不同评分系统早期预测急性胰腺炎病情严重程度的比较研究

Comparative evaluation of clinical scoring systems in prediction of severity in acute pancreatitis

目的比较胰腺炎床旁严重度指数（BISAP）评分、CT评分（CTSI）、Ranson评分以及全身炎症反应综合征（SIRS）评分4种评分系统早期预测急性胰腺炎（AP）病情严重程度、局部并发症、器官衰竭以及出现死亡结局的能力。方法回顾性分析2013年6月～2015年6月上海市第十人民医院收治的1290例AP患者的各项临床资料，使用BISAP、Ranson,CTSI以及SIRS评分系统对所有AP患者进行评分，通过受试者工作特征曲线（ROE）分析比较4种评分系统预测AP病情严重程度、局部并发症、器官衰竭以及临床死亡结局的能力。结果1290例AP患者中，轻症急性胰腺炎（MAP）680例，中度重症胰腺炎（MSAP）410例，重症急性胰腺炎（SAP）200例，出现并发症包括SIRS330例，胸腔积液300例，胰腺包裹性坏死90例，假性囊肿80例，持续性器官衰竭14例，死亡11例。BIS．AP、Ranson、SIRS以及CTS14种评分在预测SAP时的ROC曲线下面积（AUC）分别为0．905（P=0．000，95％CI：0．845～0．965）、0．938（P=0．000，95％CI：0．891～0．986）、0．812（P=0．000，95％CI：0．691～0．932）、0．834（P=0．000，95％CI：0．726～0．942），组间比较差异无统计学意义（P〉0．05）；在预测局部并发症时AUC分别为0．874（P=0．000，95％CI：0．808～0．940）、0．726（P=0．000。95％CI：0．6194）．833）、0．668（P=0．000，95％CI：0．557～0．778）、0．848（P=0．000，95％CI：0．767～0．929）．组间比较仅BISAP、Ranson与SIRS之间AUC差异有统计学意义（P〈0．05）；在预测器官衰竭时分别为AUC分别为0．904（P=0．000，95％CI：0．838～0．976）、0．917（P=0．000，95％CI：0．856～0．978）、0．758（P=0．002，95％CI：0．618～0．897）、0．849（P=0．000，95％CI：0．734～0．965），组间比较Ranson、SIRS的AUC差异有统计学意义（P〈0．05）；在预测死亡时AUC分别为0．680、0．880、0．824、0．827，4种评分系统组间比较差异均无统计学意义（P〉0．05）。结论BISAP、Ranson、SIRS、CTSI4种评分系统对预测SAP均具有较好的准确性，且能力相似；BISAP、Ranson、CTSI预测局部并发症和器官衰竭的能力相似，优于SIRS；4种评分预测临床死亡的能力无明显差异。

Objective To compare the value of four scoring systems, including BISAP, Ranson score, CTSI, SIRS score in the ability of early prediction of acute pancreatitis （AP） severity, local complications, organ failure, and the ability to assess the comparative mortality. Methods The clinical data of 1290 AP in the Tenth People＇s Hospital of Shanghai City from June 2013 to June 2015 were retrospectively analyzed. BISAP, Ranson, CTSI, SIRS score were given for all patients. The difference of distribution of each score in mild, moderately severe, sever AP were compared, the ability of BISAP, Ranson, CTSI, SIRS score in predicting severity of AP, local complications, organ failure, and mortality were analyzed by receiver operating characteristic curve （ROC）. Results In 1290 AP cases, 60 cases were mild acute pancre- atitis （MAP）, 410 cases were moderately severe acute pancreatitis （MSAP）, 200 cases were severe acute pancreatitis （SAP）, 80 cases were with pseudocyst, 90 cases were with pancreatic necrosis, 300 cases were withpleural effusion, 330 cases were with SIRS, 140 cases with sustained organ failure, 10 cases dead. BISAP, AUC of Ranson, SIRS, CTSI in in predicting SAP were 0.905 （P = 0.000, 95%CI： 0.845-0.965）, 0.938 （P = 0.000, 95%CI： 0.891-0,986）, 0.812（P = 0.000, 95%CI： 0.691-0.932）, 0.834 （P = 0.000, 95%CI： 0.726-0.942）, the difference was not statistically significant （P 〉 0.05）. The predicting local complications A UC were 0.874 （P = 0.000, 95%CI： 0.808-0.940）, 0,726 （P= 0.000, 95%CI： 0.619-0,833）, 0.668 （P = 0.000, 95%CI： 0.557-0.778）, 0.848（P =0.000, 95%CI： 0.767-0.929）, AUC of BISAP, Ranson were compared with SIRS, the differences were statistically significant （P 〈 0.05）. In predicting organ failure AUC were 0.904 （P = 0.000, 95%CI： 0.838-0.976）, 0.917 （P = 0.000, 95%CI： 0.856-0.978）, 0.758 （P = 0.002, 95%CI： 0.618-0.897）, 0.849 （P = 0.000, 95%CI： 0.734-0.965）, AUC pairwise of Ranson and SIRS were compared, the difference was statistically significant （P 〈 0.05）. The predicting the mortality AUC were 0.680, 0.880, 0.824, 0.827, four scoring system were compared, the difference was not statistically significant （P 〉 0.05）. Conclusion BISAP, Ranson, SIRS, CTSI all have good performance in predicting the severity and capability similar. BISAP, Ranson, CTSI have similar ability in predicting the local complications, they are better than SIRS score. SIRS is less powerful in predicting organ failure. There is no significance in these in predicting the mortality.