Non-invasive testing for cerebrovascular disease.

Selecting the appropriate investigation for diagnosing occlusive cerebrovascular disease depends on the availability, cost, accuracy, invasiveness and the purpose of the test. Intraarterial digital subtraction angiography remains the gold standard, but for accuracy the stenosis should be measured rather than estimated. Duplex ultrasonography is almost as accurate and can additionally analyse plaque morphology. Of 2651 duplex tests carried out in our laboratory, 12.2% were for reversible ischaemic attacks, 2.7% for amaurosis, 12.1% for cervical bruit and 4.3% for vertebrobasilar ischaemia. Duplex within 30 days of operation was carried out on 607 patients (22.9%) and surveillance on 1000 others (37.7%). Asymptomatic carotid stenosis > 60% should be confirmed by intra-arterial digital subtraction angiography, magnetic resonance angiography or spiral computed tomography angiography. For typical transient ischaemic attacks, duplex or angiography alone is adequate but when the clinical presentation is atypical, a confirming test is required. Routine preoperative brain computed tomography is not cost-effective, being equivalent to US$ 4300-11840 per perioperative stroke in our institution. Postoperative surveillance is justified, costing only US$ 505 per patient over 4 years.