Use of alternative pancreatic fistula risk score system for patients with clinical relevant postoperative pancreatic fistula after laparoscopic pancreaticoduodenectomy

Abstract

Objective: To examine the application value of alternative pancreatic fistula risk score system(a-FRS) for patients with clinically relevant postoperative pancreatic fistula(CR-POPF) after laparoscopic pancreaticoduodenectomy(LPD). Methods: Clinical data of 400 patients who underwent LPD at Department of Hepatobiliary and Pancreatic Surgery,Jilin University First Hospital,from April 2015 to August 2019 were retrospectively analyzed.There were 217 males and 183 females, with age of (M(QR)) 58 (53) years (range:26 to 93 years) and body mass index of (23.0±2.7) kg/m2 (range:19.4 to 27.1 kg/m2).Preoperative CA19-9 was (171.6±212.7) U/ml (range:32.1 to 762.6 U/ml), and preoperative CA125 was (18.6±22.9) U/ml (range:9.0 to 112.3 U/ml).Univariate analysis and multivariate Logistic regression analysis were implemented to find independent risk factors in CR-POPF.According to 3 indicators of a-FRS system(pancreatic texture,main pancreatic duct diameter,and body mass index),receiver operator characteristic curve was used to prospectively analyze the clinical value of CR-POPF. Results: CR-POPF occurred in 60 patients(15.0%) among the 400 LPD patients,including 54 patients(13.5%) with grade B pancreatic fistula and 6 patients(1.5%) with grade C pancreatic fistula.Univariate and multivariate Logistic regression analysis results showed that soft pancreas,diameter of main pancreatic duct ≤3 mm,and body mass index>23 kg/m2 were the independent risk factors for CR-POPF after LPD.The incidence of CR-POPF was 1.9% in the group with low pancreatic fistula risk(0 to 5%),5.9% with moderate pancreatic fistula risk(>5% to 20%),and 80.7% with high pancreatic fistula risk(>20%).a-FRS prospectively predicted the sensitivity and specificity of CR-POPF after LPD was 76.7% and 96.8%,positive predictive value was 80.7%,negative predictive value was 95.9%,positive likelihood ratio was 23.66,negative likelihood ratio was 0.24,and area under the curve was 0.735(95%CI:0.668-0.799). Conclusion: a-FRS system has great clinical application value in predicting CR-POPF after LPD,which can provide basis for early risk prediction of CR-POPF and timely related clinical intervention.