妊娠相关急性肾损伤预后因素模型建立和预测  

Modeling and prediction of pregnancy related acute kidney injury

作　　者：周宇涵[1] 董金菊[1] 樊伍峰[1] 张微[1] 黄玉琴[1] 习红丽[1] 

机构地区：[1]湖北医药学院附属襄阳医院产科,湖北襄阳441000

出　　处：《新乡医学院学报》2016年第5期406-410,共5页Journal of Xinxiang Medical University

摘　　要：目的探讨妊娠相关急性肾损伤(PR-AKI)的预后因素,建立预后预测模型,评估模型的预测效果。方法122例PR-AKI患者于产后1、3、6个月和1 a进行血清肌酐(SCr)、胎盘生长因子(PLGF)、内皮素-1(ET-1)及尿液中肾损伤分子-1(KIM-1)、β_2-微球蛋白(β_2-MG)水平检测,采用logistic向后逐步回归方法筛选与预后相关的因素,建立模型预测患者产后1 a的预后结局,并评估模型的预测效果。结果随着产后时间的延长,预后不良组患者血清ET-1及尿液中KIM-1、β_2-MG水平逐渐降低(P<0.05),血清PLGF水平逐渐升高(P<0.05),但血清SCr水平下降无统计学意义(P>0.05)。随着产后时间的延长,预后良好组患者血清SCr、ET-1及尿液中KIM-1、β_2-MG水平逐渐降低(P<0.05),血清PLGF水平逐渐升高(P<0.05)。产后1个月,预后良好组患者血清ET-1及尿液中KIM-1水平显著低于预后不良组(P<0.05),但2组患者血清SCr、PLGF及尿液中β_2-MG水平比较差异均无统计学意义(P>0.05)。产后3、6个月和1 a,2组患者各个指标比较差异均有统计学意义(P<0.05)。Logistic向后逐步回归分析显示,SCr、KIM-1、ET-1和产后时间与PR-AKI患者的预后相关(P<0.05),用以上筛选的指标拟合出最佳模型logit(P)=-21.202+0.113X1+0.354X2+0.105X5+0.879X6,模型拟合效果好(Hosmer-Lemeshow,P=0.284),模型的决定系数Cox-Snell R2=0.513,Nagelkerke R2=0.721,预测准确率为90.0%。该模型预测PR-AKI患者产后1 a的预后结局,灵敏度为0.857 1,特异度为1.0,Youden指数为0.8571,阳性预测值为1.0,阴性预测值为0.9773,受试者工作特征曲线下面积为0.939。结论 Logistic回归模型的拟合效果较好,对肾损伤预后结局的预测准确度较高,SCr、KIM-1和ET-1可以用于PR-AKI的筛查,对患者病情监测、干预及预后评估有重要意义。Objective To explore the prognostic factors of pregnancy related acute kidney injury（ PR-AKI）,establish a model to predict prognosis and evaluate the prediction effect of the model. Methods The levels of serum creatinine（ SCr）,placenta growth factor（ PLGF）,endothelin-1（ ET-1） and urine kidney injury molecule 1（ KIM-1）,β_2-microglobulin（ β_2-MG）of 122 patients with PR-AKI were detected at the time points of 1 month,3 months,6 months and 1 year after delivery. The prognostic factors were screened by logistic backward stepwise regression method. The model of predicting the prognosis of patients 1 year after delivery was established. The prediction effect of the model was evaluated. Results With the extension of postpartum time,the levels of serum ET-1 and urine KIM-1,β_2-MG in unfavourable prognosis group decreased gradually（ P 0. 05）,and serum PLGF level increased gradually（ P 0. 05）,but the decrease of serum SCr level was not statistically significant（ P 0. 05）. With the extension of postpartum time,the levels of serum SCr,ET-1 and urine KIM-1,β_2-MG in eusemia group decreased gradually（ P 0. 05）,and serum PLGF level increased gradually（ P 0. 05）. The levels of serum ET-1 and urine KIM-1 in eusemia group were significantly lower than those in unfavourable prognosis group 1 month after delivery（ P 0. 05）. But there was no significant difference in the levels of serum SCr,PLGF and urine β_2-MG between the two groups（ P 0. 05）. There were significant differences in each index between the two groups at the time points of 3 months,6 months and 1year after delivery（ P 0. 05）. Logistic backward stepwise regression showed that SCr,KIM-1,ET-1 and postpartum time were related to the prognosis of patients with PR-AKI（ P 0. 05）. The best model logit（ P） =- 21. 202 ＋ 0. 113X1＋ 0. 354X2＋0. 105X5＋ 0. 879X6 was fitted with the above indexes. The Model fitting effect was good（ Hosmer-Lemeshow test,P = 0. 284）with the coefficient of determination Cox-S

关 键 词：妊娠相关急性肾损伤 肌酐 肾损伤分子-1 Β2-微球蛋白 胎盘生长因子 内皮素-1 预后评估 

分 类 号：R714.25[医药卫生—妇产科学]