Background Anastomotic leakage after esophagectomy is a life-threatening complication. No comparative outcome analyses for the different treatment regimens are yet available. Methods In a single-center study, data from all esophagectomy patients from January 1995 to January 2012, including tumor characteristics, surgical procedure, postoperative anastomotic leakage, leakage therapy regimens, APACHE II scores, and mortality, were collected, and predictors of patient survival after anastomotic leakage were analyzed. Results Among 366 resected patients, 62 patients (16 %) developed an anastomotic leak, 16 (26 %) of whom died. Therapy regimens included surgical revision ( n  = 18), endoscopic endoluminal vacuum therapy ( n  = 17), endoscopic stent application ( n  = 12), and conservative management ( n  = 15). APACHE II score at the initiation of treatment for leakage was the strongest predictor of in-hospital mortality ( p  < 0.0017). Conservatively managed patients showed mild systemic illness (mean APACHE II score 5) and no mortality. In systemically ill patients matched for APACHE II scores (mean, 14.4), endoscopic endoluminal vacuum therapy patients had lower mortality (12 %) compared to surgically treated (50 %, p  = 0.01) cases and patients managed by stent placement (83 %, p  = 00014, log rank test). No other clinical or laboratory parameters significantly influenced patient survival. Conclusions Endoscopic endoluminal vacuum therapy was the best treatment of anastomotic leakage in systemically ill patients after esophagectomy in this retrospective analysis. It should therefore be considered an important instrument in the management of this disorder.