The aim of this study was to critically evaluate acute and long-term complications of hand-sewn and semimechanical cervical esophagogastric anastomosis following resection of primary esophageal adenocarcinoma. Between February 1991 and 2001, 91 consecutive patients underwent subtotal esophagectomy (transthoracic, n=49; transhiatal, n=42), transposing a gastric tube based on the right gastroepiploic artery. All esophagogastric anastomoses were performed in the left neck using a hand-sewn technique (n=53) and, from September 1997, a side-to-side semimechanical technique (n=38). Outcomes evaluated were anastomotic leak rates, length of stay, and development of strictures. Postoperative mortality was 4.4% (all cardiopulmonary causes). Fifty-eight patients (63.7%) had an uncomplicated postoperative course, with a median postoperative length of stay of 10 days (vs. 20 days with associated morbidity; P </= 0.001). Anastomotic leaks were the leading cause of postoperative morbidity (16.5%, 15/91), and were classified into four types based on severity. The semimechanical anastomotic technique was associated with a reduced leak rate compared with the hand-sewn technique (7.9%, 3/38 vs. 22.6%, 12/53; P=0.08), although different patterns of anastomotic failure were seen following semimechanical anastomoses, with increased mediastinal and pleural sepsis. Anastomotic strictures developed in nine (17.0%) hand-sewn and three (7.9%) semimechanical anastomoses. Our conclusion was that a semimechanical technique for cervical esophagogastrostomy is associated with reduced anastomotic leak rates compared with hand-sewn anastomoses, resulting in a shorter postoperative stay. Patterns of anastomotic failure varied between each technique, possibly as a consequence of a longer cervical esophageal segment required for construction of a semimechanical anastomosis. The association between anastomotic technique and stricture development was not clear from this study.