Accuracy of duplex sonography scans after renal artery stenting

Abstract

Reports of duplex sonography scan criteria for recurrent renal arterial (RA) stenosis after endoluminal stenting have suggested that criteria for native arteries may overestimate recurrent disease. This retrospective report examines the utility of renal duplex sonography (RDS) scans to define the presence of significant (ie, ≥ 60%) renovascular disease (RVD) after percutaneous angioplasty and endoluminal stenting (PTAS). Demographic, duplex, and angiographic data were reviewed and compared. RDS was obtained. Peak systolic velocities (PSV) were obtained after PTAS from multiple sites along the main RA from both anterior and flank approaches. Comparable images from digital subtraction angiography were independently examined for restenosis. Percent diameter stenosis was determined from the site of maximal stenosis compared with the normal RA distal to the stent. Sensitivity and specificity were estimated and 95% confidence intervals (CIs) were computed after adjusting for within patient "clustering" of observations applying native RA RDS criteria using angiography as the gold standard. Receiver operating characteristic (ROC) curves were used to estimate the optimal RDS values for recurrent stenosis. From October 2003 to June 2009, 49 patients had angiographic imaging after PTAS. There were 30 patients (18 women, 12 men; mean age, 71 ± 9 years) provided technically adequate paired angiographic and RDS assessment after PTAS for 66 RAs. Paired analysis was performed for 23 RAs after primary PTAS and 43 RAs after secondary treatment. The prevalence of significant restenosis was 35% (23 of 66 RAs). RAs with greater than 60% diameter restenosis had higher peak systolic velocity (PSV) compared to those without (2.48 ± 1.15 millisecond vs 1.44 ± 0.58 millisecond; P < .001). Compared to angiography, RA-PSV ≥ 1.8 millisecond with distal RA turbulence demonstrated a sensitivity of 73% (95% CI, 54%, 91%), specificity of 80% (95% CI, 67%, 93%), and an overall accuracy of 77% (95% CI, 67%, 88%) with a positive predictive value of 64% (95% CI, 46%, 82%). Optimal RDS value estimated by ROC curve resulted in RA-PSV of 2.5 millisecond which was associated with a sensitivity of 59% (95% CI, 36%, 82%), specificity of 95% (95% CI, 89%, 100%), an accuracy of 83% (95% CI, 74%, 92%), and a positive predictive value of 87% (95% CI, 68%, 100%). Renal duplex sonography has utility to detect significant restenosis after PTAS. RDS criteria for significant native RA stenosis compare favorably with optimal RDS criteria for restenosis estimated by ROC curves.