A Comparative Evaluationof Clinical Scoring Systems In Prediction Of Severity In Hyperlipidemic Acute Pancreatitis

Object: To investigatethe clinical characteristics of hyperlipidemic acute pancreatitisand compare the value of five scoring systems,including BISAP score, Ranson score,CTSI,SIRS,HAPS, in the ability of early prediction of AP severity,localcomplications,organ failure,and the ability to assess the comparative mortality. As well asto explore the advantages and disadvantages of these scoring systems in predicting thehyperlipidemic acute pancreatitis severity.Methods: This study retrospectively analyzed data of909acute pancreatitis caseswhich collected from Shanghai First People’s Hospital and Shanghai Tenth People’sHospital during September2008to March2014. Scored BISAP, Ranson, CTSI, SIRS,HAPS for all APpatients.Compared the clinical characteristics between hyperlipidemicacute pancreatitis and non hyperlipidemicacute pancreatitis.Compared the difference ofdistribution of each score in mild,moderately severe, sever acute pancreatitis.By receiveroperating characteristic curve (ROC), the ability of BISAP, Ranson, CTSI, SIRS inpredicting SAP, local complications, organ failure, and mortality was analyzed inhyperlipidemic acute pancreatitis group and non hyperlipidemicacute pancreatitis group.The ability of predict SAP of BISAP, Ranson, CTSI scoring systems was constrastedbetween hyperlipidemic acute pancreatitis group and non hyperlipidemicacute pancreatitisgroup. By kappa identity test, the ability of harmless pancreatitis predicting mild acutepancreatitis was evaluatd in hyperlipidemic acute pancreatitis group andnon hyperlipidemicacute pancreatitis group.Results: In909AP cases, there are129cases(14.2%) diagnosed as hyperlipidemicacute pancreatitis including mild acute pancreatitis(6cases), moderately severe acutepancreatitis(41cases) and severe acute pancreatitis(20cases). In which,8cases withpseudocyst, pancreatic necrosis(9cases), pleural effusion(30cases), SIRS(33cases),14 with sustained organ failure and1death of patients.The scores of five scoring systems inmild, moderately severe, severe acute pancreatitis statistically significant differences inhyperlipidemic acute pancreatitis(P<0.05). BISAP,Ranson,SIRS, CTSI in the area underthe receiver operating characteristic curve in predicting severe acute pancreatitis were0.905,0.938,0.812,0.834,pairwisedifferences insignificant(P>0.05),Cutoff were1,2,2,4.Youden index were0.604,0.608,0.640,0.668.When predicting local complicationsAUC were0.874,0.726,0.668,0.848, AUC pairwise comparison only BISAP, Ranson withSIRS exist statisticallysignificant differences,Youdenindex were0.631,0.402,0.294,0.540.Inpredicting organ failure AUC were0.904,0.917,0.758,0.849,AUC pairwise comparison Ranson and SIRS statistically significant(P<0.05). Youdenindex were0.634,0.661,0.509,0.616.When predicting the mortality, four scoring systemwere statistically insignificant (P>0.05). In hyperlipidemic acute pancreatitis group,theHAPS has poor consistency with mild acute pancreatitis according to graded by Atlantaclassification(P>0.05).Conclusion: The hyperlipidemic acute pancreatitis patients have a younger onsetage, higher recurrence rate characteristics. Hyperlipidemic acute pancreatitis has a higherrate in pancreatic necrosis, SIRS, organ failure. BISAP, Ranson, SIRS,CTSI all have goodperformance in predicting the severity and capability similar. BISAP, Ranson, CTSIhavesimilar ability in predicting the local complications, were better than SIRS. SIRS is lesspowerful in predicting organ failure.There was no significant in these in predicting themortality.HAPS is not suitable for the prediction of mild hyperlipidemic acutepancreatitis.