摘要
目的基于多因素Cox回归分析探讨行非生物型人工肝治疗肝衰竭病人的预后影响因素并构建列线图预后模型,以指导临床决策。方法回顾性收集2014年1月至2022年12月于山西医科大学第一医院感染科收治的275例行非生物型人工肝治疗肝衰竭病人的临床资料,依据病人3个月的病情随访结局分为生存组(177例)和不良结局组(98例),对观察指标进行单因素与多因素Cox回归分析,基于特征变量绘制列线图预测模型并评价模型。利用X-tile 3.6.1软件将列线图风险得分进行分组进一步探讨其临床价值。结果生存组的年龄、血清总胆红素[(364.98±151.22)μmol/L比(434.69±144.45)μmol/L]、凝血酶原时间[25.00(20.80,31.45)s比29.58(23.95,44.33)s]、国际标准化比值[2.17(1.91,2.82)比2.85(2.17,3.91)]低于不良结局组(P<0.05);凝血酶原活动度[33.12(24.10,39.20)%比25.73(19.40,33.78)%]、动态MELD评分(终末期肝病预后模型)高于不良结局组(P<0.05);生存组的病因[肝炎病毒58(32.8)%,药物性48(27.1)%,酒精性33(18.6)%及其他38(21.5)%]与不良结局组[40(40.8)%,12(12.2)%,11(11.2)%及35(35.7)%]相比差异有统计学意义(P<0.05)。多因素Cox回归分析显示,病因-药物性及酒精性、年龄、血清总胆红素、凝血酶原时间、凝血酶原时间活动度、国际标准化比值及动态MELD评分为非生物型人工肝治疗肝衰竭病人预后的独立影响因素,基于上述指标构建列线图模型;该模型的一致性指数及AUC分别为0.75、0.78,校正曲线及决策分析曲线提示该模型预测性能良好,低风险病人的生存率明显优于中、高风险组(P<0.001)。结论基于病因-药物性及酒精性、年龄、血清总胆红素、凝血酶原时间、凝血酶原时间活动度、国际标准化比值及动态MELD评分构建的列线图预后模型对于行非生物型人工肝治疗肝衰竭病人的临床决策有较高指导意义。
Objective To investigate the prognostic factors and construct a nomogram prognostic model based on multifactorial Cox regression analysis in liver failure patients treated with non-bioartificial liver support system to guide clinical decision-making.Methods Clinical data of 275 liver failure patients treated with non-bioartificial liver support system in the Department of Infection,First Hospital of Shanxi Medical University from January 2014 to December 2022 were retrospectively collected,and patients were assigned into survival group(177 cases)and poor prognosis group(98 cases)based on their 3-month follow-up outcomes.The observed indicators were analyzed sequentially by univariate and multifactor Cox regression analysis,and the nomogram model was constructed based on the characteristic variables,and then the model was evaluated.Next,the nomogram risk scores were grouped using X-tile 3.6.1 software to further explore their clinical value.Results Age,serum total bilirubin[(364.98±151.22)mmol/L vs.(434.69±144.45)mmol/L],plasminogen time[25.00(20.80,31.45)s vs.29.58(23.95,44.33)s],and international normalized ratio[2.17(1.91,2.82)vs.2.85(2.17,3.91)]in survival group were lower than those in the poor prognosis group(P<0.05);and plasminogen activity[33.12(24.10,39.20)%vs.25.73(19.40,33.78)%],and ambulatory MELD score in survival group were higher than those in the poor outcome group(P<0.05);the etiology of the disease in the survivor group[hepatitis viruses 58(32.8)%,pharmacologic 48(27.1)%,alcoholic 33(18.6)%,and other 38(21.5)%]was higher than those in the poor prognosis group[40(40.8)%,12(12.2)%,11(11.2)%and 35(35.7)%],and the difference was statistically significant(P<0.05).Multi-factor Cox regression analysis showed that etiology-drug and alcohol,age,total serum bilirubin,prothrombin time,prothrombin time activity,international normalized ratio and dynamic MELD score were independent factors affecting the prognosis of patients with liver failure treated with abiotic artificial liver,and a nomogram model was constructed based on these indicators.The consistency index and AUC of the model were 0.75 and 0.78,respectively,and the calibration curve and decision analysis curve suggested that the model had good predictive performance,and in addition,the survival rate of low-risk patients was significantly better than that of the medium-and high-risk groups(P<0.001).Conclusion The nomogram prognostic model based on etiology-drug and alcohol,age,total serum bilirubin,prothrombin time,prothrombin time activity,international normalized ratio and dynamic MELD score has high clinical guideline significance in liver failure patients treated with non-bioartificial liver support system.
作者
史诗
王一帆
杨艳芬
刘元丽
余红梅
邓春青
SHI Shi;WANG Yifan;YANG Yanfen;LIU Yuanli;YU Hongmei;DENG Chunqing(School of Public Health,Shanxi Medical University,Taiyuan,Shanxi 030001,China;Department of Infection,The First Hospital of Shanxi Medical University,Taiyuan,Shanxi 030001,China)
出处
《安徽医药》
2025年第4期798-804,共7页
Anhui Medical and Pharmaceutical Journal
基金
北京肝胆相照基金会“2022年度人工肝专项基金”项目(iGandanF-1082022-RGG005)
关键词
非生物型人工肝
肝衰竭
多因素Cox回归
列线图模型
预后
Non-bioartificial liver support system
Liver failure
Multifactorial Cox regression
Nomogram model
Prognosis
