A comprehensive review of anastomotic technique in 432 esophagectomies

Background: Little consensus exists and varying outcomes are reported when the 4 most common esophagogastric anastomotic techniques are compared: circular stapled (CS), hand sewn (HS), linear stapled (LS) (longitudinally stapled anastomosis), and modified Collard (MC) (combined linear and transverse stapled anastomosis). This report analyzes outcomes of these anastomotic techniques. Methods: From July 2004 through December 2008, all intrathoracic and cervical esophagogastric anastomoses at our institution were reviewed. Results: There were 432 patients (358 men, 74 women) who underwent primary esophagogastric operations. Median age was 64 years (range, 23-90 years). The approach was an Ivor Lewis esophagectomy in 254 patients (59%), transhiatal esophagectomy in 115 patients (27%), McKeown (3-hole) esophagectomy in 49 (11%) patients, minimally invasive esophagectomy in 9 (2.1%) patients, and thoracoabdominal esophagectomy in 6 (1.4%) patients. There were 268 intrathoracic (62%) and 164 cervical (38%) anastomoses. Anastomotic techniques included LS in 260 (60%) patients MC in 67 (16%) patients, HS in 57 (13%) patients, and CS in 48 (11%) patients. Operative mortality was 3.7%. Anastomotic leak occurred in 50 patients (11%). Grade III or IV leaks occurred in 21 patients (4.9%), including 13 in the chest (4.8%) and 8 in the neck (4.9%). Grade III or IV leaks occurred in 12 patients (4.6%) with LS anastomoses, in 4 (7.0%) patients with HS anastomoses, in 3 (6.2%) patients with CS anastomoses, and in 2 (3.0%) patients with MC anastomoses. HS anastomoses had the highest odds of leakage (p = 0.01) and LS anastomoses had the lowest risk of stricture (p = 0.006). Conclusions: When performing an esophagogastric anastomosis, clinically significant leaks occur with similar frequency in both cervical and intrathoracic locations. The HS technique has the highest leak rate and the LS technique had the lowest rate of stricture formation.