General anesthesia was used in both cases. In order for us to easily access the airway without interfering with ventilation or with the flexible bronchoscopy, a 6F introducer sheath (Cordis; Miami, FL) was placed, under bronchoscopic guidance, between the vocal cords, parallel to the endotracheal tube. Because the tip of the dilator is sharp, a guidewire was placed between the vocal cords, and the sheath-dilator combination was advanced over the wire. Once the sheath reached beyond the vocal cords, the dilator and wire were removed. An access valve for the flexible bronchoscope was attached to the endotracheal tube. Bronchography was then performed by advancing an angled 4F cobra catheter (Merritt Medical; Cleveland, OH) through the “tracheal” sheath over a guidewire (Terumo; Somerset, NJ) to the trachea, while the measuring device stays at the bronchoscope tunnel. The angle of the catheter facilitated the entry of the guidewire into the various bronchi. The catheter was always advanced over the wire to prevent damage to the airway mucosa. One milliliter of contrast medium (Ultravist; Bayer, Germany) was diluted with an equal volume of saline solution, and 1 mm of the mixture was drawn into a 10-mL syringe and injected under cineangiography guidance. During the injection, positive-pressure ventilation was administered. In this manner, a dynamic bronchogram could be recorded, delineating the bronchial stump and the multifenestrated fistula. Using a marker catheter, we precisely measured the bronchial stump diameter, so that the circular disk of the Amplatzer occluder would completely cover the whole of the stump and occlude all of the fistulous connections.