Utility of a risk score in the prediction of clinically relevant post-pancreaticoduodenectomy pancreatic fistula

Abstract

Introduction: Pancreatic fistula (PF) is a life-threatening complication after pancreaticoduodenectomy (PD). The pancreatic fistula risk score devised by Casadei was utilized in this study, and was determined by body mass index, pancreatic duct diameter and pancreatic pathology. A cut-off cumulative risk score ≥6 was considered a predictor for PF after PD. The risk score was applied to our local hospital data to assess its utility as a clinical tool. Methods: Data for PD patients were retrieved from a prospectively collected database. The risk score was calculated and clinically relevant pancreatic fistula (Grade B/C fistula according to the 2016 ISGPF definition) was recorded. Benign and malignant pancreatic diseases requiring PD were included in the study. Patients with incomplete risk score data were excluded. Results: The study period was between June 2012 and July 2017. Eighty patients underwent PD (male, n = 46) with mean age 62.9 years (range 35–83). Biochemical leak occurred in sixteen patients (20%) and grade B/C fistula occurred in eleven patients (13.8%). In the patients with risk score <6, there were three grade B/C fistulas. For patients with risk score ≥6, there were eight grade B/C fistulas. The sensitivity, specificity, positive predictive and negative predictive values for a risk score ≥6 in the prediction of clinically relevant PF after PD in our patient cohort were 0.73, 0.64, 0.24 and 0.94 respectively. Conclusion: The clinical utility of the pancreatic fistula score ≥6 in predicting clinically relevant PF was limited by the low sensitivity and positive predictive value when applied to our local PD patient cohort.