Purpose Vessel deconstruction of the ophthalmic artery is occasionally required in the treatment of periophthalmic internal carotid artery (ICA) intracranial aneurysms or more rare lesions such as dural arteriovenous fistulas (AVF), arteriovenous malformations or symptomatic intraorbital ophthalmic artery aneurysms. Multiple external carotid artery (ECA) to ophthalmic artery collaterals can prevent ocular ischemia in 90% of cases. However, despite visualization of ECA collateral supply to the distal ophthalmic–retinal artery and preservation of a choroidal blush on angiographic assessment, vision loss can still occur due to retinal ischemia or delayed thromboembolic complications, presumably due to inadequate collateral retinal perfusion following ophthalmic artery sacrifice.

Materials and methods We studied three separate endovascular interventions resulting in proximal ophthalmic artery deconstruction: patient No 1 with a Cognard type 3 dural AVF treated with n-butyl cyanoacrylate glue embolization; patient No 2 with a fusiform intraorbital ophthalmic artery aneurysm treated with coil embolization; and patient No 3 with a giant periophthalmic ICA aneurysm incorporating the ophthalmic artery treated with ethylene vinyl alcohol (Onyx) liquid embolization. In patient Nos 2 and 3, ICA balloon test occlusion studies were performed across the ophthalmic artery origin prior to sacrifice. Photopic electroretinography was recorded using surface or needle infraorbital electrodes by measuring both averaged a and b waves before and after ICA balloon inflation and intermittently during ophthalmic artery sacrifice. Additionally, diffuse flash visual evoked potentials were recorded from scalp electrodes over the occipital cortex.

Results Following endovascular treatment, compromise of the ophthalmic artery was confirmed on angiography in each patient with ECA–ophthalmic artery collaterals reconstituting the retinal arteries. An intact choroidal blush was present either during ICA balloon test occlusion study or post-ophthalmic artery deconstruction. All three patients were visually and neurologically intact postprocedure. However, patient No 1 returned 5 days later with blurred vision and a central retinal artery occlusion. For patient Nos 2 and 3, visual evoked potentials remained intact throughout the procedure. However, electroretinography demonstrated 30–50% reduced baseline b values for patient No 2 during the balloon test occlusion study and after ophthalmic artery sacrifice, in contrast with preserved values in patient No 3. Patient No 2 returned 4 days later with spontaneous vision loss and branch retinal artery occlusion. Patient No 3 has remained visually intact.

Conclusion We present a novel intraoperative application of electroretinography to assess retinal perfusion following ophthalmic artery sacrifice. During a balloon test occlusion study across the ophthalmic artery origin, concomitant electroretinography may assist in predicting retinal ischemia or delayed thromboembolic complications with greater sensitivity than the angiographic presence of ECA–ophthalmic artery collaterals. Although we do not anticipate this technique to significantly alter treatment planning for ophthalmic artery deconstruction, implications remain for postprocedure management such as fluid augmentation, hypertensive therapy and utilization of dual antiplatelet medication.