摘要
目的研究终末期肝病模型(MELD和MELD-Na)评分系统预测慢性乙型重型肝炎患者预后的价值。方法采用药物治疗的慢性乙型重型肝炎住院患者42例,记录人院时、1、2和3周的MELD和MELD-Na评分,应用c-statistic(等同于ROC曲线下面积)评估每一预测模型的准确性。观察的终点时间为3个月。结果患者入院时MELD和MELD-Na评分c-statistic分别为0.728和0.731;入院1周后c-statistic均为0.883。入院1周后判断患者3月内预后的最佳临界值分别是26.14(敏感性91.7%,特异性81.8%)和29.02(敏感性100.0%,特异性77.3%)。结论MELD和MELD-NA评分系统可以预测慢性乙型重型肝炎短期预后,入院后1周评分优于入院时评分。
Objective To investigate the predicting the prognosis of model for end-stage liver disease (MELD and MELD-Na) of chronic severe hepatitis B under chemotherapy. Method 42 cases of chronic severe hepatitis B were sellected, and the MELD and MELD-Na scores were calculated at the day of admission, and 1, 2 and 3 weeks after hospitalization. The calculation lasted for 3 months, and the validity of the model was determined by means of the concordance(c) statistic. Results The c-statistic of MELD was 0. 728 and 0. 731 for MELD-Na at the day of admission. Both of the c-statistic was 0. 883 at 1 week later. The optimal cut-off values predicting the prognosis of patients at 1 week after adnission for MELD was 26. 14 and 29.02 for MELD-Na. Conclusion MELD or MELD-Na score could relatively predict the short-term prognosis of patients with chronic severe hepatitis B. The predictive capability of MELD at 1 week after admission is better than the day of admission.
出处
《寄生虫病与感染性疾病》
CAS
2007年第2期64-66,共3页
Parasitoses and Infectious Diseases