终末期肝病模型评分在预测肝移植患者术后急性肾损伤的发生及预后中的价值

Model for end-stage liver disease score in predicting acute kidney injury and prognosis in patients receiving liver transplantation

目的评价终末期肝病模型(MELD)评分在预测尸体肝移植患者术后急性肾损伤(AKI)的发生及预后中的价值。方法选择上海交通大学医学院附属仁济医院首次接受原位经典尸体肝移植手术的成人患者193例,收集患者的MELD评分等围术期资料。应用AKI网络工作组标准评估患者术后AKI的发生情况,观察MELD评分对AKI发生的影响。应用受试者操作特征(ROC)曲线评价MELD评分预测术后各期AKI的准确性。观察AKI患者的28 d及1年预后,运用ROC曲线评价术前MELD评分预测术后患者死亡的准确性。结果193例肝移植患者中共有116例(60.1%)发生术后AKI,其中AKI 1、2、3期分别占AKI患者的50.0%、21.6%和28.5%。术前MELD评分≥16.87分的患者术后AKI的发生率为71.6%,显著高于术前MELD评分<16.87分的患者(51.8%,P<0.05)。AKI患者的术前MELD评分为(19.53±12.08)分,显著高于非AKI患者的(12.86±9.71)分(P<0.05)。术前MELD评分预测术后AKI发生的曲线下面积(AUC)为0.671。术前MELD评分预测术后AKI 1、2、3期发生的AUC分别为0.469、0.575和0.775。以MELD评分18.50分作为诊断截点时,预测术后AKI 3期发生的敏感性和特异性分别为72.7%和71.2%。生存分析显示,AKI患者术后28 d的病死率为15.5%,死亡患者的术前MELD评分为(30.99±11.00)分,显著高于存活患者的(15.42±10.74)分(P<0.05),MELD评分预测肝移植术后AKI患者术后28 d死亡的AUC为0.842。以MELD评分24.84分作为诊断截点时,预测术后早期死亡的敏感度和特异度分别为77.8%和72.2%。术后1年AKI患者的病死率为25.9%,死亡患者的术前MELD评分为(23.22±14.08)分,显著高于存活患者的(15.65±10.78)分(P<0.05),MELD评分预测肝移植术后AKI患者1年死亡的AUC为0.651。结论肝移植术后AKI患者的术前MELD评分较高,术前MELD评分是术后发生AKI 3期的一种较为敏感的预测指标,是术后AKI患者早期死亡的非常敏感的预测指标,其在预测患者术后1年预后中的价值还有待于进一步研究。

Objective To evaluate the value of model for end-stage liver disease （MELD） score in predicting acute kidney injury （AKI） and its prognosis in patients receiving liver transplantation. Methods A total of 193 patients who received single donor orthotopic liver transplantation （OLT） in Renji transplantation center, Shanghai Jiaotong University School of Medicine from 2004 to 2006, were enrolled in this study. The MELD scores and other peri-operation data were collected. Post-OLT AKI was diagnosed by using AKIN criteria, and the outcomes of AKI patients were observed 28 days and 1 year later. The diagnostic and prognostic accuracy of MELD in prediction of AKI were analyzed by using receiver operating characteristic （ROC） curve. Results One hundred and sixteen （60.1%） patients developed post-OLT AKI in our group, with the incidences of AKI stage 1, 2 and 3 being 50.0%, 21.6%, and 28.5%, respectively. The AKI incidence was higher in patients with MELD score≥16.87 than that in patients with MELD scored 16.87 （71.6% vs. 51.8%, P〈0.05）. The MELD score in patients with AKI was higher than those without AKI （ 19.53± 12.08 vs. 12.86±9.71, P〈0.05）. The area under curve （AUC） of pre-operation MELD scores in predicting incidence of AKI was 0. 671; the AUC values in predicting AKI stage 1, 2 and 3 were 0.469, 0. 575 and 0. 775, respectively. When the cutoff of MELD was 18.50, the sensitivity and specificity in predicting AKI stage 3 were 72.7% and 71.2%, respectively. In AKI patients, post- OLT mortalities at 28 days and 1 year after operation were 15.5% and 25.9%, respectively. The MELD score of patients who died within 28 days after operation was significantly higher than that in the survived patients （30.99 ±11.0 vs. 15.42 ± 10.74, P〈0.05）. The MELD scores of patients who survived for one year after operation were significantly higher than that in the survived patients （23.22±14.08 vs. 15.65±10.78, P〈0.05）. The AUC values of MELD scores in predicting the 28 days and 1 year mortality were 0.842 and 0.651, respectively. Conclusion Patients with AKI post-OLT have higher pre-OLT MELD scores. The pre-OLT MELD score is a sensitive predictor for predicting AKI stage 3 and early mortality after liver transplantation. The value of the MELD scores in predicting long-term prognosis needs to be further investigated. （Shanghai Meal J, 2009, 32：195-200）