终末期肝病模型评分及血清钠在肝衰竭近期预后判断中的意义

目的 比较基线及动态终末期肝病模型（MELD）及其联合血清钠（MELD-Na）在评价肝衰竭近期预后中的价值.方法 回顾性分析2003年4月至2012年4月新疆医科大学第一附属医院322例肝衰竭住院患者的资料,计算患者确诊时和一周后的MELD、MELD-Na评分,并计算△MELD 、△MELD-Na分值.应用受试者工作特征（ROC）曲线评价各评分系统预测患者3个月预后的价值,同时绘制Kaplan-Meier（K-M）生存曲线.结果 急性和亚急性、慢加急性、慢性肝衰竭患者3个月时的病死率分别为77.4％（24/31）,41.7％（50/120）和56.1％（96/171）,比较差异有统计学意义（x2=14.273,P ＜0.01）.对于急性和亚急性肝衰竭患者,各评分系统预测患者近期预后的ROC曲线下面积（AUC）为0.699～0.836,差异无统计学意义（Z=0.507,0.622,0.712,0.727,0.779,0.599,P＞0.05）.对于慢加急性肝衰竭,△MELD和△MELD-Na的AUC分别为0.889和0.897,均高于基线MELD和MELD-Na的AUC（Z=3.110和3.500,P＜0.05）,但△MELD和△MELD-Na的AUC比较差异无统计学意义（Z =0.310,P ＞0.05）;K-M生存曲线显示,当△MELD＞ 3.5分时,患者3个月内病死率为87.8％,平均生存时间为34.05 d.对于慢性肝衰竭,△MELD预测患者近期预后的AUC为0.871,优于△MELD-Na（Z=4.229,P＜0.05）;K-M生存曲线显示,当△MELD＞ 4.5分时,患者3个月内病死率为89.9％,平均生存时间为29.08 d.结论 对于急性和亚急性肝衰竭,各评分预测效果均可;对于慢加急性肝衰竭,△MELD和△MELD-Na均优于相应的基线评分系统;对于慢性肝衰竭,△MELD的预测能力最佳.

Model for end-stage liver disease score and serum natrium level in predicting short-term prognosis of liver failure

Objective To evaluate the model for end-stage liver disease （MELD） and MELD combined with serum natrium level （MELD-Na） in predicting short-term prognosis of liver failure.Methods Clinical data of 322 patients with liver failure admitted in the First Affiliated Hospital of Xinjiang Medical University from April 2003 to April 2012 were retrospectively analyzed.MELD and MELD-Na scores were calculated at diagnosis and one week after the diagnosis,and then △MELD and △MELD-Na were determined.Receiver operating characteristics （ROC） curve and Kaplan-Meier survival curve were used to evaluate the value of the above scores in predicting 3-month prognosis.Results The 3-month mortality rates of acute/sub-acute,acute-on-chronic and chronic liver failure were 77.4% （24/31）,41.7% （50/120） and 56.1% （96/171）,respectively,and the difference was of statistical significance （x2 =14.273,P 〈0.01）.For acute/sub-acute liver failure,the areas under ROC curve （AUCs） were 0.699-0.836 for each scoring system in predicting short-term prognosis,and no significant difference was observed （Z =0.507,0.622,0.712,0.727,0.779 and 0.599,P 〉0.05）.For acute-on-chronic liver failure,AUCs were 0.889 and 0.897 for △MELD and △MELD-Na in predicting short-term prognosis,which were higher than those of MELD and MELD-Na scores at the baseline （Z =3.110 and 3.500,P 〈 0.05）,but no significant difference was observed between △MELD and △MELD-Na （Z =0.310,P 〉 0.05） ; Kaplan-Meier survival curve showed that the 3-month mortality rate for patients with △MELD 〉 3.5 was 87.8%,and the average survival time was 34.05 d.For chronic liver failure,AUC of △MELD was 0.871 in predicting short-term prognosis,which was higher than that of △MELD-Na （Z =4.229,P 〈0.05） ; Kaplan-Meier survival curve showed that the 3-month mortality rate for patients with △MELD 〉 4.5 was 89.9%,and the average survival time was 29.08 d.Conclusion For acute/sub-acute liver failure,MELD,MELD-Na,△MELD and △MELD-Na are all satisfactory in predicting short-term prognosis; for acute-on-chronic liver failure,△MELD and △MELD-Na are better than MELD,MELD-Na scores at the baseline; and for chronic liver failure,△MELD is the best indicator.