Objective: To compare single-layered hand-sewn cervical end-to-side (ETS) anastomosis with end-to-end (ETE) anastomosis in a prospective randomized fashion. Background: The preferred organ used for reconstruction after esophagectomy for cancer is the stomach. Previous studies attempted to define the optimal site of anastomosis and anastomotic techniques. However, anastomotic stricture formation and leakage still remain an important clinical problem. Methods: From May 2005 to September 2007, 128 patients (64 in each group) were randomized between ETE and ETS anastomosis after esophagectomy for cancer with gastric tube reconstruction. Routine contrast swallow studies and endoscopy were performed. Anastomotic stricture within 1 year, requiring dilatation, was the primary endpoint. Secondary endpoints were anastomotic leak rate and mortality. Results: Ninety-nine men and 29 women underwent esophagectomy and gastric tube reconstruction. Benign stenosis of the anastomosis, for which dilatation was required, occurred more often in the ETE group (40% vs. ETS 18%, P < 0.01) after 1 year of follow-up. The overall (clinical and radiological) anastomotic leak rate was lower in the ETE group (22% vs. ETS 41%, P = 0.04). Patients with an ETE anastomosis suffered less often from pneumonia; 17% versus ETS 44%, P = 0.002 and had subsequently significantly shorter in-hospital stay (15 days vs. 22 days, P = 0.02). In-hospital mortality did not differ between both groups. Conclusion: ETS anastomosis is associated with a lower anastomotic stricture rate, compared to ETE anastomosis. However, prevention of stricture formation was at high costs with increased anastomotic leakage and longer in-hospital stay. This study is registered with the Dutch Trial Registry and carries the ID number OND1317772.