Background Whether prolonged antibiotic prophylaxis (PAP) enhances the outcomes of pancreatoduodenectomy (PD) in patients with preoperative biliary drainage (PBD) remains controversial. The aim of this retrospective, bicentric study was to determine whether PAP reduced the incidence of infection following PD. Methods Group 1 received a single dose of antibiotic prophylaxis (SDAP), while group 2 received SDAP and PAP for five days. Postoperative complication rates were compared using bipartite propensity score matching (PSM) and weighting, Cox regression and mediation analysis. Results Of the 234 eligible patients with PBD who underwent PD (SDAP, n = 147; PAP, n = 87), 173 were matched in the PSM analysis (SDAP, n = 114; PAP, n = 59). There was no significant difference between PAP and SDAP in terms of superficial SSI (30.5 v. 22.8%; P = 0.270), deep incision SSI (3.4 vs. 5.3%; P = 0.623) or deep organ SSI (6.8 vs. 8.8%; P = 0.678). Both groups had similar rates of grade B POPF (13.6 vs. 10.5%; P = 0.554), grade C POPF (11.9 vs. 5.3%; P = 0.140), PPH (15.3 vs. 13.2%; P = 0.705) and bile leakage (5.1 vs. 1.8%; P = 0.265). PAP was linked to a higher 90-day postoperative mortality rate (13.6% vs. 4.4%; P = 0.042). According to the mediation analysis, the overall pure natural direct effect of PAP on the 90-day mortality rate had a risk ratio of 3.29 (95% CI: 1.28-25.2). Discussion Our analysis based on PSM and mediation suggested that SDAP may effectively prevent postoperative infectious complications in patients with PBD who are undergoing PD. Therefore, we do not recommend PAP routinely for such patients.
