Bottom line: CTA and gadolinium-enhanced MRA are more sensitive and specific than ultrasound in diagnosing renal artery stenosis.
Summary: Ann Intern Med. 2001 Sep 18;135(6):401-11. Vasbinder GB, et al. Diagnostic tests for renal artery stenosis in patients suspected of having renovascular hypertension: a meta-analysis.
Meta-analysis of 55 studies evaluating imaging modalities for diagnosing renal artery stenosis. Studies included if they evaluated use of ultrasound, MRA, CTA, captopril renal scintigraphy, and/or captopril test in patients referred for clinical suspicion of renovascular hypertension. Reference standard test was digital subtraction angiography. Did not report on quality of studies. RESULTS: Table on pp. 404-405 reports sensitivities and specificities of tests. Gadolinium-enhanced MRA (6 studies; N=468) sens: 88-100%, spec: 75-100%; CTA (5 studies, N=734) sens: 94-100%, spec: 92-99%); ultrasound (23 studies, N=2279) sens: 17-100%, spec: 67-97%. Fig. 1 (p. 406) displays ROC curve displays the difference between CTA and gadolinium-enhanced MRA, which were nearly identical and pass through upper left quadrant of graphs indicating high sensitivity. Curve for ultrasound also passes through upper left quadrant, but indicates lower sensitivity.
Practice Recommendation: ACC/AHA recommends any of the following for detecting renal artery stenosis (Class I, Level B recommendation): duplex ultrasound, CT angiography (for patients with normal renal function), magnetic resonance angiography, and invasive catheter angiography (gold standard for diagnosis and recommended when there is high suspicion but inconclusive results from noninvasive imaging.
See Renal Artery Stenosis in DynaMed for more information