白蛋白-胆红素指数在急性胰腺炎并发多器官功能障碍综合征Nomogram风险预测中的作用

The role of albumin-bilirubin index in Nomogram prediction of acute pancreatitis complicated with multiple organ dysfunction syndrome

目的:探索白蛋白-胆红素指数(ALBI)在急性胰腺炎(AP)并发多器官功能障碍综合征(MODS)Nomogram风险预测中的作用。方法:回顾性分析华中科技大学同济医学院附属同济医院重症医学科收治的229例AP患者的临床资料,根据是否并发MODS分成MODS组和非MODS组。通过多因素logistic回归分析确定AP并发MODS的高危预测因素并通过R语言建立Nomogram图预测模型。使用受试者工作特征曲线(ROC)评价Nomogram图的预测能力,同时,使用Bootstrap自抽样内部验证法和一致性指数(C-index)验证Nomogram图风险预测效果的准确性。结果:多因素logistic回归分析结果提示ALBI指数、急性生理与慢性健康状况评分Ⅱ(APACHEⅡ)和氧合指数(PF)是AP并发MODS的高危因素。其中,ALBI指数的AUC值(0.967)、灵敏度(91.72%)和特异度(98.33%)均显著高于APACHEⅡ评分和PF。基于ALBI指数(回归系数=10.080)构建的Nomogram图风险预测模型预测AP并发MODS的C-index及AUC值均为0.997(95%CI:0.9924~1.000,P<0.05)。模型校准曲线显示预测概率与实际发生概率趋于一致。结论:以ALBI指数为核心,联合APACHEⅡ评分和PF构建的Nomogram图风险预测模型可以有效预测AP并发MODS。

Objective:To explore the role of albumin-bilirubin index(ALBI)in Nomogram prediction of acute pancreatitis(AP)complicated with multiple organ dysfunction syndrome(MODS)Methods:Clinical data of 229 patients with AP admitted to Tongji Hospital in Wuhan were retrospectively collected.The patients were divided into MODS groups and non-MODS groups based on whether MODS were concurrent or not.Multivariate logistic regression analysis was used to determine the high-risk predictors of AP complicated with MODS,and a Nomogram was drawn combined with R language.Receiver operating characteristic curve(ROC)was used to evaluate the prediction ability of the Nomogram.Meanwhile,Bootstrap self-sampling internal validation method and consistency index(C-index)were used to verify the accuracy of the prediction effect of the Nomogram.Results:Multivariate logistic regression analysis indicated that ALBI,acute physiology and chronic health status scoreⅡ(APACHEⅡ)and oxygenation index(PF)ALBI were risk factors for AP complicated with MODS.Among them,the AUC value(0.967),sensitivity(91.72%)and specificity(98.33%)of ALBI index were significantly higher than those of the APACHEⅡscore and PF.The C index and AUC value of Nomogram prediction model based on ALBI index(regression coefficient=10.080)was 0.997 in predicting AP complicated with MODS(95%CI:0.9924~1.000,P<0.05).The calibration curve showed a good fit between the predicted probability and actual probability.Conclusion:The Nomogram prediction model based on ALBI index,accompanied with APACHEⅡscore and PF can effectively predict AP complicated with MODS.