ABSTRACT
BACKGROUND: Family physicians often evaluate patients with symptoms of possible carotid artery stenosis. The gold standard for diagnosis is digital subtraction angiography (DSA), an invasive procedure with rare but serious complications of stroke and death. In this study the authors assessed the diagnostic performance of contemporary noninvasive testing with duplex ultrasound (DUS) and magnetic resonance angiography (MRA) as compared with DSA.
POPULATION STUDIED: The investigators enrolled 350 consecutive patients with symptoms of possible carotid artery stenosis from 3 medical centers in the Netherlands. They excluded patients with contraindications to MRA such as claustrophobia or metal implants. The reasons for examination were varied: 42% had symptoms of transient ischemic attacks, 36% had symptoms of stroke, and 22% had symptoms of amaurosis fugax. The authors did not describe how these patients were referred. The average age of the patients was 67 years and 76% were male; 49% of the patients were current smokers, 49% had hypertension, and 34% had either angina or a history of myocardial infarction. Thus, although no information was given regarding the race and socioeconomic status, the patients seem similar to high-risk patients in a typical US family practice.
STUDY DESIGN AND VALIDITY: In this prospective study consecutive patients underwent DUS, MRA, and DSA examination within a maximum of 4 weeks to be evaluated for carotid stenosis. The degree of stenosis on DUS was estimated from peak systolic velocity. Digital subtraction angiography and MRA images were read by an observer unaware of clinical information and the results of other tests. Severe stenosis was defined as 70% to 99% stenosis on DSA on the side of the symptoms and measured according to criteria of the North American Symptomatic Carotid Endarterectomy Trial. Digital ultrasound and MRA results were compared with results of DSA as a gold standard. Reproducibility of the estimate of stenosis was assessed by comparison of measurement by 2 independent observers on a subsample of 170 patients.
OUTCOMES MEASURED: Sensitivity, specificity, positive predictive value, and negative predictive value were calculated for DUS and MRA. Kappa statistics estimated interobserver variability for MRA and DSA. Patient-oriented outcomes that were not addressed included cost, patient satisfaction, and side effects of DUS and MRA such as discomfort.
RESULTS: The prevalence of potentially operable severe stenosis was 46% and an additional 20% of the patients had total occlusion. For diagnosing severe stenosis, DUS had a sensitivity of 87.5% (95% confidence interval [CI], 82.1–92.9), a specificity of 75.7% (95% CI, 69.3–82.2), a negative predictive value of 87.7% (95% CI, 82.3–93.0), and a positive predictive value of 75.4% (95% CI, 68.9–82.0). Magnetic resonance angiography alone had a sensitivity of 92.2% (95% CI, 86.2–96.2), a specificity of 75.7% (95% CI, 68.6–82.5), a negative predictive value of 92.0% (95% CI, 85.8–96.1), and a positive predictive value of 76.3% (95% CI, 69.6–83.0). Both DUS and MRA tended to overestimate stenosis slightly. Digital ultrasound and MRA agreed on severe stenosis in 84% of the cases, but of these cases, DSA confirmed severe stenosis in only 81%. Kappa values (κ) for DSA and MRA were excellent (κ = 0.79 for each).
This study provides good evidence that contemporary DUS and MRA are both sensitive for potentially operable severe carotid artery stenosis. Clinicians may use either procedure to rule out severe stenosis, but should keep in mind that DSA is still necessary before operation. This report did not address cost-effectiveness or establish superiority of one noninvasive test over the other.