Duplex ultrasonography, digital subtraction angiography, and conventional angiography in assessing carotid atherosclerosis.

We prospectively compared duplex ultrasonography with digital subtraction angiography (DSA) and conventional angiography in the evaluation of carotid atherosclerosis by studying 494 arteries. Arteries were graded independently as normal, mildly stenotic, moderately stenotic, severely stenotic, or occluded. There was absolute agreement between duplex ultrasonography and DSA as to disease severity in 296 (68%) of 434 vessels, good correlation in an additional 119 vessels (28%), and poor correlation in 19 vessels (4%). When duplex studies were compared with conventional angiograms, the correlations were similar: absolute for 60 (71%) of 85 vessels, good for 17 (20%) of 85 vessels, and poor for eight (9%) of 85 vessels. With conventional angiography as the standard, duplex studies tended to overestimate the severity of disease in 64% of vessels and underestimate it in 36%, v 28% and 72%, respectively, for DSA. We concluded that duplex ultrasonography is as accurate as DSA in assessing carotid bifurcation atherosclerosis. The tendency of DSA to underestimate disease severity may lead to diagnostic errors in patients with nonobstructive plaques.