摘要
目的分析影响HBV相关肝功能衰竭预后的危险因素,构建预后评估模型,评价其预测能力。方法采用回顾性队列研究方法,选择2007年1月至2010年12月住院确诊为HBV相关肝功能衰竭的569例患者。应用Kaplaw-Meier法进行生存分析,COX比例风险回归模型作单因素及多因素回归分析,对年龄、性别、并发症、生物化学指标、凝血指标、HBVDNA载量等进行单因素和多因素分析。构建预后评估模型,并用独立的79例验证样本评价模型预测能力,受试者工作特征(ROC)曲线评估本模型与终末期肝病模型对肝功能衰竭生存率预测的准确性。结果本组肝功能衰竭患者中位生存时间为59d,1、3、6个月的生存率分别为58.9%、46.2%和45.5%,1、3年生存率分别为44.9%、44.5%。肝性脑病、肺部感染、上消化道出血、Alb、AST、Cr、国际标准化比值(INR)是影响预后的独立危险因素(均P〈0.01),所构建的预后评估模型的预后指数(PI)=4.98×肝性脑病赋值+4.57×肺部感染赋值+4.41×上消化道出血赋值9.69×ln[Alb(g/L)]+2.46×ln[AST(g/L)]+5.18×ln[Cr(mmol/L)]+3.35×ln(INR)-15.36,判断肝功能衰竭患者90d内预后的ROC曲线下面积为0.838,而终末期肝病模型的ROC曲线下面积为0.751,两者差异无统计学意义(Z=1.085,P=0.278)。结论肝性脑病、肺部感染、上消化道出血、Alb、AST、Cr、INR作为独立危险因素所构建的预后评估模型,能够较准确地预测HBV相关肝功能衰竭患者90d的生存情况。
Objective The aim of this study was to analyze risk factors which may affect prognosis of patients with hepatits B virus (HBV) related liver failure, and to construct a model for prognostic evaluation and further assess its predictive ability. Methods In this retrospective cohort study, 569 hospitalized patients who were diagnosed with HBV-related liver failure from January 2007 to December 2010 were enrolled. All the patients were followed up and survival analysis was performed using the Kaplan-Meier method. Univariate and multivariate COX proportional hazards regression analyses were applied to variables such as age, sex, complications, biochemical markers, coagulation markers, and HBV DNA levels to construct a model for prognostic evaluation, and 79 independent cases of HBV-related liver failure were used to confirm the model's predictive ability.Accuracy of the constructed model and model for end stage liver disease (MELD) was evaluated by receiver operating characteristic (ROC) curves. Results The median survival time for all the patients was 59 days. The survival rates at 1, 3, 6 months were 58.9%, 46. 2% and 45. 5%, respectively; and survival rates at 1 and 3 years were 44.9% and 44. 5%, respectively. Hepatic encephalopathy, pulmonary infection, upper gastrointestinal bleeding ( UGIB ), albumin ( Alb ), aspartate aminotransferase (AST), creatininc (Cr), international normalized ratio (INR) were determined to be independent risk factors (all P〈0.01) which may affect survival o{ patients with HBV-related liver failure. Accordingly, the prognostic index (PI) of the constructed model for prognostic evaluation= 4.98 × assignment of hepatic encephalopathy+ 4. 57 × assignment of pulmonary infection+ 4. 41 × assignment of UGIB 9.69XlmEAlb (g/L)+2.46 ×In[AST (U/L)+5. 18×lnECr (mmol/L)]+ 3.35 × In (INR) 15.36. The area under receiver operating characteristic curve was 0. 838 for the constructed model assessing 90 d survival of the patients, and was 0,751 for model for end stage liver disease, with no significant difference between the two models (Z=1. 085, P= 0. 278). Conclusions Prognosis of patients with HBV related liver failure can be accurately predicted by the constructed prognostic assessment model, which is consisted of hepatic encepbalopathy, pulmonary infection, UGIB, Alb, AST, Cr, and INR as independent risk factors, and is able to predict the 90-d survival.
出处
《中华传染病杂志》
CAS
CSCD
北大核心
2013年第6期347-352,共6页
Chinese Journal of Infectious Diseases
关键词
肝炎病毒
乙型
肝功能衰竭
预后
比例危险度模型
回归分析
Hepatitis B virus
Liver failure
Prognosis
Proportional hazards models
Regression analysis