The authors report on eight patients with caustic esophageal burns in whom tracheoesophageal fistula (TEF) developed during dilatation programs. This study covered a period of 17 years between 1975 and 1992. The age of the patients ranged from 1.5 to 8 years (mean age, 3.4 years). TEF developed after 5 to 43 months after injury (mean, 20.05 months). In each case, after documentation of the fistula by esophagography, esophagoscopy, and/or bronchoscopy, the fistula was blocked by an intraluminal esophageal stent, a polytetrafluoroethylene (PTFE) tube with a large lumen (10-mm diameter maximum). In this period, patients were fed via a jejunostomy tube and by total parenteral nutrition (TPN) if indicated, while the existing pneumonia was being treated. In one patient, fistula closed spontaneously during the stent application program, which ended with a patent esophagus. In two patients primary closure of TEF was attempted. In one of them fistula recurred and in the other it was technically impossible to separate the esophagus from trachea safely because of the very tight adhesions. In five patients a two-stage coloesophagoplasty was performed to bypass the fistulated esophagus. In the first stage, retrosternal pull-through of the colon and coloesophagogastric anastomosis was performed. In the second stage, closure of the distal esophagus and cervical coloesophagostomy was carried out. The patient with the primary closure attempt and one patient with stage 1 coloesophagoplasty died 3 and 4 months, respectively, after the operations. The cause of death was uncontrollable pneumonia in both cases. Follow-up of the four patients showed no complications. Another fistula patient is currently on stent treatment program with pneumonia under control. The most important factor that influences survival in these patients is the degree of pneumonia caused by esophagotracheal leak. Primary repair of the fistula is technically difficult and dangerous in caustic esophageal burns because of extensive and tight adhesions between trachea and esophagus. On the other hand, an intraesophageal stent blocks the fistula, thus rendering the pneumonia curable, and also there is a good chance that the fistula will close spontaneously while the esophagus is being stented. Therefore, a stent trial is essential before a surgical attempt is made to obliterate the proximal esophagus with an esophageal bypass procedure.