BACKGROUND: Most likely because of low statistical power, no previous studies have shown any significant association between long-term survival and anastomotic leakage in patients who have undergone gastroesophageal cancer resection.

MATERIAL AND METHODS: The present study included, prospectively and consecutively, nationwide collected patients who underwent gastroesophageal cancer resection between 2003 and 2011 in Denmark. The operation was carried out as an Ivor Lewis procedure. Only patients with intrathoracic anastomosis were included in the analysis.

RESULTS: From 2003 to 2011, 1,296 patients underwent gastroesophageal resection, and 128 (9.9 %) of these experienced anastomotic leakage. The overall 5-year survival rates in patients with and without anastomotic leakage were 20 and 35 % (P < 0.0001), respectively. After exclusion of 4 weeks mortality, the 5-year survival rate in patients with leakage was 22 % compared to 36 % in patients without anastomotic leakage (P < 0.001). After exclusion of 8 weeks mortality, the 5-year survival rate was 23 % in patients with leakage and 36 % in those without (P = 0.009). The corresponding median time of survival was 74 versus 128, 87 versus 138, and 95 versus 138 weeks, respectively. The overall hazard ratios of death after anastomotic leakage, unadjusted, and after adjusting for potentially confounding factors, were 1.59 (1.27-1.99) and 1.45 (1.14-1.84). The unadjusted and adjusted odds ratios after exclusion of 4 weeks mortality were 1.51 (1.19-1.90) and 1.41 (1.10-1.81). After exclusion of 8 weeks mortality the odds ratios were 1.38 (1.08-1.77) and 1.32 (1.02-1.71).

CONCLUSIONS: This nationwide study confirms that patients experiencing anastomotic leakage after gastroesophageal cancer resection have a significantly lower long-term survival, even following full recovery after the leakage.