Analysis The Predictive Risk Factors Of Pancreatic Fistula Associated With Hemorrhage And Establishment Of A Dynamic Predictive Model After Pancreatoduodenectomy

Objective:The patients who developed pancreatic fistula associated with hemorrhage after PD usually have complicated medical condition and experienced rapid progression.However,recent studies have conventionally used some of the indexes that are constant to predict the occurrence of pancreatic fistula associated with hemorrhage after PD in a static way.Such prediction result cannot be updated according to the changes in the patient’s fluctuating state and examination results over time.In this study,we combined preoperative and intraoperative static data with time-relevant postoperative variables to dynamically predict pancreatic fistula associated with hemorrhage after PD has good discriminative ability and predictive accuracy.By established the model might help to guide clinical prevention and to reduce the incidence of pancreatic fistula associated with hemorrhage after PD.Method:From January 2016 to December 2021,the patients who had undergone PD at the First Affiliated Hospital of Naval Medical University Hospital were included in the study.Patients’medical records in the electronic database of our center were examined.Continuous variables were expressed as(X±S)or M(Q25,Q75),whereas categorical variables were expressed as numbers and percentages.The logistic regression model,with the use of univariable and multivariable analyses,was used to identify predominant risk factors.Variables with P<0.05 were entered into a multivariate logistic regression model to determine independent risk factors.All analyses were performed using IBM SPSS Statistics Version 26.0 statistical software and R Studio packages.The discriminatory ability of the predictive model was validated using the receiver operating characteristic(ROC)curve.Results:Of the 3027 PDs during the study period,92(3.04%)developed post-pancreatoduodenectomy hemorrhage,including 61(2.02%)pancreatic fistula associated with hemorrhage.Of these patients,223(21.1%)had diabetes and 128(12.1%)had chronic pancreatitis.The median age of these patients was 62.5(54,68)years and the median BMI was 22.5(20.5,24.6)kg/m~2.The risk-associated variables in the predictive model of pancreatic fistula associated with hemorrhage after PD included gender,history of transfusion,weight loss,ASA classification,TBil,IBil,TT,operation time,intraoperative blood loss,bile duct diameter,PFA1,PFA3,PFA5,PFA7,ΔCRP3,ΔCRP7,andΔPCT3(P<0.05).A multivariate logistic regression analysis revealed that TT<16.75s,intraoperative blood loss,≥1450ml,ASA classification greater thanⅡ,small-diameter bile duct,PFA5≥199 U/L,PFA7≥1305.5 U/L,ΔCRP7≥46.435 mg/L,andΔPCT3≥0.485 ng/m L were the significant predictive factors of pancreatic fistula associated with hemorrhage after PD.The ROC curve was statistically significant(AUC=0.902).Conclusion:Our result demonstrated that TT,ASA classification,intraoperative blood loss,small-diameter bile duct,PFA5,PFA7,ΔCRP7,andΔPCT3 were the significant predictive factors of pancreatic fistula associated with hemorrhage after PD.The ROC curve was statistically significant.The nomogram will useful for surgeons to estimate and prevent pancreatic fistula associated with hemorrhage after PD.