妊娠期特有性肝病85例临床分析

Pregnancy-specific liver diseases: clinical analysis of 85 cases

目的探讨妊娠期特有性肝病患者的预后因素,并使用统计学方法得出新的预后模型,并探讨终末期肝病模型(MELD)及新模型对妊娠期特有性肝病患者的预后评估价值。方法收集广州医科大学附属第三医院2015年1月1日至2017年12月31日以黄疸和肝功能不全入院,并诊断为妊娠期急性脂肪肝(ALFP)、子痫前期、HELLP综合征而入住的妊娠期患者共85例,根据预后情况分为死亡组和生存组,根据MELD、NEW模型对患者进行评分,评估不同模型与患者预后转归的关系,计算各模型的敏感度、特异度、阳性预测值和阴性预测值,并通过ROC曲线衡量各模型对妊娠期急性肝衰竭预后的预测能力。结果 85例患者中,存活组共71例,死亡组共14例,死亡率为16.5%,单因素分析显示,肝性脑病及实验室指标TBIL、DBIL、Cr、INR、ALP、TP、ALB等差异有统计学意义(P<0.05)。MELD模型和含有INR及一个常数项的新模型均表达出良好的预测价值,ROC曲线下面积分别为0.952(95%CI:0.906～0.997,P<0.05)、0.982(95%CI:0.958～1.0,P<0.05),敏感度分别为92.9%、100%,特异度分别为87.3%、94.4%。两种模型分值与患者病死率呈正相关,分值越高,患者病死率也越高。结论 MELD模型和新模型均能在不同程度上反映妊娠期特有性肝病患者的预后,具有良好的临床指导价值。

Objective To explore the prognosis factors of pregnancy-specific liver diseases and to predict mortality based on features of it,to investigate the value of the model for end-stage liver disease(MELD) to predict mortality among patients with pregnancy-specific liver diseases.Methods Eighty-five patients who were diagnosed as pregnancyspecific liver diseases such as acute fatty liver of pregnancy(ALFP),hemolysis,elevated liver enzymes and low platelet syndrome;and preeclampsia-associated liver disease in the Third Affiliated Hospital of Guangzhou Medical University from Jan.1 st 2015 to Dec.31 th 2017 were enrolled in the retrospective study,and they were divided into death and survival groups.The relationship of the MELD scores,NEW model with clinical outcomes was analyzed.Sensitivity,specificity,positive predictive value,negative predictive value of each models were calculated.The values of MELD scores and NEW model were evaluated by receiver operating characteristic curves(ROC).Results Fourteen patients died,with the mortality rate of 16.5%.Univariate analysis identified encephalopathy,serum total protein,albumin(ALB),total bilirubin(TBIL),direct bilirubin(DBIL),alkaline phosphatase(ALP),serum creatinine(CCr),and international normalized ratio(INR) were significant variables.MELD score and a model based on only 1 variable(INR) accurately predicted mortality.It was shown by receiver operating characteristic curve that the area under ROC curve(AUC) of these model were 0.952(95% CI:0.906-0.997,P<0.05),0.982(95% CI:0.958-1.0,P<0.05),the sensitivity were 92.9%,100%,the specificities were 87.3%,94.4%.The MELD scores and the NEW model had positive correlation with the prognosis of patients.The mortality was elevated gradually as the equation scores increased.Conclusion A new Logistic model based on only 1 variable(INR) is comparable with the MELD model in predicting mortality among pregnancy-specific liver diseases patients.