Incidence, Clinical Management and Outcomes of Esophageal Perforations after Endoscopic Dilatation

Abstract

Esophageal dilatation is performed for the treatment of anatomic and sometimes functional narrowing of the esophageal lumen caused by a variety of benign and malignant conditions. Esophageal perforation is the major complication associated with endoscopic dilatation. The object of this study was to assess the incidence of perforation, management and outcomes after endoscopic esophageal dilatation. All patients who underwent endoscopic esophageal dilatation at our institution from June 2001 to December 2006 were identified. Data were obtained by searching our prospectively collected electronic database (MEDOS AG, Langenselbold, Germany), which includes all endoscopic reports as well as discharge summaries of patients who have undergone endoscopy in our department. 248 patients (148 male, 100 female, mean age 58 years, range 14-87 years) underwent 365 esophageal dilatations: 74 pneumatic balloon dilatations for achalasia, 79 balloon dilatations for other reasons such as tumors, peptic stenosis, or post-radiotherapy stenosis and 212 dilatations with Savary bougies. The overall incidence of perforation based on the number of procedures was 2.2% (8 perforations/365 procedures) and 3.2% (8/248 patients) based on the number of patients. The perforation rate was 1.9% for bougie dilatation, 0% for balloon dilatation and 5-4% for pneumatic balloon dilatation for achalasia. Five patients were treated surgically and 3 patients were managed conservatively. The mean hospital stay was 14 days (7-33 days). One patient in whom the perforation was recognized 16 days after the dilatation died. Endoscopic esophageal dilatation is a safe procedure for the palliation of malignant strictures, for the management of benign strictures as well as for the therapy of achalasia. The perforation rate after these procedures is low. Early recognition of the perforation is associated with a good outcome. Small perforations can be treated conservatively.