317 Policy, Practice and Prophylaxis: Local Microbial Ecology and Surgical Site Infections After Pancreaticoduodenectomy

Abstract Aim Evidence increasingly supports the use of piperacillin–tazobactam (TZP) for prophylaxis in pancreaticoduodenectomy (PD). This study aims to characterise prescribing, microbiology, and resistance patterns to guide policy change in a UK unit. Method Forty-four PDs were grouped by antibiotic prophylaxis and surgical-site infection (SSI). ISGPS post-operative pancreatic fistula (POPF) grade was recorded prospectively. Routinely cultured samples (bile, stent), infection-driven cultures (drain, wound, blood) and susceptibilities were retrospectively reviewed. Results Superficial and organ-space SSI occurred in 8/44 (18%) and 17/44 (38%) patients. Organ-space SSI rates within each prophylaxis group were: TZP 4/11 (36%), Cefuroxime+Metronidazole 9/22 (41%), Co-amoxiclav 4/6 (67%), Teicoplanin/Ciprofloxacin/Metronidazole 0/5 (0%). POPF B/C rates showed a similar pattern (27%, 32%, 33%, 0%). Biliary stenting was not associated with higher rates of SSI. Routine (35 organisms; 20 patients) and infection cultures (43 organisms; 23 patients) were dominated by Enterobacterales; enterococci accounted for 20 % of isolates, and Staphylococcus aureus, anaerobes and yeasts each account for 5-10% of all cultures, whereas Pseudomonas was less frequent (3%). When both routine and infection samples were taken, organism concordance was 8/8. Only 13 infection isolates had susceptibility panels: 3/4 TZP-tested isolates were resistant (all in stented patients), while one resistant isolate was seen across cefuroxime, co-amoxiclav and teicoplanin. Conclusions Single-agent TZP may under-treat the reported spectrum of local flora. Multi-agent prophylaxis, especially in patients with biliary stents, should be considered in future policy and be supported by rapid reporting of routine bile/stent cultures. The optimal duration of antibiotics remains unclear and requires further research.