摘要目的:分析终末期肝病模型(model for end-stage liver disease，MELD)-肌少症评分在慢加急性肝衰竭(acute-on-chronic liver failure，ACLF)患者短期预后评价中的预测能力。方法:选取2013年1月至2019年12月于天津市第三中心医院住院的ACLF患者271例，其中合并肌少症患者157例，非肌少症患者114例，根据ACLF分型分为A组(无肝硬化基础)61例，B组(有代偿期肝硬化基础)99例，C组(既往有失代偿期肝硬化病史)111例。回顾性收集患者的基本资料、实验室检查结果、计算机断层成像检查结果和预后，并计算MELD评分、MELD-Na评分、MELD-肌少症评分。统计学分析采用多因素logistic回归、多因素Cox比例风险回归分析、Kaplan-Meier法、log-rank法和受试者操作特征曲线下面积。结果:低体重指数[比值比(odds ratio， OR)＝0.93， P<0.001]、合并肝硬化( OR＝1.14， P＝0.004)、合并肝性脑病( OR＝1.31， P<0.001)、高白细胞计数( OR＝1.18， P＝0.009)和高血小板计数( OR＝1.08， P<0.001)是ACLF患者合并肌少症的独立危险因素。高MELD评分[风险比(hazard ratio， HR)＝1.02， P＝0.001]、高MELD-Na评分( HR＝1.07， P＝0.038)、高MELD-肌少症评分( HR＝1.14， P<0.001)、高总胆红素水平( HR＝1.00， P<0.001)和高国际标准化比值水平( HR＝1.71， P<0.001)是ACLF患者死亡的独立危险因素。亚组分析中，A、B组患者中，肌少症患者累积生存率均低于非肌少症患者( χ2＝5.97、8.34， P＝0.015、0.004)，C组肌少症患者与非肌少症患者累积生存率差异无统计学意义( χ2＝4.90， P＝0.053)。在A组+B组患者中，MELD-肌少症评分预测短期预后的曲线下面积为0.87，大于MELD评分(0.78)和MELD-Na评分(0.78)，差异均有统计学意义( Z＝2.86、2.56， P＝0.004、0.011)，C组患者MELD-Na评分预测短期预后的曲线下面积(0.83)大于MELD评分(0.71)和MELD-肌少症评分(0.69)，差异均有统计学意义( Z＝2.52、2.64， P＝0.012、0.008)。 结论:与不合并肌少症的患者相比，合并肌少症的无肝硬化基础或有代偿期肝硬化基础的ACLF患者生存时间更短、预后更差；对于无肝硬化基础或有代偿期肝硬化基础的ACLF患者，MELD-肌少症评分能够较好地预测其短期预后。

abstractsObjective:To analyze the predictive ability of model for end-stage liver disease (MELD)-sarcopenia score in short-term prognosis of patients with acute-on-chronic liver failure (ACLF).Methods:Two hundred and seventy-one patients with ACLF hospitalized in Tianjin Third Central Hospital from January 2013 to December 2019 were selected, among whom 157 cases with sarcopenia and 114 cases without sarcopenia.According to ACLF classification, the patients were divided into group A (no cirrhosis basis) of 61 cases, group B (compensated cirrhosis basis) of 99 cases, and group C (previous history of uncompensated cirrhosis) of 111 cases.The basic data, laboratory examination results, computed tomography (CT) examination results and prognosis of the patients were retrospectively collected, and the MELD score, MELD-Na score and MELD-sarcopenia score were calculated. Multivariate logistic regression, multivariate Cox proportional hazards regression, Kaplan-Meier method, log-rank method and area under receiver operating characteristic curve were used for statistical analysis.Results:Low body mass index (odds ratio ( OR)=0.93, P<0.001), complicated cirrhosis ( OR=1.14, P=0.004), complicated hepatic encephalopathy ( OR=1.31, P<0.001), high white blood cell level ( OR=1.18, P=0.009) and high platelet level ( OR=1.08, P<0.001) were independent risk factors for sarcopenia in patients with ACLF. High MELD score (hazard ratio ( HR)=1.02, P=0.001), high MELD-Na score ( HR=1.07, P=0.038), high MELD-sarcopenia score ( HR=1.14, P<0.001), high total bilirubin ( HR=1.00, P<0.001) and high international normalized ratio (INR) ( HR=1.71, P<0.001) were independent risk factors for death in patients with ACLF. In subgroup analysis, the cumulative survival rate of sarcopenia patients in group A and B was lower than that of non-sarcopenia patients ( χ2=5.97 and 8.34, respectively, P=0.015 and 0.004, respectively), while there was no significant difference in the cumulative survival rate between sarcopenia patients and non-sarcopenia patients in group C ( χ2=4.90, P=0.053). In groups A and B, the area under the curve (AUC) of MELD-sarcopenia score in predicting short-term prognosis was 0.87, which was higher than MELD score (0.78) and MELD-Na score (0.78), and the differences were both statistically significant ( Z=2.86 and 2.56, respectively, P=0.004 and 0.011, respectively). The AUC of MELD-Na score in predicting short-term prognosis in group C (0.83) was higher than that of MELD score (0.71) and MELD-sarcopenia score (0.69), and the differences were both statistically significant ( Z=2.52 and 2.64, respectively, P=0.012 and 0.008, respectively). Conclusions:Patients with ACLF with no cirrhosis basis or compensated cirrhosis basis complicated with sarcopenia have shorter survival time and worse prognosis than those without sarcopenia. For patients with ACLF with no cirrhosis basis or compensated cirrhosis basis, MELD-sarcopenia score has better predictive value for the short-term prognosis.