Screening tests for renal artery stenosis: A case‐series from an Australian tertiary referral centre

Abstract

Renal artery stenosis is a common, correctable cause of hypertension and renal impairment, with multiple screening tests available to aid diagnosis. Data assessing the accuracy of screening tests are mostly derived from tight 'experimental' protocols and the application of these tests to large-scale clinical practice is not clear. Our aim was to investigate physician preferences and diagnostic accuracy of screening tests for renal artery stenosis when applied to clinical practice in a large, Australian tertiary referral centre. We investigated all renal angiograms performed at our institution between September 2002 and September 2004, as referred by renal physicians. We accessed hospital and physician records of all patients to document demographics, clinical history, screening investigations, source of screening and angiogram results. The series involved 75 consecutive patients who had 79 screening investigations (four patients had two screening tests). The case series showed that 19 (24%) patients did not have any screening investigations prior to angiography. Duplex ultrasonography was the most utilised screening test, being used in 20 (33%) of the remaining 60 screening tests. Computed tomographic angiography (CTA) was used in 19 (32%), magnetic resonance imaging in 13 (22%) and renal scintigraphy was used in four (7%) screening procedures. Magnetic resonance angiography was the most accurate screening test with a positive predictive value of 92%, followed by duplex ultrasonography with 88% and CTA was relatively inaccurate, with a positive predictive value (PPV) of 58% (P = 0.036). Clinical suspicion alone was inaccurate with a PPV of 40%, except in previously treated renal artery stenosis (PPV 89%). Duplex ultrasonography was the most utilised screening investigation amongst the physicians of our referral base. Magnetic resonance angiography and duplex ultrasonography had good positive predictive values, while CTA may not be as reliable as previously reported when applied to a large, non-selective clinical practice.