A randomized trial of arterial duct occlusion with a double umbrella (DU) or wire coil (WC) was undertaken for patients <18 years of age, weighing >10 kg with isolated ducts < or = 3 mm in diameter. Baseline, procedural, and outcome characteristics were compared in an intention-to-treat analysis according to randomization group. From 40 consecutively screened patients, 2 were not enrolled due to a ductal diameter of >3 mm on initial aortography, 38 patients were randomized to either the DU (n = 20) or WC (n = 18) groups. The groups did not differ significantly with respect to age, weight, gender, duct size, type, or branch pulmonary artery diameters. Crossover occurred only in the DU group, where 4 patients (20%) had a ductal diameter of < or = 1 mm and could not be entered for umbrella placement. All remaining DU group patients had ductal diameters of > or = 1.3 mm (p <0.0001). There were no embolizations or secondary implants in the DU group, but in the WC group there was 1 early and 1 late embolization, with 6 patients (33%) with > or = 2 coils. Mean times for the procedure (DU 68+/-19 minutes; WC 65+/-27 minutes; p = 0.70) and fluoroscopy (DU 14+/-4 minutes; WC 11+/-6 minutes; p = 0.22) did not differ significantly. Angiographic duct closure was documented in 4 of 13 patients (31%) of the DU group and 4 of 18 patients (22%) of the WC group (p = 0.69). Combined with an echocardiogram, closure in 11 of 17 patients with DU (65%) and 13 of 18 patients with WC (72%) (p = 0.64) was documented before hospital discharge. One WC group patient received thrombolytic therapy for a femoral artery thrombus. Follow-up at a median of 6.5 months (range 3.2 to 37) showed closure by Doppler echocardiography in 15 of 19 patients with DU (79%) versus 14 of 18 patients with WC (78%) (p = 1.0). Thus, with a tendency toward similar procedural characteristics and outcomes, the higher cost of the DU system compared with coil implants favors the use of coils for closure of the small arterial duct.