Anastomotic leakage after resection and bypass for esophageal cancer: lessons learned from the past.

A retrospective study of anastomotic leakage has been undertaken in 730 patients who had resection or bypass for carcinoma of the esophagus during the period 1964-1982 at the Department of Surgery, University of Hong Kong, Queen Mary Hospital, Hong Kong. Partial or complete gangrene of the substitute loop also resulting in anastomotic disruptions were excluded from this series. Anastomotic leakage due to suture line failure occurred in 182 patients (24.9%). Two factors were found by multivariate analysis to influence leakage: the type of operation and the choice of organ used as substitute. Leakage occurred more than twice as often in bypass (42.7%) than in resection (18.3%). When the substitute used for reconstruction was viable, jejunum was associated with the lowest incidence of leakage followed by whole stomach, distal stomach, and colon in that order. The risk of leakage for any combination of the type of operation (resection or bypass) and substitute loop used was calculated. The probability of leakage was lowest when a resection was performed and jejunum was used as substitute. In view of the simplicity and relative safety of using the whole stomach, esophagectomy followed by gastric reconstruction is still the procedure of choice for the majority of patients. A bypass procedure using colon as substitute has the highest leakage rate. A low leakage rate should now be obtained, otherwise nonoperative therapy has a legitimate claim as the preferred alternative treatment modality.