急性胰腺炎继发多器官功能障碍综合征预测模型的建立与评价(1)
【摘要】目的 研究早期预测急性胰腺炎继发多器官功能障碍综合征模型的建立方法及其效能。方法 回顾收集2008年10月至2011年10月苏州大学附属第一医院的113例急性胰腺炎住院患者临床资料，对若干发病早期(48 h内)指标进行单因素、多因素二分类Logistic回归分析并建立模型，通过ROC曲线对模型进行评价。结果 单因素分析表明血清乳酸脱氢酶(LDH)、血清肌酐(CR)、血清白蛋白(ABL)有统计学意义(P<0.05),纳入多因素分析得到回归模型：ln (P/1-P) = 2.243+0.002×LDH+0.014×CR-0.186×ABL，似然比检验(χ2=66.46，P=0.000)，ROC曲线分析：曲线下面积0.927(95%CI：0.875～0.980)，最佳截断值0.177(灵敏度82.14%，特异度85.88%，正确率84.96%)。结论 通过对急性胰腺炎继发多器官功能障碍综合征的早期危险因素进行多因素回归分析，可以建立高效能的预测模型。
【关键词】急性胰腺炎；多器官功能障碍综合征；Logistic回归模型；ROC曲线；乳酸脱氢酶；白蛋白；肌酐；危险因素
, 百拇医药
Establishment of regression model of acute pancreatitis complicated with secondary multiple organs dysfunction syndrome for prediction and evaluating its efficacy CHEN Du, LU Shi-qi, XU Feng. Department of Emergency, the First Affiliated Hospital of Soochow University，Suzhou 215006，China
Corresponding author: LU Shi-qi, Email: lushiqi2004@126.com
【Abstract】Objective To study the way for establishing prediction model of acute pancreatitis with secondary multiple organs dysfunction syndrome (MODS) in the early stage. Methods The linical data were collected from 113 hospitalized patients with acute pancreatitis admitted from October 2008 to October 2011. Some biomarkers within 48 h after the onset of acute pancreatitis were statistically analyzed by univariate and multivariate binary logistic regression analysis in order to establish the prediction model evaluated by the ROC curve. Results There were statistically significant differences in lactate dehydrogenase (LDH), serum creatinine (CR), serum albumin (P<0.05) identified by univariate regression analysis. These independent risk factors in the regression model were: ln(P/1-P) = 2.243+0.002×LDH+0.014×CR-0.186×ABL, likelihood ratio test(χ2=66.46，P=0.000); in the ROC curve analysis：AUC 0.927(95%CI：0.875-0.980)，the best cut-off value of the model predicting probability was 0.177 (sensitivity 82.14%，specificity 85.88%，correctly classified 84.96%). Conclusions Efficient prediction model could be set up by the logistic regression analyzing the early risk factors in multiple organ dysfunction syndrome following acute pancreatitis.
, 百拇医药
【Key words】Acute pancreatitis; Multiple organ dysfunction syndrome; Logistic regression model; ROC curve; Lactate dehydrogenase; Albumin; Creatinine; Risk factors
急性胰腺炎(AP)是指多种病因引起的胰酶激活，继以胰腺局部炎症反应为主要特征，伴或不伴有其他器官功能改变的疾病。20%～30%患者临床经过凶险。总体病死率为5%～10%[1]。胰酶入血引起远处脏器和全身酶系统损伤，产生大量炎症介质和细胞因子，可引起全身炎症反应综合征和多器官功能障碍综合征(MODS)[2-3]。AP患者一旦出现MODS，病死率明显升高，预后极差。因此，尽早筛选出高危患者，对提高诊疗效果降低病死率有重要意义。多因素Logisitic回归分析表明AP发病早期(48 h内)出现血清乳酸脱氢酶>400 U/L (P=0.025,OR=7.469)、肌酐肌酐>120 μmol/L (P=0.001, OR=21.042) 、白蛋白<30 g/L(P=0.018, OR=9.221)为AP导致MODS的早期独立危险因素[4]。本文旨在建立预测模型，为早期筛选出MODS高危患者提供帮助。, 百拇医药(陈都 陆士奇 徐峰)
【关键词】急性胰腺炎；多器官功能障碍综合征；Logistic回归模型；ROC曲线；乳酸脱氢酶；白蛋白；肌酐；危险因素
, 百拇医药
Establishment of regression model of acute pancreatitis complicated with secondary multiple organs dysfunction syndrome for prediction and evaluating its efficacy CHEN Du, LU Shi-qi, XU Feng. Department of Emergency, the First Affiliated Hospital of Soochow University，Suzhou 215006，China
Corresponding author: LU Shi-qi, Email: lushiqi2004@126.com
【Abstract】Objective To study the way for establishing prediction model of acute pancreatitis with secondary multiple organs dysfunction syndrome (MODS) in the early stage. Methods The linical data were collected from 113 hospitalized patients with acute pancreatitis admitted from October 2008 to October 2011. Some biomarkers within 48 h after the onset of acute pancreatitis were statistically analyzed by univariate and multivariate binary logistic regression analysis in order to establish the prediction model evaluated by the ROC curve. Results There were statistically significant differences in lactate dehydrogenase (LDH), serum creatinine (CR), serum albumin (P<0.05) identified by univariate regression analysis. These independent risk factors in the regression model were: ln(P/1-P) = 2.243+0.002×LDH+0.014×CR-0.186×ABL, likelihood ratio test(χ2=66.46，P=0.000); in the ROC curve analysis：AUC 0.927(95%CI：0.875-0.980)，the best cut-off value of the model predicting probability was 0.177 (sensitivity 82.14%，specificity 85.88%，correctly classified 84.96%). Conclusions Efficient prediction model could be set up by the logistic regression analyzing the early risk factors in multiple organ dysfunction syndrome following acute pancreatitis.
, 百拇医药
【Key words】Acute pancreatitis; Multiple organ dysfunction syndrome; Logistic regression model; ROC curve; Lactate dehydrogenase; Albumin; Creatinine; Risk factors
急性胰腺炎(AP)是指多种病因引起的胰酶激活，继以胰腺局部炎症反应为主要特征，伴或不伴有其他器官功能改变的疾病。20%～30%患者临床经过凶险。总体病死率为5%～10%[1]。胰酶入血引起远处脏器和全身酶系统损伤，产生大量炎症介质和细胞因子，可引起全身炎症反应综合征和多器官功能障碍综合征(MODS)[2-3]。AP患者一旦出现MODS，病死率明显升高，预后极差。因此，尽早筛选出高危患者，对提高诊疗效果降低病死率有重要意义。多因素Logisitic回归分析表明AP发病早期(48 h内)出现血清乳酸脱氢酶>400 U/L (P=0.025,OR=7.469)、肌酐肌酐>120 μmol/L (P=0.001, OR=21.042) 、白蛋白<30 g/L(P=0.018, OR=9.221)为AP导致MODS的早期独立危险因素[4]。本文旨在建立预测模型，为早期筛选出MODS高危患者提供帮助。, 百拇医药(陈都 陆士奇 徐峰)