Open Access Research article

Selective decontamination of the gastrointestinal tract in patients undergoing esophageal resection

Franziska Näf, René Warschkow, Walter Kolb, Michael Zünd, Jochen Lange and Thomas Steffen*

Author Affiliations

Department of Surgery, Kantonsspital St. Gallen (KSSG), CH-9007, St. Gallen, Switzerland

For all author emails, please log on.

BMC Surgery 2010, 10:36  doi:10.1186/1471-2482-10-36

Published: 16 December 2010



Selective decontamination of the digestive tract (SDD) to eliminate gram-negative bacteria is still not widely accepted, although it reduces the incidence of nosocomial infections. In a previous retrospective study, a clear benefit to perioperative morbidity, and a reduction in nosocomial infections were found in patients who underwent an esophageal anastomosis. Thus, SDD was applied routinely for esophageal anastomoses. We report the outcome of a cohort of 81 patients who underwent this treatment.


From 2002, patients who underwent an esophageal anastomosis (esophagojejunostomy) were prospectively recorded. Perioperatively, patients received polymyxin, tobramycin, vancomycin and nystatin by mouth four times a day. Outcome was compared to a control group that was treated before 2002 (68 patients without SDD and 53 patients with SDD). Postoperative morbidity and mortality were assessed.


Between 2002 and 2007, 81 patients who underwent an esophageal anastomosis received SDD. Compared to a retrospective control group, patients with SDD had significantly less pneumonia (OR 0.06 (0.01-0.46), p < 0.001) and lower morbidity (OR 0.16 (0.05-0.49), p < 0.001). Furthermore, fewer anastomotic insufficiencies and complications were found. Similar results were found in the analysis of the patients treated before 2002.


SDD significantly reduces perioperative morbidity and mortality in patients who undergo a distal esophageal anastomosis compared to a historical control group. In patients with an anastomotic leakage, there was a strong tendency of SDD to reduce postoperative mortality.