One hundred one consecutive patients underwent an esophagectomy with gastric interposition for benign and malignant processes from January 1982 through July 1990. Seventy-seven underwent transhiatal esophagectomy and 24, transthoracic esophagectomy. Multivariate analysis was performed comparing the hospitalization experience of the two groups. There was no significant difference found between the mean intraoperative blood loss for transhiatal esophagectomy (770 +/- 105 mL) and that of transthoracic esophagectomy (700 +/- 175 mL). There was a significant difference between operative time, with transhiatal esophagectomy averaging 5.4 hours and transthoracic esophagectomy averaging 7.3 hours. Postoperative stay was not significantly different although there was a wide range of values for the transthoracic esophagectomy group. An 8% operative mortality was experienced by both groups. There were a significant number of minor anastomotic leaks at the cervical anastomotic level for the transhiatal esophagectomy group, but all responded to nonoperative management. The highest morbidity and mortality were seen in the subgroup of transhiatal esophagectomies done for laryngocervical malignancies. The lowest morbidity and mortality were seen in the subgroup of 12 patients who underwent transhiatal esophagectomy for nonmalignant esophageal conditions. Transhiatal esophagectomy appears to be a safe alternative for early intrathoracic esophageal malignancies at any level, for bulky distal esophageal lesions, and for benign conditions requiring total esophagectomy.