Cys-C、β2-MG、Scr和BUN水平在预测慢加急性肝衰竭患者合并肝肾综合征中的价值

Value of cystatin C, β2 macroglobulin, serum creatinine, and blood urea nitrogen in predicting hepatorenal syndrome in patients with acute-on-chronic liver failure

目的探究慢加急性肝衰竭(acute-on-chronic liver failure,ACLF)合并肝肾综合征(hepatorenal syndrome,HRS)患者的胱抑素C(cystatin C,Cys-C)、β2微球蛋白(microglobulin,β2-MG)、血肌酐(serum creatinine,Scr)和血尿素氮(blood urea nitrogen,BUN)水平及其临床价值分析.方法对2014-02/2017-12于绍兴市中心医院就诊的36例ACLF合并HRS患者(HRS组)进行回顾性分析,选择同时间段内36例单纯ACLF患者(ACLF组)和50例慢性肝病患者(CLD组)作为对照组,比较3组的Cys-C、β2-MG、Scr和BUN水平等临床资料间差异,应用受试者工作曲线(receiver operating characteristic curves,ROC)评价应用Cys-C、β2-MG、Scr和BUN水平预测ACLF合并HRS的价值,并计算4种指标单独和联合预测的诊断效能.结果三组患者的Cys-C、β2-MG、Scr和BUN水平等指标间存在统计学差异(F=47.330、23.693、41.220、26.715;均P=0.000);与CLD和ACLF组相比,HRS患者的Cys-C(t=9.386、4.807,P=0.000、0.000)、β2-MG(t=30.265、4.116,P=0.000、0.000)、Scr(t=7.457、7.415,P=0.000、0.000)和BUN(t=6.608、5.014,P=0.000,0.000)水平均显著升高.ROC曲线显示,应用4种指标单独预测HRS时,Scr的AUC(0.799)和Cys-C(AUC=0.789)较高,β2-MG(AUC=0.741)次之,BUN(AUC=0.587)最低;应用Cys-C、β2-MG、Scr联合诊断后的诊断效能(AUC=0.910)明显高于单独诊断.以ROC曲线的最佳截点作为预测指标,3种指标联合预测HRS的诊断准确率80.33%,灵敏度91.67%,特异度75.58%,阳性预测值61.11%,阴性预测值95.59%,联合预测的灵敏度显著高于单独诊断(χ~2=10、8.692、7.432、3.956;P=0.002、0.003、0.006、0.047).结论慢加急性肝衰竭合并肝肾综合征患者Cys-C、β2-M G、Scr和BUN水平显著升高;应用Cys-C、β2-M G和Scr等3项指标联合预测慢加急性肝衰竭合并肝肾综合征的敏感度较高.

AIM To assess the predictive value of cystatin C （Cys-C）, β2 macroglobulin （β2-MG）, serum creatinine （Scr）, andblood urea nitrogen （BUN） for hepatorenal syndrome （HRS） in patients with acute-on-chronic liver failure （ACLF） METHODS Thirty-six ACLF patients with HRS （HRS group） treated at our hospital from February 2014 to December 2017 were analyzed retrospectively. Thirty-six patients with ACLF without HRS were selected as an ACLF group, and 50 patients with chronic liver disease （CLD） were selected as a CLD group. Cys-C, β2-MG, Scr, and BUN were compared between the three groups. The receiver operating characteristic （ROC） curve analysis was used to evaluate the diagnostic efficacy of Cys-C, β2-MG, Scr, and BUN, alone or in combination, in predicting HRS in patients with ACLF. RESULTS The levels of Cys-C, β2-MG, Scr, and BUN in the three groups were statistically different （F = 47.330, 23.693, 41.220, 26.715; P = 0.000 for all）. Compared with the CLD and ACLF groups, Cys-C （t = 9.386, 4.807, P = 0.000 for both）, β2-MG （t = 30.265, 4.116; P = 0.000 for both）, Scr （t = 7.457, 7.415; P = 0.000 for both）, and BUN （t = 6.608, 5.014; P = 0.000 for both） were significantly increased in the HRS group. ROC curve analysis showed that Scr had the highest AUC （0.799）, followed by Cys-C （AUC= 0.789）, β2-MG （AUC = 0.741）, and BUN （AUC = 0.910）. The combination of Cys-C,β2-MG, and Scr （AUC = 0.910） performed significantly better than any of the four indexes alone. Using the best cutoff point of the ROC curve as the predictive index, the diagnostic accuracy rate of the combination of Cys-C,β2-MG, and Scr for HRS was 80.33% （sensitivity, 91.67%; specificity, 75.58%; positive predictive value, 61.11%; negative predictive value, 95.59%）. The sensitivity of combined indexes was significantly higher than any of the four indexes alone （χ^2 = 10, 8.692, 7.432, 3.956; P = 0.002, 0.003, 0.006, 0.047）. CONCLUSION The levels of Cys-C, β2-MG, Scr, and BUN in ACLF patients with HRS significantly increase. The combination of Cys-C,β2-MG, and Scr has higher accuracy for predicting HRS in ACLF patients.