Analysis of pancreatic fistula according to the International Study Group on Pancreatic Fistula classification scheme for 294 patients who underwent pancreaticoduodenectomy in a single center

The purposes of this study were to validate the value of the International Study Group on Pancreatic Fistula (ISGPF) classification scheme for pancreatic fistula (PF) and to identify predictive factors for clinically significant PF. From January 2000 to December 2007, 294 consecutive patients underwent pancreaticoduodenectomy in a single medical center. Pancreatic fistula was evaluated by the ISGPF criteria and Johns Hopkins Hospital's definition (JHH). Then, logistic regression analysis was performed to identify predictive factors for PF development. Our own management strategies with PF were also discussed. The overall incidence of PF was 19.4% (57/294) according to the ISGPF criteria, and 8.8% (26/294) using the JHH definition. Thirty-one patients with PF classified by the ISGPF were missed by the JHH definition. By logistic regression analysis, we found that besides the lack of cardiovascular disease and malignant diseases, our single-layer continuous circular invaginated pancreaticojejunostomy was another independent factor for the lowered incidence of PF. The ISGPF classification scheme was accurate for evaluating PF. Single-layer continuous circular invaginated pancreaticojejunostomy may be a promising method that may have been responsible for the lower incidence of PF in this study.