摘要
目的了解慢加急性肝衰竭(ACLF)并发肝肾综合征(HRS)的临床特点,探讨预测ACLF患者发生HRS的因素。方法收集2012年1月-2014年5月在中山大学附属第三医院就诊的ACLF患者806例,根据有无并发HRS分为两组,分析两组间的临床指标和实验室检查结果,计量资料两组间比较采用t检验,计数资料两组间比较采用χ2检验。采用多因素Logistic回归分析法筛查预测ACLF患者发生HRS的独立指标,建立预测模型,绘制受试者工作特征曲线,评估模型预测能力的准确性。结果 ACLF患者中HRS组229例,无HRS组577例,单因素Logistic回归分析显示,肝性脑病、腹膜炎、感染、年龄、胱抑素C(Cys-C)、血肌酐(SCr)、血尿素氮(BUN)、前白蛋白(PA)、TBil、DBil、总胆固醇(TC)、K+、Na+、P、凝血酶原时间(PT)、凝血酶原活动度(PTA)、国际标准化比值(INR)、红细胞压积(Hct)为对预测HRS发生有意义的因素。多因素Logistic回归分析显示合并腹膜炎(OR=3.155,P<0.01)、Cys-C(OR=30.773,P<0.01)、SCr(OR=1.062,P<0.01)、HCO3-(OR=0.915,P<0.05)为预测ACLF患者发生HRS的独立影响因素。建立的预测模型具有良好的预测价值。结论合并腹膜炎、Cys-C、SCr、HCO3-是ACLF患者发生HRS的有效预测因素。
Objective To learn the clinical characteristics of concmxent acute - on - chronic liver failure （ ACLF ） and hepatorenal syndrome （HRS） , and to investigate the predictive factors for HRS in patients with ACLF. Methods A total of 806 patients with AClJF who were admitted to our hospital from January 2012 to May 2014 were selected and divided into two groups according to the incidence of concurrent HRS. Clinical indices and laboratory test results were analyzed in the two groups, and the multivariate logistic regression analysis was used to figure out independent indices for the prediction of HRS in patients with ACLF. A prediction model was established and the ree.eiver operating characteristic curve was drawn to evaluate the accuracy of the prediction model. Comparison of continuous data between the two groups was made by t test, and comparison of categorical data between the two groups was made byx2 test. Results In all patients with ACLF, 229 had HRS and 577 had no HRS. The univariate logistic regression analysis showed that hepatic encephalopathy, peritonitis, i.- feetion, age, eystatin C （Cys -C）, serum creatinine （SCr）, blood urea nitrogen, albumin, prealbumin, total bilirubin, direct bilirubin, total cholesterol, K ＋ , Na ＋ , phosphorus, Ca2 ＋ , prothrombin time, prothrombin activity, international normalized ratio, and henlatocrit were significant predictive factors for HRS. The muhivariate logistic regression analysis showed that concurrent peritonitis, Cys - C, SCr, and HCO3 - were independent predictive factors for HRS in patients with ACLF （OR =3. 155, P 〈0. 01 ; OR =30. 773, P 〈0. 01 ; OR = 1. 062, P 〈 0. 01 ; OR = 0. 915, P 〈 0.05 ）. The model was proved of great value in prediction. Conclusion Concurrent peritonitis, Cys - C, SCr, and HCO3 - are effective predictive factors for HRS in patients with ACLF.
出处
《临床肝胆病杂志》
CAS
2015年第9期1422-1425,共4页
Journal of Clinical Hepatology
