乙型肝炎相关肝功能衰竭患者预后危险因素及预后模型建立

摘要

目的 探讨影响乙型肝炎相关肝衰竭患者预后的危险因素并建立其预后模型.方法 回顾性收集2006年6月至2008年12月在我科收治的178例乙型肝炎相关肝衰竭患者的临床资料.采用x 2和t检验和非参数检验进行单因素分析,Logistic回归进行多因素分析.结果 年龄、腹水、感染、消化道出血、肝性脑病、肝肾综合征、甲胎蛋白、PT、WBC、TBil、DBil、Cr、BUN、血清钠在生存组与死亡组之间差异均有统计学意义(P＜0.05).Logistic多因素分析进一步显示,肝性脑病、感染、PT、TBil是影响乙型肝炎相关肝衰竭患者预后的独立危险因素.同时对所得出的独立危险因素建立该人群的预后判断模型,通过计算预后指数并绘制ROC曲线,计算其曲线下面积(AUC)为0.931(95％CI:0.893～0.970).其评估价值优于CTP分级(0.862)、MELD评分(0.807)及MELD-Na评分(0.774).结论 肝性脑病、感染、PT、TBil是影响乙型肝炎相关肝衰竭患者预后的独立危险因素.建立的预后模型能够较为准确地预测乙型肝炎相关肝衰竭患者的短期预后,是一个较为理想的乙型肝炎相关肝衰竭预后评估系统.%Objective To investigate the risk factors of the prognosis for HBV related liver failure and thus to establish a prognosis model. Methods Retrospective analysis of the clinical data of 178 patients with HBV related liver failure in Ruijin hospital from June 2006 to December 2008. Quantitative data were analyzed by using t test and rank test, and qualitative data were analyzed by using Chi-square test. Then Logistic regression analysis was used for selecting the independent risk factors of the prognosis for HBV related liver failure. Based on independent risk factors from Logistic regression analysis, prognostic model for the patients with HBV related liver failure was established. Results The differences of age, ascites, infection, upper gastrointestinal bleeding, hepatic encephalopathy, hepatorenel syndrome, alpha fetoprotein (AFP), prothrombin time (PT), white blood cell count (WBC), total bilirubin (Tbil), direct bilirubin (Dbil), serum creatinine (Cr), blood urea nitrogen (BUN) and serum sodium (Na) were statistically significant in the patients with HBV related liver failure by univariate analysis (P＜0. 05). Furthermore, hepatic encephalopathy, infection, PT and Tbil were the independent risk factors with HBV related liver failure by Logistic regression analysis. Area under the ROC curve (AL7C) for the prognostic model was 0. 932(.95%Cl, 0. 893-0. 970). It is better than CTP, MELD and MELD-Na. Conclusion Hepatic encephalopathy, infection, PT and Tbil are the independent risk factors for the prognosis of HBV related liver failure. The Logistic regression model established in this study has good predictability. It is of significant value in the prognosis of HBV related liver failure.