BACKGROUND: The association between bacterial infection and clinically relevant postoperative pancreatic fistula (CR-POPF) after pancreatoduodenectomy (PD) has not been fully elucidated. METHODS: Microbiological data for intraoperative bile culture (BC) and drain culture on postoperative day 4 (DC) were collected. The study population was classified into the following three groups: P1 (positive BC and DC), P2 (positive BC and negative DC), and N (negative BC). A total of 209 patients (P1: 38, P2: 72, and N: 99) who underwent PD between May 2013 and May 2018 met the inclusion criteria of the study. RESULTS: The rate of CR-POPF was significantly higher in the P1 group (34.2%) than in the P2 (12.5%; P = .007) and N groups (14.1%; P = .008). Between P1 and P2 groups, a significant difference was observed in the proportion of Enterococcus faecalis grown in BC (42.1% vs 4.2%; P < .001). Multivariable logistic regression analysis showed that the presence of E faecalis in bile was independently associated with CR-POPF. CONCLUSIONS: The detection of both bile and postoperative abdominal fluid infections can be more accurate in predicting CR-POPF. Furthermore, the presence of E faecalis in bile may serve as a novel surrogate marker for CR-POPF.
BACKGROUND: The association between bacterial infection and clinically relevant postoperative pancreatic fistula (CR-POPF) after pancreatoduodenectomy (PD) has not been fully elucidated. METHODS: Microbiological data for intraoperative bile culture (BC) and drain culture on postoperative day 4 (DC) were collected. The study population was classified into the following three groups: P1 (positive BC and DC), P2 (positive BC and negative DC), and N (negative BC). A total of 209 patients (P1: 38, P2: 72, and N: 99) who underwent PD between May 2013 and May 2018 met the inclusion criteria of the study. RESULTS: The rate of CR-POPF was significantly higher in the P1 group (34.2%) than in the P2 (12.5%; P = .007) and N groups (14.1%; P = .008). Between P1 and P2 groups, a significant difference was observed in the proportion of Enterococcus faecalis grown in BC (42.1% vs 4.2%; P < .001). Multivariable logistic regression analysis showed that the presence of E faecalis in bile was independently associated with CR-POPF. CONCLUSIONS: The detection of both bile and postoperative abdominal fluid infections can be more accurate in predicting CR-POPF. Furthermore, the presence of E faecalis in bile may serve as a novel surrogate marker for CR-POPF.