HPB SO02 - Cutting surgical site infection rates in perihilar cholangiocarcinoma with informed changes to perioperative antibiotic prophylaxis: A quality improvement project in a UK hepatopancreatobiliary tertiary centre

Abstract Background Perihilar cholangiocarcinoma (pCCA) is the most common cancer of biliary epithelia. While potentially curative, surgical resection is associated with high morbidity and mortality. Surgical site infections (SSIs) are common and pernicious. Preoperative biliary instrumentation is universal: to confirm histological diagnosis and deploy stents to relieve jaundice. Patients receive multiple antibiotic courses as periprocedural prophylaxis and to treat cholangitis. They are at high risk of harbouring bacteria not quelled by perioperative prophylactic antibiotics given for other general surgical procedures. Studying the biliary microbiome, this quality improvement project sought to reduce SSIs through informed changes to local guidelines on perioperative antibiotic prophylaxis. Method Health records of all patients undergoing resection for pCCA in a UK hepatopancreatobiliary centre from February 2009 to January 2024 were analysed retrospectively (n=118). For all patients, we routinely sent intraoperative bile swabs for culture and sensitivity. Informed by these microbiology results, multidisciplinary discussion led to changes to local guidelines in March 2019. Formerly, cefuroxime and metronidazole were given on anaesthetic induction, with subsequent antibiotic choice and duration according to surgeon preference. New guidance stipulated teicoplanin and piperacillin/tazobactam be given on induction and for 72 hours postoperatively. We studied the impact of these changes on SSIs within 30 days. Results 75 patients received cefuroxime and metronidazole; 43 patients were given teicoplanin and piperacillin/tazobactam after the policy change. Proportions of patients who developed any SSI substantially fell following the guideline change (55% v 28%; p=0.0049), primarily owing to reductions in intraperitoneal collections (39% v 21%; p=0.047). Rates of deep (8% v 2%; p=0.21) and superficial incisional infections (8% v 5%; p=0.49) were not significantly affected. Associated secondary outcomes were drops in post-hepatectomy liver failure (17% v 5%; p=0.047), and 90-day mortality (16% v 5%; p=0.067), albeit the latter was not statistically significant. Conclusion Invasive investigative work-up, disease complications and exposure to multiple courses of antibiotics prior to resection of pCCA subject patients’ biliary systems to colonisation with antimicrobial-resistant bacteria. We urge centres that undertake these operations to introduce intraoperative bile cultures as a part of routine practice to inform local guidance on perioperative antimicrobial prophylaxis. We demonstrated effective antibiotics in the perioperative period can have enormous benefits to patient outcomes.