The failed gastrointestinal anastomosis: An inevitable catastrophe?

Background: There is a great deal of conflicting data regarding risk factor s for anastomotic leakage, with most studies being small and looking only a t anastomoses performed at one level of the gastrointestinal (GI) tract. Su rgeons have looked at patient and technical variables with inconsistent fin dings. The purpose of this study was to evaluate the incidence, possible pr edictive factors, and results of treatment of anastomotic dehiscence in pat ients undergoing operations at all levels of the GI tract. Study Design: We evaluated the records of 2,842 patients undergoing esophag ogastrectomy, total or partial gastrectomy, enterectomy, and partial or sub total colectomy over a 12-year period. Complete demographic data, comorbidi ty, and details regarding anastomotic technique were collected on all patie nts sustaining leaks along with diagnostic methods used, treatment modaliti es, and outcomes data. Using age and gender-matched case control methodolog y, we compared patients sustaining an anastomotic leak to those undergoing successful anastomoses. Results: Fifty-one of 2,842 patients (1.8%), ranging from 1.1% of enterecto my patients to 4.8% of total gastrectomy patients, sustained an anastomotic dehiscence. Foregut procedures were accompanied by a significantly increas ed rate of leakage, and depending on location, diagnosis was made between t he 6th and 9th postoperative day. For each procedure, deaths from factors o ther than leakage far exceeded deaths from leaks, Standard risk stratifiers did not predict occurrence of leakage. Overall, 24% of patients sustaining a leak died, and this complication necessitated multiple reoperations and significantly increased length of hospital stay. Conclusions: In view of these findings, standard preoperative strategies to prepare these patients for operation may prove unsuccessful, because minim izing the incidence of anastomotic leaks will have little overall impact on survival. In addition, efforts to accomplish early hospital discharge may prove hazardous, because many of these patients manifest their leaks later in the postoperative period than is generally assumed. Improved management of GI tract disruption, including aggressive attempts at diagnosis, ICU car e, antibiotics, and nutritional support may further increase survival in th ese patients. (J Am Coll Surg 1999;188:473-482. (C) 1999 by the American Co llege of Surgeons).