Ultrasound diagnosis of spontaneous carotid dissection with isolated Horner syndrome.

BACKGROUND AND PURPOSE Isolated Horner syndrome without associated cranial nerve palsies or ischemic symptoms is an important presentation of spontaneous internal carotid artery dissection (sICAD). Ultrasound is often used as a screening method in these patients because cervical MRI is not always available on an emergency basis. Current knowledge on ultrasound findings in patients with sICAD presenting with isolated Horner syndrome is limited. METHODS Patients were recruited from prospective cervical artery dissection databases of 3 tertiary care centers. Diagnosis of sICAD was confirmed by cervical MRI and MR angiography or digital subtraction angiography in all patients. Data on Doppler sonography and color duplex sonography examinations performed within 30 days of symptom onset were analyzed. RESULTS We identified 88 patients with Horner syndrome as the only sign of sICAD. Initial ultrasound examination was performed in 72 patients after a mean time interval from symptom onset to examination of 11 (SD 8) days. The overall frequency of false-negative ultrasound findings was 31% (22 of 72 patients). It showed stenosis >or=80% or occlusion in 34 (47%) patients, and stenosis <or=80% in 16 (22%). It visualized mural hematoma in 7 patients and intimal flap in one. Patients with normal ultrasound were less frequently smokers (9% versus 28%, P=0.034), and had less frequently migraine without aura (9% versus 43%, P=0.012) or neck pain (18% versus 62%, P=0.003) than those with pathological ultrasound findings. CONCLUSIONS Nearly one third of patients with Horner syndrome as the only sign of sICAD presented with normal ultrasound findings. These results indicate that ultrasound is not a reliable method to diagnose sICAD in patients with isolated Horner syndrome.