Anastomotic leaks: what is the best diagnostic imaging study?

PURPOSE: Postoperative anastomotic leaks are one of the most devastating consequences of colorectal surgery. Diagnostic imaging for upper gastrointestinal anastomotic leaks has been evaluated and reported on extensively. No study has compared the utility and effectiveness of CT scans and water-soluble enemas for the identification of postoperative lower gastrointestinal anastomotic leaks. The present study was designed to evaluate and compare these two common radiographic imaging modalities in detecting lower gastrointestinal anastomotic leaks.
METHODS: A retrospective chart review was performed that identified 36 patients during a seven-year period who underwent reoperative surgery for a lower gastrointestinal anastomotic leak. Patient's imaging studies were classified as positive if extravasation of contrast material was demonstrated. When negative, a study was retrospectively reviewed in an attempt to identify findings suggestive of an anastomotic leak.
RESULTS: There were 36 patients identified with a postoperative lower gastrointestinal leak requiring surgical intervention. There were 28 of 36 patients (78 percent) re-explored on the basis of a radiologic study demonstrating an anastomotic leak. A total of 27 CT scans were performed, of which 4 (14.8 percent) were considered positive for an anastomotic leak. On review of the remaining negative CT scans, nine (33.3 percent) were considered descriptive positive with a large amount of fluid or air in the peritoneal cavity but without obvious extravasation of contrast. Eighteen patients were evaluated with a water-soluble enema and 15 (83.3 percent) demonstrated extravasation of contrast material. In the 26 patients with a distal anastomotic leak, 17 water-soluble enemas were performed, with 15 (88 percent) demonstrating a leak. In contrast, only 2 of 17 (12 percent) CT scans were positive in this group of patients (P < 0.001). There were ten patients who initially had a CT scan followed by a water-soluble enema. Of these patients, eight of nine (88 percent) initially had a negative CT scan but were considered to be clinically suspicious of having an anastomotic leak and subsequently had a leak demonstrated on a water-soluble enema.
CONCLUSIONS: Early intervention in patients who develop an anastomotic leak can be shown to improve the ultimate outcome, especially with respect to mortality. It is usually necessary to obtain objective tests of anastomotic integrity because of the nonspecificity of clinical signs. Our study supported the superiority of water-soluble enema to CT imaging in patients in whom both modalities were used. This difference was most pronounced for distal anastomotic leaks, whereas no radiologic imaging study proved effective in evaluating proximal anastomoses.