Complications of cerebral angiography for patients with mild carotid territory ischaemia being considered for carotid endarterectomy.

Abstract
It is essential to image the carotid bifurcation adequately in patients with symptomatic carotid territory ischaemia if they are being considered for carotid endarterectomy. Optimal resolution is achieved by selective intraarterial contrast angiography which is an invasive procedure carrying some risk. The overall risk-benefit of carotid endarterectomy is currently being investigated in several large randomised trials in Europe and North America. Because cerebral angiography is a prerequisite for carotid endarterectomy, the risks of cerebral angiography will need to be added to those of surgery when considering whether carotid endarterectomy is effective in the management of these patients. This study evaluated prospectively 382 patients with symptomatically mild carotid ischaemia who had cerebral angiography to visualise a potentially resectable lesion at the carotid bifurcation. Complications followed 14 cerebral angiograms in 13 patients (3.4%); two complications were local (0.5%), two systemic (0.5%) and 10 were neurological (2.6%). The neurological complications were transient (TIA 1, generalised seizure 1) in two patients (0.5%), reversible (stroke) in three (0.8%) and permanent (stroke) in five patients (1.3%). There were no deaths. The significant risk factors for post angiographic stroke were (1) stroke before angiography compared with transient ischaemic attacks of the eye or brain and (2) the presence of greater than or equal to 50% diameter stenosis of the symptomatic internal carotid artery; unfortunately it may be the latter patients who are most at risk of stroke as part of the natural history of their disease and therefore most in need of prophylactic carotid endarterectomy (which requires cerebral angiography). The absolute risk of post-angiographic stroke of patients for cerebral angiography using clinical evaluation and Duplex carotid ultrasound screening.