE-2 agonists Ang-1 and Ang-2 present. AKB-9778 is provided in a pre-filled syringe and meant for self-administered subcutaneous injection. It is stable at room temperature and has high ocular availability but is rapidly absorbed into and eliminated from the systemic circulation.

The Phase II TIME-2 study evaluated the safety and efficacy of daily subcutaneous injections of AKB-9778 administered with and without monthly intravitreal injections of ranibizumab for three months in patients with DR, with or without DME.

The percentages of patients who achieved an improvement of two or more steps on the diabetic retinopathy severity scale from baseline were 8.8% in the ranibizumab alone group, 10% in the AKB-9778 alone group, and 11.4% in the ranibizumab plus AKB-9778 group. In addition, AKB-9778 was shown to have an effect on the fellow eye. Evaluation of AKB-9778 is ongoing.

CONCLUSION

In the next part of this series of articles, I will review the additional products currently in the pipeline for the treatment of DME. RP

REFERENCES