O18 – Microbial Analysis to Predict Anastomotic Leak in Humans

Author(s):

Background: Our laboratory has generated compelling evidence confirming that intestinal microbes are key causative agents in the pathogenesis of intestinal anastomotic leak. In a rat model, we identified two organisms, Pseudomonas aeruginosa and Enterococcus faecalis, that express a “leak-inducing phenotype” characterized by their capacity to produce collagenase and cleave host tissue matrix metalloproteinase MMP-9 that predicts leak.

Hypothesis: The aim of this study was to determine if patients undergoing colon surgery harbor strains expressing the “leak-inducing phenotype” at the time of anastomotic construction.

Methods: Eleven (11) patients undergoing colon surgery were studied under IRB protocol approved by the University of Chicago. To isolate microbes present at anastomotic tissues during surgery, we swabbed distal and proximal ends of resected colon segments immediately after removal. Microbial analysis was performed using standard culture techniques. Phenotype analysis of cultured microbes was performed using a fluorescein-based collagenase assay and a human recombinant MMP-9 cleavage assay developed in our lab. Finally, we performed 16S rRNA analysis to survey the background microbial community in which the cultured strains coexisted.

Results: All 11 patients received IV Cefoxitin prophylaxis according to the SCIP recommendations and all recovered from surgery. Among cultured strains tested, P. aeruginosa (only present in patient #10) and E. faecalis (only present in patient #11) produced high collagenase (n=68, p <0.001). Of strains tested, only P. aeruginosa and E. faecalis cleaved human recombinant MMP-9 from its inactive pro-form to its active collagen degrading form. 16S rRNA analysis at the phyla and genera level demonstrated a balanced and diverse community of microbiota in all except 1 patient (patient #10), where Gammaproteobacteria were abundant and the health-promoting Bacteroidetes were proportionally low. Only patient #10 developed a clinical course consistent with an anastomotic leak (collection on CT, percutaneous drainage, antibiotics).

Conclusions: Data from the present study suggest that analysis of the species and phenotype of microbes that colonize anastomotic tissues, when coupled with an understanding of the background microbiota in which they coexist, might be a useful method to predict patients at risk for anastomotic leak. A larger clinical trial is underway that will be adequately powered to formally test this hypothesis.