Partially versus fully covered self-expandable stent for esophageal perforations and fistula: A randomized controlled study

Introduction: Previous results on stent therapy for benign esophageal perforations, anastomotic leaks and fistula (1) have led to increased use of self-expandable stents (SEMS). To date no prospective randomized trials have been published comparing different types of removable SEMS. We therefore compared the clinical efficacy of fully covered (FC) versus partially covered (PC) SEMS in patients with benign esophageal leaks. Methods: Between 2008 and 2010, we conducted a prospective randomized controlled, clinical trial in subsequent patients referred for treatment of benign esophageal leaks. After informed consent, patients were randomized to placement of an FC Hanaro stent (MI tech, Korea) or a PC Ultraflex stent (Boston Scientific, USA). SEMS removal was scheduled after a time-interval of 4 weeks. On removal the position of the upper flange was measured in cm from the dental verge and compared to the position at insertion to establish stent movement. The primary outcome measure was reintervention rate prior to scheduled removal. Secondary outcome measures included procedural complication rates, stent displacement, and primary stent removal success. Results: Thirty-eight patients (82% male); median age 65 years (range: 26-82 years) were randomized to the FC group (n=19) or PC group (n=19). Esophageal leaks included iatrogenic perforations (n=17), surgical anastomotic leaks (n=14), Boerhaave's syndrome (n=6), and a fistula of unknown origin (n=1). Placement was technically successful in all patients (100%). Thirty-five patients (92%) underwent SEMS removal. Two (5%) patients died 12 and 30 days after SEMS placement, 1 (3%) patient had progression of underlying disease impeding further intervention or stent removal. In the remaining 35 (92%) patients, no significant difference was found in reintervention rate between FC (8/18) and PC (5/17) SEMS (p=0.36). Stent movement occurred more frequent in the FC group than PC (FC n=12/18; PC n=7/17, p=0.13). Four (11%) patients needed endoscopic stent repositioning due to migration uncovering the leak. In 7 (20%) the SEMS was replaced by a second SEMS, one SEMS was replaced by a duodenal feeding tube. All 35 patients succesfully underwent SEMS removal after a median of 28 days (range 1 - 60 days). During one SEMS (PC) a selflimiting bleeding occurred and subsequently a stenosis requiring repeated dilation (n.s.). Conclusions: Temporary SEMS placement is effective in sealing benign esophageal fistulae and perforations. Although SEMS movement occurs much more frequently with fully covered as compared to partially covered SEMS, no difference was found in the reintervention rate. Since fully covered SEMS are technically easier to remove placement of a fully covered SEMS is preferred in these patients.