Use of Corticosteroids in the Treatment of Patients With Diabetic Macular Edema Who Have a Suboptimal Response to Anti-VEGF: Recommendations of an Expert Panel

Introduction

Diabetic macular edema (DME) is characterized by accumulation of fluid and retinal thickening within the macula. The pathogenesis of DME is multifactorial. Vascular endothelial growth factor (VEGF) has a central role in the breakdown of the blood-retinal barrier that leads to DME,¹ but other pro-inflammatory factors and processes also appear to be involved.^2,3 DME is the most common cause of vision loss in patients with diabetes.⁴ It is estimated to affect approximately one in 25 individuals with diabetes older than age 40 in the United States,⁵ and the number of cases of DME can be expected to increase because of the growing prevalence of diabetes and its complications.⁶

The American Academy of Ophthalmology Preferred Practice Patterns-recommended first-line treatment for center-involved DME is intravitreal anti-VEGF therapy, with or without adjunct focal laser treatment.⁷ Ranibizumab 0.3 mg (Lucentis; Genentech, South San Francisco, CA) and aflibercept 2 mg (Eylea; Regeneron, Tarrytown, NY) are U.S. Food and Drug Administration (FDA)-approved anti-VEGF treatments for DME. Bevacizumab 1.25 mg (Avastin; Genentech, South San Francisco, CA) is commonly used via an intravitreal injection in an off-label fashion for treatment of DME; a formulation approved for systemic cancer therapy is repackaged in syringes by a compounding pharmacy for off-label intravitreal ophthalmic use. Many studies have demonstrated that intravitreal anti-VEGF treatment has a favorable tolerability profile and is effective in reducing central retinal thickness and improving best-corrected visual acuity (BCVA) in patients with DME.⁸ Even with monthly administration of anti-VEGF, however, the response to anti-VEGF therapy is suboptimal in a substantial proportion of patients.^9–11 After 2 years of monthly ranibizumab 0.3 mg injections for the treatment of DME in the RISE/RIDE registration trials, central foveal thickness remained greater than 250 μm on time-domain optical coherence tomography (OCT) in 24.8% of patients, and BCVA was worse than 20/40 in 42.8% of patients.⁹

Corticosteroids are also available for the treatment of DME. Although intraocular corticosteroid treatment is commonly associated with adverse effects of cataract and increases in intraocular pressure (IOP), intravitreal corticosteroids have demonstrated efficacy in improving central retinal thickness and BCVA in DME and, therefore, are viable treatment options.^7,12,13 In the United States, dexamethasone intravitreal implant 0.7 mg (DEX implant) (Ozurdex; Allergan plc, Dublin, Ireland) is FDA-approved for treatment of DME, and intravitreal fluocinolone acetonide insert (FAc) (Iluvien; Alimera Sciences, Alpharetta, GA) is approved for treatment of DME in patients who have been previously treated with a course of corticosteroids and did not have a clinically significant rise in IOP. In addition, intravitreal triamcinolone acetonide (TA) has been used off-label for many years and has demonstrated beneficial effects in patients with DME.¹⁴

There is a need for guidelines on how DME should be treated to provide patients with the best possible outcomes when the response to anti-VEGF therapy is inadequate. To address this need, a panel of clinical experts in retinal ophthalmology was brought together to develop consensus recommendations for DME treatment through a modified Delphi process. The objective was to develop a clinically recommended treatment paradigm for DME, with particular focus on how corticosteroids fit within the treatment paradigm.

Patients and Methods

The Delphi technique is a widely used and accepted method for building consensus from experts in a field through iterations of a survey and anonymous feedback; in each round of the survey, the participants review and assess the results and feedback from the previous round, then respond to the survey again, and the process is repeated until consensus is reached.¹⁵ A modified Delphi approach incorporating a face-to-face meeting led by a moderator has been used in previous studies to develop recommendations for the assessment and management of ophthalmic diseases^16–21 and was used in this study to develop a recommended treatment paradigm for DME. A panel of nine retina specialists with expertise in the treatment of diabetic retinopathy and DME were invited and agreed to participate in the development of the treatment guidelines. After a literature review was conducted to identify current treatment patterns in DME, a survey including 22 multipart, multiple-choice, and open-ended questions regarding treatment scenarios and clinical decision-making was developed by Endpoint Outcomes and Allergan (See Appendix A below). The questions were designed to elucidate key elements of the treatment paradigm including assessment of the response to anti-VEGF therapy, the role of corticosteroids in the treatment paradigm, differentiation of available corticosteroid treatment options, and corticosteroid use and side effects.

The panel responded to the survey in two formal rounds of feedback. The first round of the survey was conducted on the SurveyMonkey online survey platform ( www.surveymonkey.com). Nine panelists participated in the survey and provided anonymous feedback. Following the first round, the results were summarized and provided to the panelists. Eight of the panelists then participated in the second round of the survey at a “real-time” Delphi panel meeting on June 4, 2016, in Philadelphia, Pa. At the meeting, the anonymous summary results of the first round were presented to the panelists. Answers to open-ended questions in the first round were used to frame multiple-choice answers to survey questions for the second round (See Appendix B below). For each survey question, if consensus was not reached in the first round, the results were discussed and debated by the panelists. The panelists then voted for their chosen response. An electronic audience response system (ResponseCards; Turning Technologies, Youngstown, OH) was used to capture anonymous votes during the Delphi panel discussion, and all panelists were required to vote before results were displayed. Discussion was continued and voting was repeated until consensus was reached, or it was clear that it was impossible to achieve consensus. During this process, modification of survey questions and possible responses, based on opinions expressed during the discussion, was allowed to facilitate the achievement of consensus. Consensus was defined as agreement by at least six of the nine panelists in the first round (66.7%), or by at least six of eight panelists at the Delphi panel meeting (75%).

Results

The panelists reached consensus on many issues in the treatment of DME (Table 1) and developed a recommended treatment paradigm for center-involved DME in patients who have an incomplete response to anti-VEGF treatment (Figure).

Table 1: Delphi Panel Consensus on Issues in the Treatment of DME

Figure. Recommended treatment paradigm for center-involved DME. BCVA = best-corrected visual acuity; DME = diabetic macular edema; VEGF = vascular endothelial growth factor.

The key aspects of the treatment paradigm are as follows.

Role of Corticosteroids in the Treatment Paradigm

Consensus was reached that intravitreal corticosteroid treatment is appropriate in multiple scenarios of DME (Table 2). Phakic as well as pseudophakic patients who have an inadequate response to anti-VEGF therapy can be candidates for corticosteroid treatment, and corticosteroid treatment can be used in patients with severe edema and those who have undergone vitrectomy. Corticosteroid treatment can also be considered in DME patients with co-existing glaucoma as long as IOP is controlled on one or two medications and the optic nerve is healthy. Several panelists indicated that patients who have had successful filtering surgery can be candidates for corticosteroid therapy as long as the cup-to-disc ratio does not exceed 0.8.

Table 2: Delphi Panel Consensus on Candidates for Corticosteroid Treatment of DME

Discussion

The introduction of anti-VEGF treatment has greatly improved visual outcomes in patients with DME, but a significant proportion of patients have an incomplete response to anti-VEGF treatment.^9–11 Corticosteroids are a rational approach to the treatment of DME because inflammatory mediators and pathways in addition to VEGF appear to be involved in the development of DME. However, there has been limited practical guidance on when and how to use corticosteroids in the treatment regimen. The consensus recommendation of this panel is that corticosteroids be used in patients with an inadequate response to anti-VEGF treatment, and guidance is provided on how to identify patients who are inadequate responders to anti-VEGF, and when to progress to corticosteroid therapy.

The recommendation of the panel is that center-involved DME with some degree of decreased visual acuity should be treated promptly with anti-VEGF therapy. Most patients with DME respond well to anti-VEGF treatment. Currently bevacizumab is used more frequently than ranibizumab or aflibercept for treatment of ophthalmic disease in clinical practice²² because of its lower cost. In the DRCR.net Protocol T study comparing these anti-VEGF agents in patients with DME, after 1 year of treatment, aflibercept was more effective than ranibizumab or bevacizumab in improving BCVA in patients with baseline BCVA worse than 20/40, whereas the anti-VEGF agents showed comparable efficacy in patients with baseline BCVA of 20/32 to 20/40.²³ After 2 years of treatment, aflibercept and ranibizumab provided similar improvement in BCVA, and aflibercept continued to be more effective than bevacizumab in improving BCVA in patients with baseline BCVA worse than 20/40.²⁴ Although the efficacy of switching among anti-VEGF therapies has not been well studied,²⁵ favorable outcomes have been reported after patients with suboptimal response switched from bevacizumab to ranibizumab,²⁶ or from bevacizumab or ranibizumab to aflibercept.²⁷ Therefore, if a patient has a poor early response to anti-VEGF, and the initial anti-VEGF agent used was not aflibercept because of considerations such as cost and reimbursement, the clinician may consider switching the patient to aflibercept. The consensus of the panel, however, is that whether or not there is a switch among anti-VEGF therapies, continued edema should be tolerated for no longer than 6 months.

There are three fundamental reasons why a patient with an inadequate response to anti-VEGF therapy typically should remain on anti-VEGF treatment for no longer than 6 months before proceeding to a corticosteroid. First, there is evidence that anti-VEGF treatment is most effective in early disease.^28–30 In the RISE/RIDE registration studies of ranibizumab, patients in the initial sham group who crossed over to ranibizumab treatment at year 2 had less favorable outcomes after 12 months of ranibizumab treatment than patients in the initial ranibizumab groups,²⁸ and patients with a longer duration of DME at baseline required more as-needed ranibizumab injections during the open-label study extension after year 3,²⁹ suggesting that prompt treatment of shorter-term DME with ranibizumab results in better outcomes. Similarly, in the VIVID/VISTA registration studies of aflibercept, patients in the initial laser treatment group achieved a smaller gain in BCVA after receiving rescue aflibercept treatment compared with patients in the initial aflibercept treatment groups, suggesting that delaying aflibercept treatment results in poorer outcomes and that aflibercept treatment is more effective in shorter-term DME.³⁰ Second, continuing anti-VEGF treatment when the early response to treatment is poor usually does not result in favorable outcomes. The EARLY study analysis of results from the DRCR.net Protocol I study showed that the majority of patients who had a less-than-five-letter gain in BCVA from baseline at week 12 (after three monthly ranibizumab injections) continued to have less-than-five-letter gains in BCVA from baseline at week 156 (after 3 years of monthly ranibizumab injections).¹¹ Third, delaying effective treatment increases the risk of permanent vision loss, because chronicity of DME is associated with foveal atrophy and photoreceptor damage.³¹

Members of the panel agreed that the balance of VEGF versus other inflammatory cytokines appears to be an important factor contributing to the different levels of anti-VEGF response observed in patients with DME. The importance of this balance goes hand-in-hand with the importance of the chronicity of DME in the variability observed in patient response to anti-VEGF treatment, because it is generally believed that DME is primarily mediated by VEGF in its early phase, whereas the involvement of other cytokines and inflammatory mechanisms becomes more important in chronic disease.² Low-grade inflammation has been postulated to be responsible for the retinal damage that occurs in chronic DME.²

It may be important to treat the inflammatory component of DME in patients who do not respond adequately to anti-VEGF treatment. Studies have demonstrated the effectiveness of corticosteroid treatment in long-standing and refractory DME. In the MEAD registration studies for DEX implant, the treatment effects of DEX implant relative to sham were similar in subgroups defined by the duration of DME at baseline: DEX implant had beneficial effects in longer-term (duration of more than 3 years) as well as short-term (less than 1 year in duration) DME.^12,32 Furthermore, in the FAME registration studies for FAc, the percentage of patients with 15-letter or greater gains in BCVA from baseline was superior with FAc relative to sham only in the subgroup of patients with long-standing DME (duration of 3 years or longer),³³ consistent with an important role of inflammatory pathways amenable to corticosteroid treatment in the pathophysiology of persistent DME.

The panel recommends that a patient who has a less than 50% reduction from baseline in the excess macular thickness after three to four monthly anti-VEGF injections, or whose BCVA has not improved to 20/40 after three to six monthly injections because of edema, be switched to corticosteroid treatment. The definitions of inadequate response include a range in the number of anti-VEGF injections, because the decision of whether to change treatment should take into account the trajectory of the response. A patient with continual improvement in edema after three or four sequential injections typically would be re-treated with anti-VEGF, whereas a patient with no improvement in edema after three or four anti-VEGF injections typically would be switched to a corticosteroid.

The panelists recommend that patients on anti-VEGF therapy receive frequent (monthly) injections and have monthly assessments to evaluate the trajectory of the response to treatment and identify inadequate responders. However, they recognize that some practitioners may see patients monthly for anti-VEGF injections but wait until after three to four injections before doing a full examination and OCT, rather than doing full exams at every visit. With this approach, the decision of whether to continue anti-VEGF treatment or switch to a corticosteroid is still made 3 to 4 months after the initial anti-VEGF injection, consistent with the suggestion from the EARLY study that this is an appropriate timeframe for deciding whether to switch therapy, since BCVA outcomes at week 12 in the Protocol I study were predictive of long-term outcomes.¹¹ Although this approach does not consider the trajectory of the treatment response, there are cost savings associated with fewer examinations, and there is little risk of overtreatment, since edema is unlikely to resolve after one to three anti-VEGF injections. The panelists also recognize that the frequency of anti-VEGF injections required can be burdensome for patients,^34,35 but in the experience of the panelists, patients are likely to adhere to frequent injections if they are educated on the alternative of risk of permanent loss of vision. The consensus opinion of the panel is that frequent monitoring is also needed for patients who have switched to corticosteroid implant treatment, but the recommended interval between visits for these patients is longer (1 to 2 months) than that recommended for patients on anti-VEGF therapy.

The panel's consensus choice of corticosteroid for treatment of DME in inadequate responders to anti-VEGF is DEX implant, primarily because DEX implant has a more favorable safety profile compared with FAc and intravitreal TA. The increases in IOP associated with intraocular corticosteroid use are less frequent, less severe, and more easily managed with DEX implant than with FAc or intravitreal TA.^{12,13,36–38} Accordingly, the FDA-approved indication for FAc in DME is limited to patients who have been demonstrated to have no clinically significant IOP response to a corticosteroid. Disadvantages of use of intravitreal TA include its lack of FDA approval for treatment of DME and the possible need for more frequent injections. DEX implant has been demonstrated to be effective in patients with previous vitrectomy,³⁹ patients with previously treated DME,⁴⁰ and patients with DME that has been demonstrated to be resistant to anti-VEGF treatment.^41–45

The primary role of corticosteroids in the treatment paradigm for DME is as alternative treatment after a suboptimal anti-VEGF response, but corticosteroids may also be considered for treatment in patients who discontinue anti-VEGF therapy because of systemic safety concerns. Although intravitreal anti-VEGF treatment is generally believed to be systemically safe, it has been difficult to determine whether anti-VEGF might increase the risk of stroke, arteriothrombotic events, or cardiovascular events because studies have been underpowered and patients with DME may be at increased risk of these events regardless of treatment.⁴⁶ However, the majority of panelists agreed that they typically would likely consider a change in therapy if a patient being treated with anti-VEGF had a recent stroke or cardiovascular event. In a recent meta-analysis that pooled systemic safety data from patients treated with ranibizumab 0.5 mg or aflibercept in clinical studies, the risk of cerebrovascular accidents and vascular deaths was significantly higher in patients treated monthly for 2 years with anti-VEGF injections than in patients treated with sham,⁴⁷ suggesting that frequent anti-VEGF injections over the long term may increase the risk of serious systemic adverse events.

There is a lack of evidence from clinical studies concerning some aspects of care in DME, such as the benefits of using combination therapy with an anti-VEGF agent and a corticosteroid. In their discussion at the Delphi meeting, panelists indicated that they generally use combination anti-VEGF and corticosteroid therapy only for patients with severe and persistent DME. Furthermore, based on their clinical experience, the panelists believe that a patient with a suboptimal response to anti-VEGF has better improvement in retinal drying if the patient is switched to a corticosteroid than if the patient is switched to a different anti-VEGF agent, but supporting data from clinical trials are lacking.

The clinically recommended treatment paradigm for DME presented here includes recommendations for the medical treatment of DME, and it was developed with the understanding that the recommendations should be general and applicable in 80% of cases. The authors recognize that there are exceptions and that there is a need for individualized patient care, as well as a need for surgery in cases of epiretinal membrane or vitreomacular traction. Some patients will lose vision irretrievably even if the recommendations are followed. The recommendations are based on evidence from reported studies and the clinical experience of the panelists, and they may be modified when more data become available. It is anticipated that results of the ongoing DRCR.net Protocol U study ( clinicaltrials.gov identifier: NCT01945866), comparing ranibizumab monotherapy to combination therapy with ranibizumab and DEX implant in patients with an incomplete response to three monthly injections of ranibizumab, may further help guide corticosteroid use in the treatment of DME.