中文摘要：

目的：建立影响核苷（酸）类似物治疗代偿性乙肝肝硬化疗效的预测模型。方法：204例代偿性乙肝肝硬化患者给予核苷（酸）类似物治疗，治疗48周后随机选取136例作为建模组，68例作为验证组，根据治疗效果将建模组又分为代偿组87例，失代偿组49例。对影响代偿组与失代偿组预后的相关因素进行单因素及Cox回归分析，建立预测方程。结果：多因素Cox比例风险回归分析得出实际较优模型：h（t、x）=h0（t）exp（0．5502x16＋0．3247x19-0．0149x8-0．0130X11-0．0125x14），由模型可知ALT（x8）、胆红素（X11）、PT（X14）、透明质酸（X16）、肝硬度值（X19）对核苷（酸）类似物治疗代偿性乙肝肝硬化的预后影响较大；代偿性乙肝肝硬化发展成为失代偿性乙肝肝硬化的概率模型（P）=1／[1＋e—h（t、x）]，受试者工作特征（ROC）曲线下面积（AUC）为0．8519，回归模型预测能力良好。验证组中失代偿性乙肝肝硬化组ALT、胆红素、PT阳性，透明质酸〉200μg／L，肝硬度值〉25kPa比例均高于代偿性乙肝肝硬化组，差异比较有统计学意义（P〈0．05）。结论：核苷（酸）类似物可有效抑制代偿性乙肝肝硬化患者HBV-DNA病毒复制，并促进HBeAg转阴，但其治疗效果受ALT、胆红素、PT、透明质酸、肝硬度值等的影响。

英文摘要：

Objective： To establish predictive models on nucleoside （acid） analogue treatment efficacy decompensated liver cirrhosis. Methods： 204 cases of hepatitis B patients with decompensated cirrhosis were given nucleoside （acid） analogue treatment. After 48 weeks of treatment, 136 cases were randomly selected as a modeling group, and the other 68 cases as a verification group. According to the therapeutic effects, patients in modeling group were divided into compensated group （87 cases） and decompensated group （49 cases）. The factors affecting prognosis of compensation group and decompensated group were analyzed by univafiate and Cox regression to get prediction equation. Results： Multivariate Cox proportional hazards regression analysis showed the actual optimum model as h（t, x） =h0 （t）exp （0.5502x16＋0.3247x19-0.0149x8-0.0130x11-0.0125x14）. The model showed that ALT （XS）, bilirubin （X11）, PT （X14）, hyaluronic acid （X16）, liver stiffness values （X19） had greater prognostic impact on nucleoside （acid） analogue treated compensation hepatitis B cirrhosis. Compensated cirrhosis developed into decompensated cirrhosis probability model （P） = 1/ [1＋ e-h （t, x）]. The receiver operating characteristic area under the curve （ROC） （AUC） was 0.8519, and the regression model showed good predictive ability. In the decompensated cirrhosis group, the positive rate of ALT, bilirubin and PT, the proportion of hyaluronic〉 200 μg / L and liver stiffness values〉 25 kPa were higher than in the compensated liver cirrhosis group, and the differences were statistically significant （P 〈0.05）. Conclusion： Nucleoside （acid） analogues can effectively inhibit HBV-DNA viral replication in compensated cirrhosis patients, and promote HBeAg negative, but its therapeutic effect was impacted by ALT, bilirubin, PT, hyaluronic acid and liver stiffness value.