The classic approach for esophagectomy is via a combined thoracic and abdominal approach. Concerns persist regarding the adequacy of this approach as a cancer operation. A study was carried out to compare these approaches, with particular reference to complication rates and long-term survival. The charts of all adult patients undergoing esophagectomy for carcinoma at the University of Miami/Jackson Memorial Hospital between July 1991 and June 1996 were reviewed. Patients who had transabdominal resections alone or colon interpositions were excluded. Of 65 esophageal resections, 38 (58%) were performed transhiatally (THE) and 27 (42%) were performed via the transthoracic (TTE) route. Treatment groups were matched for age and site, stage, and histology of tumor. Similarly, the treatment groups were homogeneous with respect to distribution of neoadjuvant chemotherapy/radiation. The number of patients experiencing any postoperative complication was similar in both treatment groups, occurring in 22 THE (58%) and 17 TTE (63%) patients (P>0.05). Anastomotic leak occurred in five THE patients (13%) and one TTE patient (4%) (P>0.05). The single TTE patient with a leak died within 3 months without leaving the hospital. All five THE patients who developed a leak left the hospital. Although there was a tendency toward a higher percentage of patients in the TTE group to suffer respiratory failure and sepsis and a higher percentage of THE patients to experience anastomotic leak, these did not reach statistical significance. Again, although perioperative mortality tended to be higher in the TTE group, this did not reach statistical significance. Four and 5-year survival rates were similar in both groups. Whereas a 4-year cumulative survival difference of 42% for THE patients and 31% in TTE patients extended at 58 months to 28% and 8%, respectively, these did not reach statistical significance. Similarly, analysis by stage and preoperative treatment type (+/- neoadjuvant chemotherapy/radiation) failed to demonstrate any survival difference between the two groups. These findings demonstrate that there is little difference in operative morbidity and mortality between THE and TTE routes. Anastomotic leaks that occur after cervical anastomosis tend to run a more benign course. Survival data do not support routine TTE as a superior oncological operation, despite the theoretical benefit of better lymphatic clearance. We continue to advocate THE because it allows a cervical anastomosis without thoracotomy and we feel it is better tolerated by patients.