Back to Annual Meeting Posters

Background: Anastomotic leak is a devastating complication following gastrointestinal (GI) surgery. Few studies have evaluated the role of CT-guided percutaneous drainage (PD) in the management of these leaks. The aim of this study was to define predictive clinical, laboratory, radiographic, or operative factors for CT-guided PD failure of symptomatic anastomotic leaks after GI surgery.
Methods: A 10-year retrospective review of an interventional radiology database was conducted to identify patients with symptomatic anastomotic leak after undergoing GI surgery. Inclusion criteria were patients having small bowel or colorectal surgery, the operating surgeon documenting clinical concern for postoperative anastomotic leak, a supporting CT demonstrating a fluid collection adjacent to an anastomosis, and the use of PD as initial therapy. Exclusion criteria included patients undergoing foregut surgery, concomitantly undergoing hepatobiliary or pancreatic anastomoses, and/or solid organ resection. Patient characteristics (clinical, laboratory, radiographic, operative) following a technically successful PD who then failed and required reoperation for anastomotic leak were compared to those successfully treated with PD. Fisher's exact test was used to analyze categorical variables and a Mann-Whitney test used for continuous variables. A forward logistic regression was utilized to identify factors independently associated with the need for reoperation using all available covariates, with a p-value <0.05 considered to be statistically significant.
Results: 61 patients met study inclusion criteria. Median age of the study cohort was 45 (range, 14-92) years, and included 32 males (53%). Indications for surgery were inflammatory bowel disease (n=25), cancer (n=15), small bowel obstruction (n=9), and other conditions (n=12). Anastomotic types included enterocolic (n=17), enteroenteric (n=16), enteroanal (n=14), colorectal (n=8), colocolonic (n=4) and coloanal (n=2). Twenty-two patients (36%) had an infraperitoneal anastomosis. 50 patients (82%) successfully underwent therapeutic PD of a perianastomotic fluid collection, with median follow-up of 16 months. 11 patients (18%), at a median interval of 16 days, required reoperation following PD. A forward logistic regression showed cardiopulmonary disease (p=0.03) and cancer surgery (p=0.01) to be factors independently associated with the need for reoperation. Level of the anastomosis, initial fecal diversion/stoma, fluid collection size and microbiology of aspirate did not predict failure of PD.
Conclusion: Cardiopulmonary disease and cancer surgery appear to be independent predictors for failure of PD and the need for reoperation following symptomatic GI anastomotic leak. For patients without these risk factors, PD is a valuable tool for managing anastomotic leaks conservatively with a high degree of success.