Association between biliary pathogens, surgical site infection, and pancreatic fistula: results of a randomized trial of perioperative antibiotic prophylaxis in patients undergoing pancreatoduodenectomy
Ryan J. Ellis*1, Brian C. Brajcich2, Kimberly A. Bertens11, Carlos H. Chan3, Carlos Fernandez-del Castillo4, Paul J. Karanicolas5, Shishir K. Maithel6, Bradley N. Reames7, Sharon M. Weber8, Roberto Vidri8, Henry A. Pitt9, Vanessa M. Thompson2, Mithat Gonen1, Adam C. Yopp10, Clifford Y. Ko2, Michael I. D'Angelica1
1Memorial Sloan Kettering Cancer Center, New York, NY; 2American College of Surgeons, Chicago, IL; 3University of Iowa Hospitals and Clinics, Iowa City, IA; 4Massachusetts General Hospital, Boston, MA; 5Sunnybrook Health Sciences Centre, Toronto, ON, Canada; 6Emory University Hospital, Atlanta, GA; 7University of Nebraska Medical Center, Omaha, NE; 8University of Wisconsin, Madison, WI; 9Rutgers Cancer Institute of New Jersey, New Brunswick, NJ; 10University of Texas Southwestern Medical Center, Dallas, TX; 11The Ottawa Hospital, Ottawa, ON, Canada
Background: Postoperative morbidity remains common for patients after pancreatoduodenectomy (PD), including high rates of surgical site infection (SSI) and clinically relevant postoperative pancreatic fistula (CR-POPF). Contaminated bile is known to be associated with SSI, but the role of antibiotic prophylaxis in mitigation of infectious risks is poorly defined.
Methods: Intraoperative bile cultures were collected as an optional adjunct to a multicenter, randomized phase 3 clinical trial comparing piperacillin-tazobactam with cefoxitin (standard of care) as perioperative prophylaxis in patients undergoing PD. Results of intraoperative bile cultures were compiled, including pathogen sensitivities and culture results of postoperative infections when available. Associations between culture results, SSI, and CR-POPF were assessed using chi-square tests adjusted for the presence of a preoperative biliary stent.
Results: In 778 patients overall (n=378 piperacillin-tazobactam, n=400 cefoxitin), the rate of SSI was lower in patients who received piperacillin-tazobactam compared to patients who received cefoxitin (19.8% vs. 32.8%; P<0.001). Patients treated with piperacillin-tazobactam also had lower rates of CR-POPF (12.7% vs. 19.0%; P=0.031). Biliary cultures were available for 247 patients (117 piperacillin-tazobactam, 130 cefoxitin), 231 of which had full speciation and 210 had antibiotic sensitivity data. Overall, 68 (27.5%) grew no organisms, 37 (15.0%) grew a single organism, and 142 (57.5%) were polymicrobial. Organisms resistant to cefoxitin but not piperacillin-tazobactam were present in 95 patients (45.2%). Biliary stents were present in 66.4% of patients and were associated with higher rates of contaminated bile (92.7% vs 32.5%, P<0.001) and resistant organisms (61.7%% vs 19.5%, P<0.001) compared to unstented patients. In patients harboring resistant organisms, the rate of SSI was significantly lower in patients treated with piperacillin-tazobactam (n=5, 13.5%) compared to cefoxitin (n=31, 53.5%; P<0.001). There was no difference in SSI when bile did not harbor resistant organisms (27.0% piperacillin-tazobactam vs 25.0% cefoxitin, P=0.812). Rates of CR-POPF were highest in patients treated with cefoxitin that harbored bacteria intrinsically resistant to cephalosporins (Enterobacter spp, 41.4%; Enterococcus spp, 33.3%). The presence of Enterobacter spp or Enterococcus spp was associated with CR-POPF in patients treated with cefoxitin (32.1% vs 12.5% if not present; P=0.014), but not those treated with piperacillin-tazobactam (15.6% vs 19.0%; P=0.681). Among 24 patients with postoperative infection culture data, 18 (66.7%) grew identical organisms on initial biliary culture.
Conclusion: Use of piperacillin-tazobactam as antibiotic prophylaxis reduces postoperative SSI after PD, with effects potentially mediated by biliary pathogens that are resistant to standard prophylaxis. The biliary microbiome may also influence development of CR-POPF.
Back to 2023 Abstracts