Do Corticosteroids Improve Refractory Diabetic Macular Edema Outcomes?

Effect of Adding Dexamethasone to Continued Ranibizumab Treatment in Patients With Persistent Diabetic Macular Edema: A DRCR Network Phase 2 Randomized Clinical Trial

Maturi RK, Glassman AR, Liu D, et al; Diabetic Retinopathy Clinical Research Network
JAMA Ophthalmol. 2018;136:29-38.

Study Summary

The Diabetic Retinopathy Clinical Research Network conducted a phase 2, multicenter, randomized, prospective clinical trial examining the effects of adding intravitreal dexamethasone to ranibizumab in 116 patients (129 eyes) with refractory diabetic macular edema (DME).

The study's primary and secondary endpoints were mean change in visual acuity and central subfield thickness (CST), respectively, at 24 weeks. All patients had to have persistent DME, for which they received at least three intravitreal anti-vascular endothelial growth factor (anti-VEGF) injections in the 20 weeks prior to study inclusion. Researchers initially only included pseudophakic eyes, but due to a slow enrollment phase, the study was opened to phakic eyes as well.

All patients underwent a run-in phase consisting of three monthly injections of ranibizumab. If the DME persisted after this, patients were then enrolled and randomly assigned to either ranibizumab 0.3 mg plus sham injection or to ranibizumab 0.3 mg plus intravitreal dexamethasone 700 ųg.

At the study's beginning, all eyes received ranibizumab and then a sham or dexamethasone injection within 8 days. Each patient was then evaluated at 4-week intervals through week 24. At weeks 4 and 8, all patients received ranibizumab. If at weeks 12 through 20 the patients met the retreatment criteria, they underwent another ranibizumab plus sham or ranibizumab plus dexamethasone injection series.

Patients received a maximum of two sham or dexamethasone injections. The mean number of ranibizumab injections was 5.6 in the combination cohort and 5.7 in the ranibizumab-only cohort. In the combination cohort, 97% of patients received a second dexamethasone injection; and in the ranibizumab-only cohort, 96% of patients received a second sham injection.

At 24 weeks, the mean improvement in visual acuity was 2.7 letters in the combination cohort and 3.0 letters in the ranibizumab-only cohort (P = .73). There was also no difference found when controlling for confounding factors such as lens status. The patients in the combination group had a greater chance of improving >15 letters (P = .03), but they also had an increased chance of losing 10 or more letters (P = .09). At 24 weeks, the patients in the combination cohort had a greater decrease in CST (P <.001). Last 29% of patients who received dexamethasone experienced an increase in intraocular pressure (IOP), defined as any visit with an IOP > 30 mm Hg, >10 mm Hg increase from baseline, or if the patient was initiated on an IOP-lowering agent (P <.001).

Viewpoint

This study did not meet its primary endpoint, as the mean visual acuity in the eyes that received dexamethasone plus ranibizumab was not significantly different from in those who received the sham injection and ranibizumab. Still, there were more eyes in the combination group that improved > three lines of visual acuity.

The dexamethasone cohort also had a significant decrease in retinal thickness at 24 weeks compared with the group who did not receive dexamethasone. The patients who received dexamethasone, however, were more likely to develop elevations in IOP.

There was limited follow-up to assess cataract formation in both groups, but the authors hypothesized that cataracts were the cause of the trend toward more patients with >10-letter vision loss. The study authors also mentioned that a subanalysis showed that pseudophakic eyes might have had increased vision in the combination cohort and that phakic eyes benefited more in the ranibizumab cohort.

This study also highlights the importance of continuing intravitreal anti-VEGF injections for at least 6 months, as one third of patients in the run-in phase of this study were ineligible for randomization given that their DME had resolved following six intravitreal anti-VEGF injections.

Follow Medscape on Twitter: @Medscape