论文部分内容阅读
目的探讨终末期肝病模型(MLED)评分系统对慢性重型肝炎患者短期(3个月)预后的预测能力及其临床应用价值。方法应用MELD模型公式对125例慢性重型肝炎患者进行评分,观察其3月内的死亡率,以受试者工作特征(ROC)曲线和曲线下面积评价该系统预测患者短期(3个月)预后的准确性,并求出作为判断患者3个月内死亡与否的MELD最佳临界值。结果 81例(64.8%)患者在3个月内死亡,死亡组MELD分值(34.6±10.5)明显高于存活组(23.5±5.8)(P<0.001);MELD<20、20≤MELD<30、30≤MELD<40及MLED≥40的患者病死率分别为14.3%、54.9%、79.5%和95.2%;应用该模型预测患者3个月内死亡与否的最佳MLED临界值为26,c-统计值为0.802,敏感度为80.2%,特异度为79.5%,准确性为80.0%。结论患者短期内(3个月)死亡的危险性随其MELD分值的增加而上升,MELD模型对慢性重型肝炎患者的病情及预后有很好的预测价值。
Objective To investigate the predictive ability and clinical value of the end-stage liver disease (MLED) scoring system in short-term (3 months) prognosis of patients with chronic severe hepatitis. Methods A total of 125 patients with chronic severe hepatitis were scored by MELD model formula and their 3-month mortality was observed. The system was used to predict the short-term (3-month) prognosis of patients with ROC curve and area under the curve And determine the best cut-off point for MELD as a measure of the patient’s death within 3 months. Results 81 patients (64.8%) died within 3 months. The MELD score of death group (34.6 ± 10.5) was significantly higher than that of survival group (23.5 ± 5.8) (P <0.001); MELD <20,20≤MELD <30 , And those with 30≤MELD <40 and MLED≥40 were 14.3%, 54.9%, 79.5% and 95.2%, respectively. The optimal MLED threshold for predicting the death within 3 months using this model was 26, c - Statistical value was 0.802, sensitivity was 80.2%, specificity was 79.5%, and accuracy was 80.0%. Conclusions The risk of death in short term (3 months) increases with the increase of MELD score. MELD model has a good predictive value for the severity and prognosis of patients with chronic severe hepatitis.