Esophagogastrectomy: the influence of stapled versus hand-sewn anastomosis on outcome.

Successful anastomosis is essential for favorable esophagogastrectomy outcomes. Before July 2002, almost all esophagogastric anastomoses at our institution were hand-sewn. We then began using linear stapled anastomotic techniques. This review compares patient outcomes with both techniques. From July 2001 to June 2004, 280 consecutive esophagogastrectomy patients (235 men and 45 women) were reviewed (median age, 65 years). The anastomosis was intrathoracic in 206 patients (74%) and cervical in 74 (26%). Anastomoses were hand-sewn in 205 patients (73%) and linear stapled in 75 (27%). Stapled anastomoses were intrathoracic in 33 patients (16%) and cervical in 42 (57%). Anastomotic leaks occurred in 30 patients (11%); 26 (12.7%) in the hand-sewn and 4 (5.3%) in the linear stapled group (P = .008). Leaks were asymptomatic in 17 patients (57%). Dilatation was required in 70 hand-sewn anastomoses (34%) and in 11 stapled (14.6%) (P = .001). Hand-sewn anastomoses were more likely to leak and require dilatation; odds ratios and 95% confidence intervals were 5.35 (1.67-19.27) and 3.58 (1.66-8.34), respectively. A linear stapled anastomosis is safe and associated with both a significantly lower leak rate and the need for dilatation compared with hand-sewn anastomosis. This nonrandomized series suggests that linear stapled anastomosis is the preferred technique regardless of anastomotic location.