The Construction Of A Risk Prediction Model For Post-operative Abdominal Infection After Pancreaticoduodenectomy

Objective:The aim of this retrospective study is to develop a predictive nomogram for post-operative abdominal infection(PAI)in patients undergoing pancreaticoduodenectomy(PD),and to provide a basis for early identification of PAI.Methods:By searching the database of our hospital,we checked the registration data of 219 PD patients who were treated for various benign and malignant diseases in the Pancreatic Surgery Department of General Hospital of Northern Theater Command from July 2017 to July 2022.219 patients undergoing PD surgery were randomly divided into the training cohort and the validation cohort at a ratio of 2:1.Preoperative,intraoperative and postoperative data of patients were recorded and sorted out.Univariate and multivariate Logistic regression analysis was performed to determine the independent risk factors of abdominal infection,and then the risk prediction model was established based on the independent risk factors.The nomogram was internally and crossly validated by the development and validation cohort.The receiver operating characteristic curve(ROC)and area under curve(AUC)of the prediction model were calculated to evaluate the model discrimination,and the calibration curve was used to complete the internal verification of the model.Finally,clinical decision curve(DCA)was used to evaluate the clinical application value of the model.Results:After PD,post-operative abdominal infection occurred in 31.9%(n=70)of patients.The nomogram showed that Diabetes mellitus(P=0.009),Intraoperative blood loss≥500ml(P=0.008),Clinically relevant pancreatic fistula(P=0.004)and Drainage time≥14d(P<0.001)were independent risk factors for post-operative abdominal infection after PD.The internal and cross validation of Receiver Operating Characteristic curve was statistically significant(AUC=0.84 and 0.71,respectively).The calibration curves showed good agreement between nomogram predictions and actual observations.The decision curves showed that the nomogram was of great clinical value.Conclusion:The nomogram prediction model based on the independent risk factors of abdominal infection in patients after pancreatoduodenectomy has good prediction efficiency,which provides a basis for early identification of postoperative abdominal infection.