Use of a pancreatic fistula risk score system for patients with clinically relevant postoperative pancreatic fistula after pancreaticoduodenectomy

Abstract

Objective To study the use of a preoperative predictive scoring system established by the Beth Israel Deaconess Medical Center, Washington University School of Medicine and Hospital of the University of Pennsylvania for patients with clinically relevant postoperative pancreatic fistula (CR-POPF) after pancreaticoduodenectomy. Methods The clinical data of 394 patients who underwent pancreaticoduodenectomy at the Provincial Hospital Affiliated to Anhui Medical University from September 2007 to December 2015 were retrospectively analyzed. The four indexes including the gland texture, pathology, pancreatic duct diameter and intraoperative blood loss were calculated for the predictive score system using the logistic regression test. The factors associated with CR-POPF were analyzed. The sensitivity and specificity of the predictive scoring system were determined by the receiver operating characteristic (ROC) curve analysis. Results Of the 70 patients who were diagnosed to have postoperative pancreatic fistula (POPF), 34 were CR-POPF, which included 36 with grade A, 23 with grade B and 11 with grade C. Univariate analysis showed that male, preoperative serum total bilirubin level ≥170 mmol/L, pancreatitis or pancreatic cancer, portal vein invasion, soft pancreatic texture, main pancreatic duct diameter ≤3 mm, and pancreaticojejunostomy were significantly related to POPF after pancreaticoduodenectomy (P<0.05). Portal vein invasion, pancreatic texture and main pancreatic duct diameter were the risk factors of CR-POPF after pancreaticoduodenectomy (P<0.05). Multivariate analysis showed the independent risk factors associated with POPF were male, preoperative serum total bilirubin level ≥170 mmol/L, soft pancreatic texture and main pancreatic duct diameter ≤3 mm (P<0.05), while soft pancreatic texture and main pancreatic duct diameter ≤3 mm were the independent risk factors of CR-POPF (P<0.05). There were significant differences in the clinical relevant postoperative pancreatic fistula rates among the negligible risk, low risk, intermediate risk, and high risk patients with CR-POPF (P<0.05). The results of ROC curve analysis showed that the sensitivity and specificity of the Fistula Risk Scoring system were 76.5% and 95.8%, respectively. The nomogram showed the area under the curve was 0.913 (95%CI: 0.858~0.968). Conclusion The preoperative predictive scoring system accurately predicted the occurrence of CR-POPF. Key words: Pancreatic fistula risk score; Pancreaticoduodenectomy; Pancreatic fistula; Prediction