摘要
目的探讨APACHE-Ⅱ＋肥胖指标（APACHE—O）、APACHEII、Ranson、BalthazarCT评分系统（CTSI）对急性胰腺炎（AP）预后的评估价值。方法收集2005年1月1日至2011年1月1Et321例AP患者临床资料，分别计算人院48h内APACHE—O、APACHE-11、Ranson及CTSI评分。用受试者工作曲线（ROC）和曲线下面积（AuC）分析比较上述评分系统判断重症AP的预测价值，并分析它们对局部并发症、全身并发症及病死的预测能力。结果取诊断阈值为7时，APACHE-O评分对重症AP预测的敏感性、特异性和准确性分别为95．4％、76．6％和79．4％，约登指数和AUC分别为0．720和0．736；取诊断阈值为8时，APACHE-Ⅱ评分对重症AP预测的敏感性、特异性和准确性分别为90．4％、81．0％和82．6％，约登指数和AUC分别为0．714和0．699；取诊断阈值为3时，Ranson评分对重症AP预测的敏感性、特异性和准确性分别为75．00A、78．1％和77．6％，约登指数和AUC分别为0．531和0．703；取诊断阈值为5时，CTSI评分对重症AP预测的敏感性、特异性和准确性分别为82．7％、91．4％和90％，约登指数和AUC分别为0．741和0．778。结果提示评分系统的准确性：CTSI〉APACHE—O〉Ranson和APACHE-Ⅱ。CTSI对AP局部并发症预测能力最高，约登指数和AUC分别为0．766和0．777；取诊断阈值为5时，敏感性、特异性和准确性为85．4％、91．2％和90．3％。APACHE—O对AP全身并发症预测能力最高，约登指数和AUC分别为0．789和0．779；取诊断阈值为8时，诊断的敏感性、特异性和准确性为91．1％、87．8％和88．2％。CTSI对AP死亡发生的预测能力最高，约登指数和AUC分别为0．952和0．847；取诊断阈值为8时，敏感性、特异性和准确性为100％、95．2％和95．3％。结论CTSI评分系统判断重症AP有较高预测能力，CTSI评分对AP局部并发症和死亡预测能力最高，APACHE—O对AP全身并发症预测能力最高。
Objective To investigate the values of the APACHE-O,APACHE-m, Ranson and Balthazar CT（CTSI） scoring systems in predicting prognosis of severe acute pancreatitis. Methods Data were collected prospectively from 321 consecutive patients who were admitted into our hospital from 2005-01-01 to 2011 01 01 with acute pancreatitis （AP）. The sensitivity, specificity and accuracy of the APACHE-O, APACHE-Ⅱ , Ranson, Balthazar CT scoring systems at different cutoff levels were calculated. The receiver-operating curves （ROC） for the prediction of severe AP in the early pe riod were calculated using the APACHE-O,APACHE-Ⅱ , Ranson and Balthazar CT scores in differ- ent cut-off levels on hospital admission. The area under the curve （AUC） was used to compare the predictive accuracy. Using ROC curves, the values in predicting systemic complications, local compli- cations and morbidities were also compared. Results At a cut-off point of 7, the APACHE-O had a sensitivity of 95.4% , a specificity of 76.6% and an overall accuracy of 79.4%. The Youden＇s index and the AUC of the APACHE-O score were 0. 720 and 0. 736, respectively. At a cut-off point of 8,the APACHE-Ⅱ had a sensitivity of 90.4%, a specificity of 81.0% and an overall accuracy of 842.6%. The Youdenrs index and the AUC of the APACHE-Ⅱ were 0. 714 and 0. 699, respectively. At a cut-off point of 3, the Ranson had a sensitivity of 75.0 %, a specificity of 78.1% and an overall accuracy of 77.6%, respectively. The Youdenrs index and the AUC of the Ranson were 0. 531 and 0. 703, respectively. At a cut-off point of 5, the CTSI had a sensitivity of 82.7%, a specificity of 91.4% and an overall accuracy of 90.0%, respectively. The Youdenrs index and the AUC of the CTSI were 0. 741 and 0. 777, respectively. The CTSI system was the best in predicting local complications with a Youdenrs index of 0. 766, and an AUC of 0. 777, respectively. At a cut-off point of 5, the CTSI had a sensitivity of 85.4%, a specificity of 91.2% and an overall accuracy of 90.3%, respectively. The APACHE-O system was the best in predicting systematic complications with a Youdenrs index of 0. 789 and an AUC of 0. 779, respectively. At a cut-off point of 8, the CTSI had a sensitivity of 91.1%, a specificity of 87.8% and an overall accuracy of 88.2%, respectively. The CTSI system was the best in predicting local complications with a Youdenrs index 0. 952 and an AUC of 0. 847, re- spectively. At a cut-off point of 8, the CTSI had a sensitivity of 100%, a specificity of 95.2%, and an overall accuracy of 95.3%, respectively. Conclusions The results suggested that the CTSI is the most useful system in predicting local complications and morbidities of severe AP in the early period. The APACHE-O is most useful in predicting systemic complications of severe AP.
出处
《中华肝胆外科杂志》
CAS
CSCD
北大核心
2012年第9期680-683,共4页
Chinese Journal of Hepatobiliary Surgery
基金
温州市科学技术局资助项目（Y20110104）
关键词
胰腺炎
急性坏死性
研究设计
预后
Pancreatitis
Acute necrotizing
Research design
Prognosis
