四种终末期肝病模型联合血清钠模型对慢加急性肝衰竭患者短期预后的评估价值

Short-term prognostic value of four types MELD combined with serum sodium in patients with acute-on-chronic liver failure

目的探讨终末期肝病模型(MELD)联合血清钠MELD-Na、MESO、iMELD模型对评估慢加急性肝衰竭患者预后的价值。方法对53例慢加急性肝衰竭患者进行回顾性分析,随访患者4、12周预后,分别应用MELD、MELD-Na、MESO及iMELD模型进行评分,用受试者特征曲线(ROC)下面积(AUC)比较各评分系统预测患者生存不同时间的准确性。计量资料应用t检验;计数资料用χ2检验;AUC比较采用正态性Z检验。结果 53例慢加急性肝衰竭患者4、12周内分别死亡31、1例。血清钠≤125 mmol/L,死亡4例;125~135 mmol/L,死亡16例;血清钠≥135 mmol/L,死亡13例;3组之间差异有统计学意义(P<0.05)。MELD评分≤25,死亡8例;MELD评分25~30,死亡12例;MELD评分≥30,死亡12例,3组之间差异有统计学意义(P<0.05)。判断患者4周的预后,MELD、MELD-Na、MESO、iMELD的AUC值分别为0.722、0.746、0.739、0.633。MELD的最佳截断值:24.5,敏感度83.3%,特异度36.4%;MELD-Na的最佳截断值:26,敏感度83.3%,特异度45.5%;MESO的最佳截断值:1.8,敏感度86.7%,特异度40.9%;iMELD的最佳截断值:42,敏感度83.3%,特异度63.6%。MELD与MELD-Na、MESO、iMELD对慢加急性肝衰竭患者不同生存时间的AUC比较差异无统计学意义(P>0.05)。结论 MELD及其联合血清钠模型均可有效地预测评估慢加急性肝衰竭患者短期预后。

Objective To investigate prognostic value of models for end‐stage liver disease （MELD） combined with serum sodium （MELD‐Na ,MESO and iMELD） in patients with acute‐on‐chronic liver failure .Methods Fifty‐three patients with acute‐on‐chronic liver failure were retrospectively analyzed .After 4 and 12 weeks follow‐up ,MELD ,MELD‐Na ,MESO and iMELD from patients were respectively performed .The area under the receiver operating characteristic （AUROC） curves was carried our to compare the accuracy of all scoring systems in prediction of different survival times of patients .Measurement data was tested by t‐test , enumeration data was tested by chi‐square test , and AUROC was compared by normal Z‐test .Results Among 53 patients with acute‐on‐chronic liver failure ,31 died within 4 weeks ,and 1 died within 12 weeks .According to the level of serum sodium ,all patients were divided into three groups ,including less than 125mmol/L ,125mmol/L^135mmol/L and more than 135mmol/L ,in which mortality was 100% ,66 .7% and 52%respectively with statistically significant differences （P〈0 .05） .According to MELD scores ,patients were also divided into three group ,including less than 25 ,25 ~ 30 and more than 30 ,in which the mortality was 34 .8% ,75% and 92 .3%respectively with statistically significant differences （P〈0 .05） .According to prognosis after 4 and 12 weeks follow up ,the AUC was judged to be more than 0 .6 after 4 and 12 weeks in four models ,and the AUC was 0 .722 ,0 .746 0 .739 and 0 .633 for MELD ,MELD‐Na ,MESO and iMELD after 4 weeks ,respectively .The optimum cut‐off value ,susceptibility and specificity were 24 .5 ,83 .3% and 36 .4% for MELD respectively ,26 ,83 .3% and 45 .5% for MELD‐Na respectively ,1 .8 , 86 .7% and 40 .9% for MESO respectively ,and 42 ,83 .3% and 63 .6% for iMELD respectively .Comparison of AUROC for different survival time of patients with acute‐on‐chronic liver failure showed that MELD was not statistically different from MELD‐Na ,MESO and iMELD （P〉0 .05） .Conclusion MELD and its combinations with serum sodium are effective in predicting short‐term prognosis of patients with acute‐on‐chronic liver failure .