The purpose of this study was to compare our experience with duplex ultrasonography (US) and computed tomography (CT) for the routine follow-up of patients after endovascular repair (EVAR) of abdominal aortic aneurysm (AAA). We reviewed the electronic charts and radiologic exams of the first 125 patients (113 males, 12 females, median age of 76 years, range 48-98 years) with AAA treated by EVAR from June 1996 to November 2001. Our follow-up protocol included serial CT and US at regular intervals after the procedure (before discharge, at 1 month, and then every 6 months). Adequacy of each exam, ability to detect endoleaks, measurements of AAA diameter, and ability to determine graft patency were compared. For endoleak detection, comparison between CT and US was done using CT as the gold standard. A total of 608 exams, 337 CTs and 271 US, were performed 1 day to 5 years after endovascular aneurysm repair; 98% of CT and 74% of US were technically adequate. Contrary to CT, the proportion of adequate US exam was significantly less in patients with higher body mass index (BMI > or = 30 = 54% vs. BMI < 30 = 81%, p < 0.001) and for pre-discharge US compared to the post-discharge US (54% vs. 88%, p = 0.0005). Concurrent scan pairs were obtained in 252 instances in 107 patients (1-8 pairs per patient). Excellent correlation between AAA diameter measured on CT and US was noted (correlation coefficient of 0.9, p < 0.0001). However, agreement was poor. CT anteroposterior (AP) and transverse measurements were on average 2.9 mm (95% limits of agreement = -7 to 13 mm) and 1.8 mm (95% limits of agreement = -9 to 12 mm) greater than US. For AAA diameter change, there was no case of increase AP diameter on CT. However, in 23% (29/128 pairs of sets) of US, an increase in AAA size that could have influenced patient management (> or = 4 mm) was reported despite no change demonstrated on CT. For endoleak detection, sensitivity and specificity of US compared to that of CT was 25% and 89%. Similar sensitivity and specificity were noted when we excluded the first set (25% and 95%), sets done prior to 2000 (30% and 89%), inadequate CT or US scans (31% and 98%), or duplicate sets of results for each patient (28% and 81%). Of the 27 endoleaks missed on US in 17 patients, 2 were type I endoleaks. None of the four endoleaks seen only on US were type I endoleak. US usefulness prior to discharge was reduced by the high rate of inadequate exam, especially in obese patients. Despite the excellent correlation in AAA diameter between US and CT, there was significant disagreement in AAA diameter measurement and diameter change. Sensitivity of nonstandardized US for endoleak was low compared to CT. CT remains our primary imaging study after EVAR, but standardization of post-EVAR US technique may improve its accuracy.