Evidence-based Treatment of RVO With Anti-VEGF Drugs

In the second of two parts, we review the evidence for treating central RVO.

DILRAJ S. GREWAL, MD • SHARON FEKRAT, MD, FACS

In the last issue of Retinal Physician, we discussed the evidence for treatment of branch retinal vein occlusion with anti-VEGF drugs. In the second part of our review, we consider the evidence in treating central RVO. We also offer some general considerations when treating RVO with drugs that block VEGF.

BEVACIZUMAB

In a rat model of CRVO, bevacizumab (Avastin, Genentech, South San Francisco, CA) prevented the upregulation of VEGF-A and pigment epithelium-derived factor, which is known to influence the development of vascular edema,¹ while decreasing upregulation of the proinflammatory cytokine interleukin 1B.

Epstein and coworkers performed a six-month, randomized, sham-controlled, double-masked trial, in which 60 eyes with macular edema secondary to CRVO were randomized to receive either 1.25 mg of bevacizumab or sham injection every six weeks for six months, after which all of the eyes received bevacizumab every six weeks until month 12.

The mean VA gain at 12 months was 16 letters in the bevacizumab/bevacizumab group, which was significantly higher than the 4.6 letters in the sham/bevacizumab group.²

PEGAPTANIB

The Macugen in CRVO Study was a randomized, dose-ranging, double-masked, sham-controlled, multicenter, phase 2 trial to evaluate the off-label use of pegaptanib sodium (Macugen, Eyetech, New York, NY) in eyes with macular edema due to CRVO in a one-year study.³

Patients who developed a nonischemic CRVO in the previous six months and had macular edema at baseline, central retinal thickness ≥250 µm and visual acuity of 20/50 to 20/400 were randomized to either 0.3 mg of pegaptanib (n=33), 1 mg (n=33), or sham injection (n=32) every six weeks for 24 weeks.³

The patients treated with 0.3 mg and 1 mg of pegaptanib experienced ≤15 letter losses of 9% and 6%, respectively, compared to 31% in sham-injected eyes. The groups receiving 0.3 mg and 1 mg of pegaptanib showed 7.1 and 9.9 mean letter improvement, respectively, compared to the 3.0-letter loss in sham-injected eyes, and they also exhibited a significantly greater decrease in CRT.

Despite these initial results, with the advent of bevacizumab, ranibizumab (Lucentis, Genentech) and aflibercept (Eylea, Regeneron, Tarrytown, NY), pegaptanib is no longer commonly used and remains off-label for RVO.

RANIBIZUMAB

The efficacy of ranibizumab in eyes with CRVO was confirmed by the double-masked, multicenter, randomized phase 3 CRUISE study,⁴ which demonstrated the efficacy of monthly 0.3- or 0.5-mg ranibizumab injection compared to sham for the treatment of macular edema secondary to CRVO in 392 eyes.

Following the six-month endpoint, the control group was treated with ranibizumab on a PRN basis for macular edema, limiting further information on the natural history of CRVO and any long-term comparisons in this study between a true control group and the treatment group. The mean VA improvements in CRUISE were 12.7 letters (0.3 mg) and 14.9 letters (0.5 mg), compared with 0.8 letters in the sham group at six months.⁴

At 12 months, the mean improvement was 13.9 letters in both the 0.3-mg and 0.5-mg ranibizumab groups, compared to 7.3 letters in the sham/0.5 mg group. Proportions of 46.2% (0.3 mg) and 47.7% (0.5 mg) of ranibizumab-treated eyes gained ≥15 letters compared to only 16.9% in the sham group, with mean reductions in CRT of 434 µm (0.3 mg) and 452 µm (0.5 mg), compared to 167.7 µm in the sham group at six months.⁴

In CRUISE, the 0.3-mg and 0.5-mg treatment groups each showed a substantial decrease in mean BCVA between months 6 and 7. This decrease occurred to a greater extent in CRVO patients, and there was gradual improvement thereafter using the PRN regimen.^5,6 The benefits of ranibizumab for CRVO observed in the first six months were generally maintained at one year.

Questions About Dosing Regimens

However, the initial decline in BCVA after instituting PRN therapy raised questions as to whether visual outcomes at one year could have been even better if the patients had continued to receive monthly injections during the second six months and whether providing additional monthly injections before a PRN regimen could eliminate this initial loss in VA.

Patients who completed the CRUISE trial were eligible for the open-label, multicenter extension study — the HORIZON trial⁷ — to determine the long-term effects of ranibizumab in CRVO. In the HORIZON-CRVO study, with a mean follow-up of 51.4 months, 44% of CRVO eyes had resolution of edema, 78% gained ≥15 letters, and 64% achieved ≥20/40 VA. However, the other 56% still required six injections during the last year of follow-up, only 33% gained ≥15 letters, and 28% had VA ≥20/40.

The RETAIN study⁸ followed 32 CRVO patients from the HORIZON study for another two years. With a mean follow-up of 49.7 months, 14 of 32 CRVO patients (44%) had edema resolution, with 71% receiving their last injection within two years of treatment initiation.

Compared with eyes with persistent CRVO, eyes with resolved disease had greater improvement in BCVA (25.2 vs 4.3 letters), and a greater percentage had a final BCVA of ≥20/40 (64.3% vs 27.8%).

This analysis showed that, despite a good visual response, approximately 50% of patients continued to need anti-VEGF injections for edema as long as four years later.

There were, however, some patients who lost vision. Some experienced ischemic damage to the fovea at the time of or soon after the occlusion, and therefore, visual potential was limited from the onset.

Other patients who had good visual potential initially lost that over time from progressive ischemic damage to the fovea, photoreceptor damage from edema, or both. Photoreceptor damage is more likely when edema is severe and prolonged, repeated, or both, but it also seems that some individuals may be more susceptible than others who maintain good vision despite repeated bouts of edema.

Beyond Resolving Edema

Resolution of macular edema should not be the sole treatment objective. The prevention of worsening retinal nonperfusion is an important treatment objective as well. Periodic fluorescein angiograms, preferably wide-angle, should be performed to monitor perfusion status.

In the SHORE study, for CRVO patients at month 12, the mean BCVA improvement was 18.9 letters, compared to 14.7 letters at month 12 for the 0.5 mg ranibizumab group in CRUISE and similar to aflibercept in the COPERNICUS, and GALILEO studies.^9,10 The SHORE trial demonstrated that a PRN regimen with monthly assessment in RVO patients who achieved disease stability after ≥7 monthly ranibizumab injections could achieve similar VA outcomes with dosing every other month during the PRN phase, compared to continued monthly treatment over 15 months.

A post-hoc analysis of CRUISE patients¹¹ demonstrated that, at month 3 of ranibizumab treatment, OCT images were predictive for CRVO patients. VA at months 6 and 12 was reduced in the 0.5 mg ranibizumab-treated CRVO eyes with persistent cystoid macular edema at month 3. The findings suggest that late or incomplete responders may need careful follow-up to recognize and treat chronic or recurrent edema.

AFLIBERCEPT

Two parallel, randomized, multicenter, double-masked, phase 3 trials, the COPERNICUS and GALILEO studies, evaluated the efficacy and safety of 2 mg of aflibercept for the treatment of macular edema secondary to CRVO,^9,10,12,13 and the results prompted its approval by the FDA for treatment of CRVO in 2012.

All eyes received monthly aflibercept during the first six months. From week 24 to week 52, the treatment protocol differed. Patients in COPERNICUS received aflibercept PRN according to prespecified retreatment criteria, and sham-treated eyes were allowed to receive aflibercept after week 24. In contrast, patients in GALILEO maintained their original randomization, receiving aflibercept or sham through week 52, with panretinal laser photocoagulation allowed at any time during the study to treat neovascularization.

During the PRN phase in COPERNICUS, an average of 2.5 injections were given in the aflibercept monthly/PRN group. Patients in the treatment arms of both studies gained 17.3 to 18 letters at week 24. At week 52, those gains were sustained.

However, the two-year COPERNICUS results¹³ demonstrated that the visual improvements after fixed-dosing through week 24, followed by PRN dosing with monthly monitoring from weeks 24 to 52, were diminished after continued PRN dosing, when monitoring frequency was reduced during weeks 52 to 100.¹³ Eyes receiving sham/aflibercept PRN did not achieve visual and anatomic improvement as robustly as those receiving aflibercept monthly from initiation, suggesting a benefit from early treatment initiation.⁹

In the COPERNICUS and GALILEO trials, there was a significantly greater reduction in CRT at six months in the aflibercept group than in the sham group. However, this difference was maintained over time only in the GALILEO trial, in which patients continued their assigned treatment up to 12 months.

In the COPERNICUS trial, patients in the sham group also received aflibercept during the extension period, which caused a similar decrease in CRT as in the original intervention group. Aflibercept did not appear to increase the incidence of ocular or nonocular adverse events compared with sham in both trials.

The CRUISE, HORIZON, COPERNICUS, and GALILEO studies demonstrated that long-term anti-VEGF treatment is necessary to control macular edema in many eyes with CRVO, likely because the continued ischemia leads to persistently increased VEGF production.

The COPERNICUS year 2 results demonstrated that the visual and anatomic improvements decreased after changing from fixed to PRN dosing, and they declined further when the monthly monitoring frequency was decreased to quarterly.

These outcomes suggest that a PRN dosing regimen with at least quarterly evaluations and treating edema only after it has recurred may not be sufficient, at least for some patients, to maintain the improvements gained after a fixed monthly dosing treatment regimen. PRN dosing may lead to fewer injections than a fixed monthly regimen, but it should come with the requirement of monthly visits.

Given that, outside of a clinical study setting, a close monitoring schedule might not always be practical, a treat-and-extend regimen or a fixed-dosing regimen of aflibercept every two months, after an initial period of monthly doses, could be a viable treatment option to reduce the monitoring burden.¹²

CORTICOSTEROIDS

Intravitreal Triamcinolone

The SCORE-CRVO trial compared the efficacy and safety of 1 mg and 4 mg of preservative-free intravitreal triamcinolone acetonide (IVTA; Kenalog, Bristol-Myers Squibb, New York, NY) to observation in eyes with vision loss associated with macular edema secondary to perfused CRVO.¹⁴ Patients in the corticosteroid medication groups received an average of two injections during the first 12 months of the study.

The trial provided the first level I evidence to support intervention for VA gain in macular edema due to CRVO, albeit at the expense of increased cataract progression and elevated intraocular pressure. These effects did necessitate further treatment in the form of increased rates of cataract surgery at 12-24 months in the triamcinolone groups and the need for IOP-lowering drops at 12 months (35% in the 4-mg IVTA group and 20% in the 1-mg IVTA group, compared with 8% in the observation group).

After one year, 27% of patients in the 1-mg group and 26% of patients in the 4-mg group experienced a visual gain of ≥15 letters, compared to 7% in the observation group. These results appeared to last up to two years, although the two-year results included a smaller number of patients.

The trial data supported the 1-mg IVTA dose, given its superior safety profile over the 4-mg dose, and it concluded that the 1-mg dose, following the retreatment criteria applied in the SCORE study, should be considered for up to one, and possibly two, years.

The proportion of patients gaining ≥15 letters was also significantly larger in the intervention groups at 12 and 24 months (25.6% compared with 6.8% and 31% compared with 9%, respectively).

The proportion of patients receiving IVTA and losing ≥15 letters was smaller (25.6%) than in the observation group (43.8%), but this difference was not statistically significant (P=.06). CRT decreased in all of the study groups, but there was no significant difference between the groups at either 12 or 24 months.

Dexamethasone Implant

In the CRVO subgroup of the GENEVA trial, which tested a dexamethasone intravitreal implant (Ozurdex, Allergan, Irvine, CA) the peak effects were seen at two months, similar to the BRVO results.

There was a substantial drop-off between two and three months. At two months, the change in BCVA was a gain of 10 letters in the 0.35-mg group, 9 letters in the 0.7-mg group, and 0 letters in the sham group.

These gains decreased to 6 letters in the 0.35-mg group, 4 letters in the 0.7-mg group, and 0 letters in the sham group at 3 months. The percentage of eyes gaining ≥15 letters at two months was 33% in the 0.35-mg group, 29% in the 0.7-mg group, and 9% in the sham group, which decreased to 24%, 18%, and 10%, respectively, at three months.^15,16

At six months, the mean change in BCVA was 2 letters in the 0.35-mg group, 0 letters in the 0.7-mg group, and a decline of 2 letters in the sham group. In the 0.35-mg group, 17% of eyes gained ≥15 letters, compared to 18% of eyes in the 0.7-mg group and 12% of sham eyes.

However, there was no significant difference in the proportion of patients gaining or losing 15 letters at either 6 or 12 months (0.35 or 0.7 mg dexamethasone). This finding may reflect the timing of the peak effect at 90 days with dexamethasone.

There was no significant difference in the reduction of CRT after six months of treatment with the 0.7-mg implant (no data were given for the 0.35-mg implant), compared with sham. In the GENEVA 12-month extension study, the 0.7 mg/0.7 mg group had a mean gain of approximately 2 letters, while the sham/0.7 mg group lost approximately 1 letter at 12 months.

The GENEVA trial did not directly address the question of the optimal retreatment interval for the 0.7-mg implant. It also provided only limited information regarding the 12-month safety and efficacy of a single implant injection because most of the patients in the study received open-label 0.7-mg implant treatment at day 180. Overall, in the subgroup of study eyes diagnosed with BRVO and CRVO, the peak improvement in mean BCVA was approximately 10 letters, and it occurred 60 days after each injection.

One of the differences of the GENEVA trial was that, while the BRAVO and CRUISE studies included patients who had a more recent onset of macular edema (more than half had a short duration), >80% of patients in the GENEVA study had macular edema for >90 days.

Pooled BRVO and CRVO analysis of patients with a short duration (<90 days) of macular edema in GENEVA showed 48% gaining >15 letters at 270 days, which was relatively similar to the improvement seen in the BRAVO (61% at six months) and CRUISE (50% at 12 months) studies.

The Table (pages 25 and 26) summarizes the study characteristics and visual outcomes of the CRVO trials.

SYSTEMIC RISK FACTORS

Anti-VEGF treatments, while effective, are temporizing measures that target downstream mediators of CRVO sequelae and not the underlying etiology. Besides age, risk factors for CRVO include a hypercoagulable state, stroke, hypertension, and diabetes mellitus.¹⁷

Younger individuals who present with a clinical picture of CRVO may have an underlying hypercoagulable or inflammatory etiology.¹⁸ This makes questions about cardiovascular disease, stroke, high cholesterol, and diabetes particularly relevant when eliciting the medical histories of RVO patients.

Identification and treatment of systemic vascular risk factors are of paramount importance.^19-22 It is unclear whether systemic anticoagulation can prevent or alter the natural history of RVO,²³ but prophylactic use may help to prevent nonocular thrombotic events.

There has been discussion that anti-VEGF agents may potentially induce systemic thromboembolic events, but none of the trials with ranibizumab or aflibercept has raised such concerns.

WHICH DRUG TO CHOOSE?

The availability of several therapeutic options for RVO-associated macular edema makes it important to understand how to choose an appropriate treatment. It poses a dilemma to rely on the trial data alone, because these studies compared treatment only to sham or observation.

No direct comparison of aflibercept, bevacizumab, ranibizumab, triamcinolone, and dexamethasone in RVO has been published. Part of the reason for this is that the FDA requires proof of the safety and effectiveness of a drug; thus, for pharmaceutical companies, the most efficient way to establish such safety is through placebo-controlled trials. Comparison of these trials is further limited by the lack of uniformity among the primary endpoints in the various RVO trials.²⁴

Bressler and Schachat proposed that comparing common clinical endpoints, such as percentages of eyes losing or gaining three lines of vision at the clinically relevant endpoint, would be useful for comparison among trials.²⁵ For example, while 56% of aflibercept-treated CRVO eyes gained ≥15 letters at six months in the COPERNICUS trial (vs 12% sham), 48% of ranibizumab eyes at six months in the CRUISE trial (vs 17% sham), 29% following dexamethasone implantation at 30 days in the GENEVA trial (vs 11% sham), and 27% of IVTA-treated patients at one year in the SCORE trial (vs 7% sham) gained ≥15 letters.

It is critical to recognize though that looking at the data in this manner does not allow for true intertrial comparisons because these trials lacked uniformity of enrollment criteria, resulting in distinct trial populations with respect to baseline VA, chronicity of disease, and proportion of ischemic eyes, among other factors. These differences are reflected by the variable responses among sham-treated eyes, preventing direct comparison of even these identical endpoints.

Despite the limitations of cross-trial comparisons, visual and anatomic outcomes from the CRUISE, COPERNICUS, and GALILEO studies were generally comparable, suggesting that similar results could be achieved with ranibizumab or aflibercept in CRVO patients.^7,9

However, it is possible that the steeper decline in VA between months 6 and 7 with 0.5-mg ranibizumab in the CRUISE study, compared with the smaller decline seen with aflibercept during the same time period in the GALILEO study, is reflective of the longer duration of aflibercept.

A meta-analysis comparing aflibercept, bevacizumab, dexamethasone, ranibizumab, and triamcinolone for the treatment of macular edema secondary to CRVO found no differences among ranibizumab, aflibercept, bevacizumab, and triamcinolone for improving vision, but it did find that dexamethasone was not as effective as ranibizumab or aflibercept.²⁶

QUESTIONS REMAIN

Despite these large trial datasets, several clinical questions remain unanswered: Could more aggressive treatment at the onset of RVO or sustained delivery approaches aimed at preventing progression of retinal nonperfusion and repeated, recurrent episodes of edema prevent visual loss in susceptible eyes?

An unanswered question from the HORIZON study regarded why edema resolution occurred in approximately half of the patients with continued treatment, whereas the other half continued to experience bouts of recurrent edema even 4 years after initiation of treatment? Older age at onset of occlusion is one factor. Another possible factor is hypertension: a significantly higher percentage of nonresolvers were hypertensive.

Perhaps the unresolved group had more retinal ischemia, resulting in higher VEGF levels that were difficult to neutralize by injections, particularly during periods of less frequent follow-up. The worsening of retinal nonperfusion and ischemia by elevated VEGF levels may cause a feedback loop that promotes chronicity.

Early diagnosis and treatment after the presentation of CRVO are important for optimal visual outcomes with anti-VEGF agents. In the CRUISE study, sham patients crossing over to 0.5 mg of ranibizumab had fewer VA gains at month 12, compared with those receiving 0.5-mg ranibizumab from study entry.⁶

Similarly, in the COPERNICUS study, sham patients treated with aflibercept PRN during weeks 24 to 100 did not gain vision as robustly as those who started the study with fixed monthly doses of aflibercept, despite receiving a higher mean number of injections.

In clinical practice, a monthly injection is often given until the edema resolves or until two or three consecutive injections provide no further improvement. At that point, depending on patient preference, a treat-and-extend or PRN treatment protocol can be adopted.

The trials have demonstrated that long-term dosing with anti-VEGF agents is necessary to control macular edema in many eyes with CRVO, likely because continued ischemia leads to continued excessive production of VEGF.

Our Experience

In the authors’ practices, an anti-VEGF agent is used as first-line therapy for BRVO-associated macular edema with steroids reserved for refractory cases. Grid-pattern laser is considered a means of increasing the treatment interval in eyes with BRVO. Macular edema associated with CRVO is typically treated first with an anti-VEGF agent as well; similarly, steroids are generally reserved for refractory cases.

The specific indications for anti-VEGF agents, corticosteroids, and laser and comparisons of their individual, as well as combined, efficacies require further investigation. There are currently no evidence-based guidelines for the selection among these options.^27-30 Current treatment, however, is directed toward downstream sequelae of RVO and remains a temporizing treatment against a chronic eye disease.

The development of anti-VEGF agents has significantly improved our ability to treat RVO, and continued experience with these agents will likely further optimize our evolving treatment protocols. As we learn more about the pathophysiology of RVO, we will be better able to develop targeted treatment plans. RP

REFERENCES