Dexamethasone implants and neovascular glaucoma in central retinal vein occlusion

Editor, A dexamethasone intravitreal implant (Ozurdex; Allergan Inc., Irvine, CA, USA) has recently been reported in a randomized control trial (the GENEVA trial) to be effective at treating macular oedema (ME) from branch retinal vein occlusion and non-ischaemic central retinal vein occlusion (niCRVO) (Haller et al. 2010, 2011). This implant is now the UK National Institute for Health and Clinical Excellence (NICE)-approved first line treatment for both of these conditions (National Institute for Health and Clinical Excellence 2011). The GENEVA study also reported that the implant reduced the rate of retinal neovascularization and suggested that it may have a beneficial impact on ischaemia and disease progression (Haller et al. 2010). This is in contrast to the SCORE study, which previously reported that the long-term incidence of neovascularization was not reduced by the administration of intravitreal triamcinolone (Chan et al. 2011). We report a retrospective case series of five patients with clinically diagnosed niCRVO treated with the dexamethasone intravitreal implant who subsequently developed neovascular glaucoma (NVG). As in the GENEVA trial, the initial diagnosis of niCRVO was made on clinical appearance, visual acuity and the absence of a relative afferent pupillary defect (RAPD); fluorescein angiography was not routinely performed. The five patients presented with visual acuities of 6/24–6/36. NVG was diagnosed as neovascularization of the angle (NVA) visible on gonioscopy associated with raised intraocular pressure. Patient one was diagnosed with steroid-induced ocular hypertension (OHT) 8 weeks following his dexamethasone implant, but developed NVG 2 weeks later (Fig. 1). Patient two developed OHT at week twelve, but was diagnosed with NVG at week sixteen. Patient three developed NVG after his second dexamethasone implant (he had relapsed 16 weeks following his first implant). He had steroid-induced OHT after the first implant, but developed NVG 10 weeks following the second implant, that is, 26 weeks from initial injection. Patient four developed NVG at week eight, and patient five developed NVG at week twelve. Patients were managed with intraocular pressure control, intravitreal bevacizumab injections, panretinal argon laser photocoagulation and, in one case, diode laser ciliary body ablation. Extensive neovascularization of the iris associated with neovascular glaucoma 10 weeks after dexamethasone implant insertion for central retinal vein occlusion. We present a small case series of five patients who developed NVG within 6 months of dexamethasone implant treatment for clinically diagnosed niCRVO. Hayreh’s natural history studies reported that 12–33% of niCRVO become ischaemic within 4 years (Hayreh 2003), but the GENEVA trial reported a significant reduction in retinal neovascularization from 2.6% to 0.7% at 6 months (Haller et al. 2010), leading the investigators to suggest that the implant might reduce the development of ischaemic complications (Haller et al. 2010, 2011). In the SCORE study, the 36-month incidence of NVG was 8.5% at 3 years, with no difference being observed between any of the observation/treatment groups (Chan et al. 2011). All five of our patients developed NVG within 8–12 weeks of their most recent dexamethasone implant. This is similar to the timeframe in which NVG developed in the Central Vein Occlusion Study (CVOS) and is also a time at which the implant would be expected to be exerting its full effect. Thus, our case series, whilst small, suggests that the dexamethasone implant neither prevents nor delays the development of NVG. Importantly, three patients were diagnosed with NVG following an earlier diagnosis of steroid-induced ocular hypertension, suggesting either that it may be easy to overlook early NVG, particularly in a situation where 15% of patients are expected to develop steroid-induced OHT (Haller et al. 2010), or that steroid-induced OHT itself promotes ischaemic conversion. One of the assumptions that has supported the adoption of the dexamethasone implant in the UK is that treated patients may require less frequent follow-up than patients managed with ranibizumab or observation (National Institute for Health and Clinical Excellence 2011), but our study suggests that frequent, careful observation is still required to guard against NVG, and that patients with ocular hypertension should be assessed carefully for the presence of subtle iris and angle neovascularization. SRJT was supported by the UK National Institute of Health Research, and this work was supported by the Imperial College NIHR Comprehensive Biomedical Research Centre. SRJT has received advisory board fees from Allergan and Novartis, but neither organisation played any role in this work.