Risk factors for surgical site infection after pancreatic surgery: a better postoperative antibiotic strategy is possible.

INTRODUCTION: Pancreatic surgery is associated with high morbidity, mainly due to infectious complications, so many centres use postoperative antibiotics (ATBpo) for all patients. However, antibiotic regimens vary according to local practices. The aims of this study were to describe the occurrence of surgical site infection (SSI) and ATBpo prescription after pancreatic surgery, and to determine the risk factors of postoperative SSI, in order to better define the clinical indications for ATBpo in this context. PATIENTS AND METHODS: All patients undergoing scheduled major pancreatic surgery from January 2007 to November 2018 were included in this retrospective study. Patients were classified into four groups according to SSI and routine ATBpo prescription: SSI+/ATBpo+, SSI-/ATBpo+, SSI+/ATBpo- and SSI-/ATBpo-. In addition, risk factors (fever and pre-operative biliary prosthesis) associated with the occurrence of SSI and ATBpo were analysed using a logistic regression model. RESULTS: Data from 149 patients (115 pancreaticoduodenectomies and 34 splenopancreatectomies) were analysed. Thirty (20.1%) patients experienced SSI and 42 (28.2%) received ATBpo. No difference was found in routine ATBpo prescription between patients with and without SSI (26.7% vs 28.6%, respectively; P=0.9). Amongst the 107 patients who did not receive routine ATBpo, 85 (79.4%) did not develop an SSI. In-hospital mortality did not differ between infected and uninfected patients (7% vs 2%, respectively; P=0.13). The occurrence of postoperative fever differed between SSI+ and SSI- patients (73.3% vs 34.2%, respectively; P<0.001), while the prevalence of pre-operative biliary prosthesis was similar (37.9% vs 26.7%, respectively; P=0.3). CONCLUSION: Non-routine ATBpo after major pancreatic surgery resulted in 85 (56%) patients being spared unnecessary antibiotic treatment. This suggests that routine ATBpo prescription could be excessive, but further studies are needed to confirm such antibiotic stewardship. Fever appears to be a relevant clinical sign for individual-based prescription, but the presence of a biliary prosthesis does not.