The Analysis Of The Risk Factors Of Post-ERCP Pancreatitis And The Establishment Of Predictive Models

Objective:To explore the risk factors of post-ERCP pancreatitis(PEP)in patients and build a predictive modelMethodsData of 312 patients who underwent ERCP in our hospital from March to June in 2014 were retrospectively investigated.The association of PEP with clinicopathological factors was analyzed using univariate and multivariate statistic analysis.The predictive accuracy of the logistic regression model was evaluated using receiver operating characteristic(ROC)curve by comparing area under the curve(AUC).ResultsThe rate of PEP was 8.33%(26/312).PEP was closely associated with liver cirrhosis,glutathione-converting enzyme level,prothrombin time,type of duodenal papilla,catheterization time of ERCP,total manipulation time,repeatedly entering the duodenal papilla,and pancreatic duct radiography,while not with age,sex,diabetes,albumin level,platelet count,dilation of bile duct,history of pancreatic disease,count of ERCP,pre-cut,incision size,balloon dilation,nasal biliary drainage(P>0.05).Multivariate analysis showed that prothrombin time,type of duodenal papilla and repeatedly entering the duodenal papilla(more than 2)were independent factors for PEP.More importantly,ROC curve analysis showed that the logistic model building by these three independent factors provided a high diagnostic accuracy of PEP(AUC,0.873;susceptibility,87.0%;specificity,93.9%).ConclusionsThe occurrence of PEP is affected by multiple independent risk factors,and the regression model composed of these independent risk factors can accurately predict the possibility of PEP.