Responsiveness of Choroidal Neovascular Membranes in Patients With R345W Mutation in Fibulin 3 (Doyne Honeycomb Retinal Dystrophy) to Anti–Vascular Endothelial Growth Factor Therapy

Doyne honeycomb retinal dystrophy (DHRD), malattia leventinese, and dominant radial drusen are allelic conditions that typically present with maculopathy of early age and autofluorescent drusen in the posterior pole with late hyperpigmentation and atrophy of the retinal pigment epithelium (RPE).¹ These conditions are caused by a single R345W mutation in the gene encoding epidermal growth factor–containing fibrillin-like extracellular matrix protein 1 (EFEMP1),² also known as fibulin 3. Development of choroidal neovascularization (CNV) is an uncommon but known sight-threatening complication¹ for which the only recently reported intervention for this condition—vitrectomy and subretinal membrane removal³—requires complex surgery. Because of phenotypic similarities but pathophysiologic differences to age-related macular degeneration, we sought to determine the effects of intravitreal treatment with bevacizumab in patients with DHRD.

Retrospective review of all patients with DHRD complicated by CNV at Moorfields Eye Hospital and studied for at least 12 months was performed. Three patients met these criteria and 2 patients received intravitreal bevacizumab (1.25 mg in 0.05 mL). All patients had confirmed disease-causing mutation in fibulin 3. The study followed the Declaration of Helsinki and was approved by the institutional review board. All patients were informed during the consent process of the off-label nature of intravitreal bevacizumab. Retreatment criteria were recurrent subretinal or intraretinal fluid with or without a decrease in visual acuity.

Best-corrected visual acuity (VA) using Early Treatment Diabetic Retinopathy Study and Snellen charts at baseline, 6 months, and final follow-up was recorded. Fluorescein angiography and spectral-domain optical coherence tomographic (OCT) imaging were used to assess macular fluid using the Topcon 3D OCT-1000 (Topcon Medical Systems, Oakland, New Jersey).

A 56-year-old woman with DHRD and a history of stable vision presented in March 2008 with a 7-week history of deteriorating vision, central scotoma, and distortion. Her VA had decreased from 20/20 to 20/250 OS (Table). The clinical examination, fluorescein angiography, and OCT findings were consistent with occult CNV (Figure 1A-C). Six weeks after the first intravitreal bevacizumab injection, VA improved to 20/160 OS and there was minimal intraretinal fluid on OCT. A second intravitreal bevacizumab injection was given. By October 2008, VA was 20/100 OS but a small hemorrhage was noted well away from the fovea without leakage on fluorescein angiography (Figure 1D-F). She was managed conservatively, and by latest follow-up in June 2009, her hemorrhage had resolved without evident leakage. Visual acuity remained stable at 20/25 OD and 20/50 OS.

A 39-year-old woman presented with a 2-week history of reduced VA in the left eye in March 2007. Visual acuity was 20/200 OS and 20/20 OD (Table). Subretinal hemorrhage and retinal thickening consistent with subfoveal occult CNV was confirmed. After the first intravitreal bevacizumab treatment, VA improved to 20/40 OS. Examination in May 2007 revealed persistent subretinal hemorrhage while fluorescein angiography confirmed central leakage (Figure 2A and B). Outer retinal fluid and subretinal/RPE material were seen on OCT during the active phase of disease (Figure 2C). After 7 intravitreal bevacizumab injections were given over 2 years, the subretinal hemorrhages resolved (Figure 2D) and VA improved to 20/20 OS by June 2009.

A 41-year-old man with an extensive family history of DHRD noticed photopsias in the right eye in October 1997. Though CNV was noted at this time, visual acuity was 20/20 OU (Table). Because of a gradual decline in vision, he was evaluated for type 2 subretinal membrane surgery in 1998 at an outside hospital but deemed a poor surgical candidate. By April 1999, VA decreased to 20/80 OD owing to a disciform scar (Figure 3A). At last visit in June 2003, VA was 20/120 OD, and examination demonstrated enlargement of pigmentary changes, increased number of drusen, and expansion of subretinal tissue in the macula (Figure 3B).

Poor visual prognosis can follow conservative management of CNV in patients with fibulin 3 mutations. In our 2 cases, treatment with a course of intravitreal bevacizumab led to resolution of fluid leakage and improvement in VA whereas VA in an eye left untreated was poor at 6 years of follow-up. No adverse effects were observed in the patients who received intravitreal bevacizumab during follow-up (15 and 25 months).

Fibulin 3, or EFEMP1, is the third in a family of 6 fibulins that are extracellular-matrix proteins widely expressed in the basement membranes of epithelia and blood vessels. The exact role of fibulin 3 has yet to be determined, but R345W knock-in mice develop age-related macular degeneration–like sub-RPE deposits that may stimulate complement activation,⁴ reinforcing the parallel phenotypes observed in subjects with DHRD and age-related macular degeneration. Similarly, the basal deposits in the EFEMP1 -R345W knock-in mouse contain increased levels of the tissue inhibitor of metalloproteinase 3,⁴ which is a binding partner of fibulin 3 and capable of inhibiting vascular endothelial growth factor–mediated angiogenesis.⁵

Several findings in this study merit comparison of DHRD with age-related macular degeneration. One salient feature found in these patients was the increase in RPE changes and atrophy on follow-up examination. This occurred in all 3 patients, whether CNV existed (Figure 3C and D) or the patients were treated for their CNV and in the context of improved VA in those treated with CNV. In patient 1, the decrease in retinal edema and restoration of foveal contour on OCT (Figure 1F) helped elucidate this. Treatment with vascular endothelial growth factor inhibition helped restore retinal architecture allowing improved VA despite alterations in the RPE. Also, the presence of symptoms with relatively good VA at presentation (as in patient 3) suggests there is a window of opportunity in which intervention could be considered; however, poor VA at presentation should not by itself dissuade the clinician from considering treatment, especially if latency from time of symptoms or diagnosis is short (as in patient 2).

Our observations are interesting for 2 principal reasons. First, we show that CNV in DHRD is sensitive to treatment with intravitreal bevacizumab, providing indirect evidence that vascular endothelial growth factor may play a role in this disease. This is significant because patients who develop CNV have a poor prognosis and our observations may help guide other clinicians who are presented with this condition. Second, we show that VA recovered well following resolution of CNV, suggesting that the loss of retinal function may be reversible. This observation needs to be confirmed with a larger number of patients but, if consistent, would argue strongly for an early intervention with anti–vascular endothelial growth factor treatments when CNV complicates DHRD.

Correspondence: Dr Sohn, Ophthalmology/Retina, University of Iowa, 200 Hawkins Dr PFP, Iowa City, IA 52242 (elliott.sohn@gmail.com).