Objective To compare the value of the scoring systems, including the scoring system for end-stage liver disease (MELD), MELD-Na, MELD to Sna ratio (MESO) index and integrated MELD (iMELD) in the prediction of the prognosis of patients with chronic severe hepatitis B. Methods Clinical data of 122 inpatient with chronic severe hepatitis B were recorded. The patients were divided into survivor group and non-survivor group according to the 3 months of follow-up. All the patients were evaluated according to MELD, MELD-Na, MESO and iMELD scoring systems and the differences of each scoring system in the two groups were compared. The receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC) were used to evaluate the predictive ability of each system. The total predictive agreement rate and Youden index were calculated according to the optimum cut-off values which were determined by the respective ROC curve. Results The average scores of MELD, MELD-Na, MESO and iMELD in the non-survivor group were 27.6±7.0, 34.4±10.7, 2.1±0.6 and 35.8±9.1, respectively, which were greater than those in the survivor group (18.1±4.9, 20.4±7.8, l.5±0.3, 25.0±6.2), and the differences between them were significant; AUC of each scoring system were 0.839, 0.860, 0.847 and 0.832, respectively. There were no significant differences in the predictive ability and total predictive agreement rates, which were 75.41%, 81.15%, 78.69% and 76.23%, respectively. Youden index of the 4 systems were 0537, 0.629, 0.598 and 0.541, respectively. Conclusions MELD, MELD-Na, MESO and iMELD scoring models have better effect on predicting short-term prognosis of the patients with chronic severe hepatitis B. The 4 scoring models have the similar predictive values, and the differences are not significant.%目的 比较终末期肝病模型( model for end-stage liver disease,MELD)评分、MELD联合血清钠模型(MELD-Na)评分、MESO评分/血清钠比值(MELD to SNa ratio,MESO)和integrated MELD(iMELD)评分系统预测慢性乙型重型肝炎患者预后的价值.方法 记录住院后确诊慢性乙型重型肝炎122例患者的临床资料,并随访3个月的生存情况,分为生存组和死亡组,并采用MELD、MELD-Na、MESO和iMELD评分系统,比较各评分系统在两组间的差异,通过受试者工作特征曲线(receiver operating characteristic curve,ROC曲线)和曲线下面积(area under the ROC curve,AUC)评价每一个模型的预测能力,同时用各自ROC曲线确定判断重型肝炎患者生存与否的评分模型的最佳临界值,并算出总预测正确率和Youden指数.结果 死亡组的MELD、MELD-Na、MESO、iMELD评分分别是(27.6±7.0)、(34.4±1 0.7)、(2.1±0.6)和(35.8±9.1)分,均高于生存组的(18.1±4.9)、(2 0.4±7.8)、( 1.5±0.3)和(25.0±6.2)分,差异有统计学意义；各评分系统的AUC分别是0.839、0.860、0.847和0.832,差异无统计学意义；总正确率分别是75.41%、81.15%、78.69%和76.23%,差异亦无统计学意义.Youden指数分别为0.537、0.629、0.598和0.541.结论 MELD、MELD-Na、MESO和iMELD评分模型对慢性乙型重型肝炎患者短期预后判断有较好的作用,4种模型预测价值无统计学差异,预测价值相当.
展开▼
