The clinicopathologic factors in 50 gastrectomized and 831 nongastrectomized patients with thoracic esophageal carcinoma were compared. The differences in various factors were not statistically significant except for the average duration of operation, which suggests that the association of esophageal carcinoma and previous gastrectomy is a chance finding. Thirty-four gastrectomized patients underwent abdominal lymphadenectomy and 16 did not. The postoperative survival curves for the two groups did not differ. In comparing the clinicopathologic factors between the two groups, only the tumor location differed significantly (P < .05). Nine patients (26.5%) with lesions in the mid- or lower thoracic esophagus in the gastrectomized group had positive nodes in the abdomen; the 5-year survival rate was 39.4%. Among them, 8 had gastrectomy for peptic ulcer. Abdominal lymphadenectomy is recommended for those patients with mid- or lower thoracic esophageal carcinoma who have had previous gastrectomy without lymphadenectomy.