Proposal for a sub-classification of hepato-biliary-pancreatic operations for surgical site infection surveillance following assessment of results of prospective multicenter data

Background: Surgical site infection (SSI) surveillance in Japan is based on the National Nosocomial Infection Surveillance system, which categorizes all hepato-biliary-pancreatic surgeries, except for cholecystectomy, into "BILI." We evaluated differences among BILI procedures to determine the optimal subdivision for SSI surveillance. Methods: We conducted multicenter SSI surveillance at 20 hospitals. BILI was subdivided into choledochectomy, pancreatoduodenectomy, hepatectomy, hepatectomy with biliary reconstruction, pancreatoduodenectomy with hepatectomy, distal pancreatectomy and total pancreatectomy to determine the optimal subdivision. The outcome of interest was SSI. Univariate and multivariate analyses were performed to determine the predictive significance of variables in each type of surgery. Results: 1,926 BILI cases were included in this study. SSI rates were 23.2 % for all BILI; for choledochectomy 23.6 %, pancreatoduodenectomy 39.3 %, hepatectomy 12.8 %, hepatectomy with biliary reconstruction 41.9 %, pancreatoduodenectomy with hepatectomy 27.3 %, distal pancreatectomy 31.8 %, and total pancreatectomy 20.0 %. SSI rates for hepatectomy were significantly lower than those for non-hepatectomy BILI. Risk factors for developing SSI with hepatectomy were drain placement and long operative duration, while for non-hepatectomy BILI, risk factors were use of intra-abdominal silk sutures, SSI risk index and long operative duration. Conclusions: Hepatectomy and non-hepatectomy BILI differ with regard to the incidence of and risk factors for developing SSI. These surgeries should be assessed separately when conducting SSI surveillance.