Spontaneous vertebral dissection: clinical, conventional angiographic, CT, and MR findings

The purpose of this study was to determine if typical clinical and neuroradiologic patterns exist in patients with spontaneous vertebral artery (VA) dissection.

The medical records and neuroradiologic examinations of 14 patients with spontaneous VA dissection were reviewed. The medical records were examined to exclude patients with a history of trauma and to record evidence of a nontraumatic precipitating event ("trivial trauma") and presence of possible risk factors such as hypertension. All patients underwent conventional angiography, 13 either CT or MRI (11 both CT and MRI), and 3 MRA. Conventional arteriograms were evaluated for dissection site, evidence of fibromuscular dysplasia, luminal stenosis or occlusion, and pseudoaneurysm formation. CT examinations for the presence of infarction or subarachnoid hemorrhage. MR examinations for the presence of infarction or arterial signal abnormality, and MR angiograms for abnormality of the arterial signal column.

Seven patients had precipitating events within 24 h of onset of symptoms that may have been causative of dissection and five had hypertension. At catheter angiography, two patients had dissections in two arteries (both VAs in one patient, VA and internal carotid artery in one patient), giving a total of 15 VAs with dissection. Dissection sites included V1 in four patients, V2 in one patient, V3 in three patients, V4 in six patients, and both V3 and V4 in one patient. Luminal stenosis was present in 13 VAs, occlusion in 2, pseudoaneurysm in 1, and evidence of fibromuscular dysplasia in 1. Posterior circulation infarcts were found on CT or MR in five patients. Subarachnoid hemorrhage was found on CT in two patients and by lumbar puncture alone in two patients. Abnormal periarterial signal on MRI was seen in three patients. MRA demonstrated absent VA signal in one patient, pseudoaneurysm in one, and a false-negative examination in one. Repeat catheter angiography of nine VAs at an interval ranging from 2 weeks to 1 year showed progression to occlusion in two arteries, unchanged appearance in 4, and angiographic resolution in three, which did not closely correlate with clinical outcome.

No preferred site of dissection along the course of the VA was found in this study. CT and MR examinations of the head are frequently normal in patients with VA dissections. No correlation between clinical outcome and findings at repeat angiography was demonstrated.