From 1985 to 1995, 60 patients with a mean age of 52 +/- 12 years [24-78] underwent colon interposition for esophageal replacement. Indications were esophageal cancer (n = 37), benign stricture (n = 13), iatrogenic esophageal fistula (n = 5), achalasia with megaesophagus (n = 3), and necrosis of a previous substitute (n = 2). Colon interposition represented only 18.5% of all operations performed for oesophageal replacement during the same period. The colon was selected because of inadequate stomach in 33 cases (55%). Long-segment conduit based on the ascending branch of the left colonic artery was the preferred method and could be used in 52 patients (86.7%). The colon was placed in the esophageal bed in 38 patients (63.3%), substernally in 21 (35%), and subcutaneously in 1. Overall operative mortality and morbidity were 8.3% and 65% respectively. Five-year survival rate was 9% in the 37 patients with esophageal cancer. Seven patients (13.5%) required one or more dilatations of the esophagocolonic anastomosis. At last follow-up, 34 patients (65.4%) had no difficulty eating. Multivariate analysis identified the conduit position in the posterior mediastinum as an independent predictor of good functional result (p = 0.0018). We conclude that colon interposition for esophageal replacement provides satisfactory and durable function; however, early mortality and morbidity are substantial.