ObjectivesSevere acute pancreatitis is a common emergency of digestive system which it's overall mortality was 20%,and the risk factors of death of SAP are not completely identify. Ranson scores, APACHE II scores and CTIS scores of Balthazar scoring system are commonly used to assess the severity of acute Pancreatitis. However, there aren't any reports both here and abroad by using the mention scores criterion to predict death of SAP. In our study, 214 patients who suffered from severe acute p ancreatitis were collected and analyzed to identify risk factors related to the mortality of severe acute pancreatitis.The aim is to evaluate the prognostic ability of the risk factors.Methods2383 patients with acute pancreatitis were admitted from four general hospitals(Including Guangzhou First Municipal People's Hospital, Shaoguan Yuebei People's hospital, Shantou central hospital and Meizhou People's hospital )from January 1990 to January 2005, including 248cases of SAP (14.6%) , 214 patients with SAP who met the criterion were divided into death group and survival group according to their outcome. The characteristics, the results of blood tests, functions of organs, Ranson scores, APACHE II scores and CTIS scores of Balthazar scoring system were analyzed, and those with significant difference between two groups were further analyzed by logistic regession. Reciever operating characteristic (ROC) curve of each independent death risk factor found by logistic regression was drawn to judge the accuracy of prediction, and the optimal cut~off value was decided.Results①Serum creatinine(Scr), serum calcium(Ca), serum glucose, aspartate aminotransferase(AST), Ranson scores, APACHE II(24h) scores, APACHE II(48h) scores and CTIS scores, the incidence rates of gastrointestinal dysfunction, respiratory dysfunction, renal dysfunction, cardiac dysfunction, Pancreatic encephalopathy, sepsis, upper gastrointestinal hemorrhage and shock of death group were significantly different from those of survival group;②Logistic regression showed that independent death risk factors of SAP included APACHE II(48h) scores, Scr and Ca;③ROC curve analysis found that Predicting accuracy of APACHE II(48h) scores was the best, the area under ROC curve was 0.797[95%CI(0.727~0.866)], and its optimal cut-off value was 17.50, with the sensitivity 0.538(0.395~0.678) and Specificity 0.926(0.874~0.961); And the AUC of Ranson scores was 0.745[95%CI(0.659~0.830)], and its optimal cut-off value was 5.5, with the sensitivity 0.538(0.395~0.678) and Specificity 0.932(0.882~0.966);And the AUC of CTIS scores was 0.702[95%CI(0.627~0.778)], and its optimal cut-off value was 4.5, with the sensitivity 0.923(0.815~0.979) and Specificity 0.404(0.327~0.484) ; And the AUC of serum calcium was 0.676 [95%CI(0.583~0.769)], and its optimal cut-off value was 16.40umol/L, with the sensitivity 0.538(0.395~0.678) and Specificity 0.864(0.802~0.913);And the AUC of Serum creatinine(Scr) was 0.820 [95%CI(0.756~0.885)], and its optimal cut-off value was 113.50umol/L, with the sensitivity 0.538(0.395~0.678) and Specificity 0.864(0.802~0.913).ConclusionsAPACHE II(48h) scores, Scr and Ca were the independent death risk factors of SAP,and theirs oPtimal cut~off values were respectively 17.50,>113.50μmol/L and <1.64mmol/L. |