A clinical prediction rule for renal artery stenosis

Abstract

Background: Renal artery stenosis is a rare cause of hypertension. The gold standard for diagnosing renal artery stenosis, renal angiography, is invasive and costly.

Objective: To develop a prediction rule for renal artery stenosis from clinical characteristics that can be used to select patients for renal angiography.

Design: Logistic regression analysis of data from a prospective cohort of patients suspected of having renal artery stenosis. A prediction rule was derived from the regression model for use in clinical practice.

Setting: 26 hypertension clinics in The Netherlands.

Patients: 477 hypertensive patients who underwent renal angiography because they had drug-resistant hypertension or an increase in serum creatinine concentration during therapy with angiotensin-converting enzyme inhibitors.

Results: Age, sex, atherosclerotic vascular disease, recent onset of hypertension, smoking history, body mass index, presence of an abdominal bruit, serum creatinine concentration, and serum cholesterol level were selected as predictors. The regression model was reliable (goodness-of-fit test, P > 0.2) and discriminated well between patients with stenosis and those with essential hypertension (area under the receiver-operating characteristic curve, 0.84). The diagnostic accuracy of the regression model was similar to that of renal scintigraphy, which had a sensitivity of 72% and a specificity of 90%.

Conclusions: In the diagnostic workup of patients suspected of having renal artery stenosis, the clinical prediction rule can be considered as an alternative to renal scintigraphy. It can help to select patients for renal angiography in an efficient manner by reducing the number of angiographic procedures without the risk for missing many renal artery stenoses.