Accuracy of Computed Tomographic Angiography and Magnetic Resonance Angiography for Diagnosing Renal Artery Stenosis

Abstract

Background: Timely, accurate detection of renal artery stenosis is important because this disorder may be a potentially curable cause of hypertension and renal impairment. Objective: To determine the validity of computed tomographic angiography (CTA) and magnetic resonance angiography (MRA) compared with digital subtraction angiography (DSA) for detection of renal artery stenosis. Design: Prospective multicenter comparative study conducted from 1998 to 2001. Two panels of 3 observers judged CTA and MRA image data and were blinded to all other results. Digital subtraction angiography images were evaluated by consensus. Setting: 3 large teaching hospitals and 3 university hospitals in the Netherlands. Patients: 402 hypertensive patients with suspected renal artery stenosis were included. A group of 356 patients who underwent all 3 diagnostic tests was used for analysis. Measurements: Reproducibility was assessed by calculating interobserver agreement. Diagnostic performance was evaluated in terms of sensitivity, specificity, and other diagnostic variables. Atherosclerotic stenoses of 50% or greater and fibromuscular dysplasia were considered clinically relevant. Results: Twenty percent of patients who underwent all 3 tests had clinically relevant renal artery stenosis. Moderate interobserver agreement was found, with κ values ranging from 0.59 to 0.64 for CTA and 0.40 to 0.51 for MRA. The combined sensitivity and specificity were 64% (95% CI, 55% to 73%) and 92% (CI, 90% to 95%) for CTA and 62% (CI, 54% to 71%) and 84% (CI, 81% to 87%) for MRA. Limitations: Eighteen percent of the patients were included nonconsecutively. Digital subtraction angiography may be an imperfect reference test. Conclusion: Computed tomographic angiography and MRA are not reproducible or sensitive enough to rule out renal artery stenosis in hypertensive patients. Therefore, DSA remains the diagnostic method of choice. *For a list of the other investigators and research coordinators who participated in RADISH, see the Appendix.