V. G. Aveson1, R. Brower-Sinning1, B. Firek1, B. Boone1, J. Steve1, M. Hogg1, A. Zureikat1, H. J. Zeh1, M. J. Morowitz1 1University Of Pittsburgh,Department Of Surgery,Pittsburgh, PA, USA
Introduction:
Important recent studies have detailed surprising ways in which the human microbiome contributes to human health and disease. However, contributions of the microbiome to surgical outcomes have not been studied. Here, we conducted a pilot study to investigate bacterial populations present at multiple body sites in patients undergoing pancreatico-duodenectomy (PD) for pancreatic ductal adenocarcinoma (PDA).
Methods:
We collected intraoperative swabs of pancreatic fluid, bile, and jejunal contents (PB&J) as well as postoperative fecal samples from 12 patients undergoing PD for PDA. After isolation of bacterial DNA, 16S ribosomal RNA gene sequences were sequenced on the Illumina MiSeq platform. Analysis of sequencing results with QIIME software allowed for taxonomic identification and statistical analysis of microbial communities. For each patient, we recorded preop clinical variables (biliary stent, neoadjuvant chemotherapy, antibiotics) and postop outcomes (anastomotic leak, abscess formation, wound infection, cancer recurrence, death).
Results:
Bacterial diversity was surprisingly high in PB&J samples (mean Shannon index 3.7, 2.7, and 2.7 respectively). We detected and characterized bacterial DNA within PB&J samples from 8 of 12 patients. The most abundant taxa within PB&J samples were gram-negative pathogens from the family Enterobacteriaceae (e.g. Citrobacter and Enterobacter), oral pathogens (e.g. Fusobacterium and Aggregatibacter), and intestinal commensals (e.g. Bacteroides). Correlations between PB&J samples varied. One patient had a marked abundance of Fusobacterium in each of the PB&J samples. Interestingly, a high number of DNA sequences from an Enterobacter-like species was found in the PB&J samples from a patient that developed a postop Enterobacter wound infection. Fecal samples from the 12 PDA patients were notable for a relative lack of bacterial diversity (mean Shannon diversity index 3.1), perhaps reflecting perioperative antibiotic therapy. Additionally, several of the fecal samples contained unusually large populations of gram-negative pathogens (e.g. Citrobacter and Enterobacter). One fecal sample contained a highly unusual dominant population of Enterococcus (relative abundance >90%). Six patients developed a pancreatic anastomotic leak or postop abscess. Due to small sample size, we did not identify associations between the microbiome and either postop complications or preop exposures.
Conclusion:
Microbial diversity in biologic samples collected from PD patients has not previously been studied. Notable findings in this study were the high diversity within bile and pancreatic duct fluid and the presence of oral pathogens known to cause extraoral infections. We are currently enrolling additional study subjects. Future analyses may allow us to identify associations between intraop microbial profiles and postop outcomes.