Validation of subclavian duplex velocity criteria to grade severity of subclavian artery stenosis

Abstract

BackgroundValidation of subclavian duplex ultrasound velocity criteria (SDUS VC) to grade the severity of subclavian artery stenosis has not been established or systematically studied. Currently, there is a paucity of published literature and lack of practitioner consensus for how subclavian duplex velocity findings should be interpreted in patients with subclavian artery stenosis. ObjectiveThe objective of the present study was to validate SDUS measurements using subclavian conventional or computed tomography angiogram (subclavian angiogram [SA])-derived measurements. Secondary objectives included measuring the correlation between SDUS peak systolic velocities and SA measurements, and to determine the optimal cutoff value for predicting significant stenosis (>70%). MethodsThis is a retrospective review of all patients with suspected subclavian artery stenosis and a convenience sample of carotid artery patients who underwent SDUS and SA from May 1999 to July 2013. SA reference vessel and intralesion minimal lumen diameters were measured and compared with SDUS velocities obtained within 3 months of the imaging study. Percent stenosis was calculated using the North American Symptomatic Carotid Endarterectomy Trial method for detecting stenosis in a sufficiently large cohort. Receiver operating characteristic curves was generated for SDUS VC to predict >70% stenosis. Velocity cutoff points were determined with equal weighting of sensitivity and specificity. ResultsWe examined 268 arteries for 177 patients. The majority of the arteries were for female patients (52.5%) with a mean age of 66.7 ± 11.1 years. Twenty-three arteries had retrograde vertebral artery flow and excluded from further analysis. For the remaining 245 arteries, the average peak systolic velocity was 212.6 ± 110.7 cm/s, with a range of 45-626 cm/s. Average stenosis was 25.8% ± 28.2%, with a range of 0% to 100%. Following receiver operating characteristic analysis, we found a cutoff value of >240 cm/s to be most predictive of >70%. Area under the curve was 0.94 with 95% confidence intervals of 0.91 to 0.97. The sensitivity and specificity for predicting >70% stenosis was 90.9 and 82.5%, respectively. ConclusionsIn patients with known or suspected disease involving the great vessels, a subclavian artery flow velocity exceeding 240 cm/s seems to be predictive of significant subclavian stenosis. Thus, we propose new SDUS VC, for predicting subclavian artery stenosis. However, because of the use of a convenience sample, it is possible that the current proposed cutoff point might need to be adjusted for other populations.