OBJECTIVE

To develop and test a scoring system to predict the risks of postoperative pancreatic fistula (POPF) after pancreaticoduodenectomy(PD).

METHODS

Clinic data and postoperative complications of the 445 consecutive patients who underwent a PD procedure between January 2008 and April 2015 in Peking University First Hospital were retrospectively collected and analyzed.The patients were randomly selected to modelling and validation sets at a ratio of 3∶1, respectively.The patient data were subjected to univariate and multivariate analysis in the modelling set of patients.A score predictive of POPF was designed and tested in the validation set.

RESULTS

POPF occurred in 88 of 334 patients(26.4%) in the modelling set.The multivariate analysis showed that body mass index (BMI, P<0.01) and pancreatic duct width(P=0.001) are associated with POPF independently.A risk score to predict POPF was constructed based on these factors and successfully tested.The area under the receiver operating characteristic curve were 0.829(95% CI: 0.777-0.881) on the modelling set and 0.885(95% CI: 0.825-0.945) on the validation set, respectively.

CONCLUSIONS

BMI and pancreatic duct width were associated with POPF after PD. The preoperative assessment of a patient's risk for POPF is feasible.The present risk score is a valid tool to predict POPF in patients undergoing PD, to make the selection on anastomosis types, and to take precautions against POPF.