尿毒症行维持性透析并发脑出血患者预后的影响因素及风险模型构建

Risk factors and risk model construction for prognosis of uremia patients complicated with intracerebral hemorrhage undergoing maintenance dialysis

目的探讨影响尿毒症行维持性透析并发脑出血患者预后的因素,并构建风险预测模型。方法尿毒症行维持性透析并发脑出血患者54例,均继续血液透析或腹膜透析治疗,并给予降颅压、血压,止血等对症治疗,随访6个月评定Rankin评分,根据Rankin评分分为预后良好组(Rankin评分≤2分)和预后不良组(Rankin评分>2~6分)。比较2组临床资料,多因素logistic回归分析尿毒症行维持性透析并发脑出血患者预后的影响因素,根据影响因素权重构建风险模型,并采用ROC曲线评估风险模型的价值。结果预后良好组年龄[(44.34±12.44)岁]较预后不良组(65.92±10.01)岁]小,有糖尿病比率(11.54%)、平均动脉压[(120.53±11.76)mm Hg]、空腹血糖[(5.92±1.88)mmol/L]较预后不良组[25.00%、(129.27±14.89)mm Hg、(8.16±2.38)mmol/L]低,脑出血量[(13.20±6.79)mL]较预后不良组[(31.06±11.56)mL]少,血红蛋白[(100.88±20.72)g/L]、白蛋白[(38.97±3.96)g/L]水平较预后不良组[(86.18±16.70)、(34.12±8.31)g/L]高(P<0.05);2组性别比例、有心血管疾病比率、脑出血破入脑室比率、脑出血部位、透析方式、透析龄、血小板计数、血小板与淋巴细胞计数比值、凝血酶原时间、活化部分凝血酶原时间、血肌酐、尿素氮比较差异无统计学意义(P>0.05);多因素logistic回归分析结果显示,年龄(OR=1.224,95%CI:1.027~1.459,P=0.024)、平均动脉压(OR=1.051,95%CI:1.006~1.098,P=0.027)、脑出血量(OR=1.320,95%CI:1.081~1.612,P=0.007)、白蛋白(OR=0.899,95%CI:0.818~0.989,P=0.028)是维持性透析并发脑出血尿毒症患者预后的影响因素;根据logistic回归分析构建危险模型,ROC曲线分析结果显示,模型Youden指数最大值为0.816,最佳截断值为0.764,预测预后不良的AUC为0.952,灵敏度为87.3%,特异度为98.7%。结论年龄、平均动脉压、脑出血量、白蛋白是尿毒症行维持性透析治疗并发脑出血患者预后的影响因素,构建的风险模型拟合效度好,预测价值高。

Objective To investigate the risk factors for the prognosis of intracerebral hemorrhage(ICH)patients with uremia undergoing maintenance dialysis and to construct the risk models.Methods Fifty-four ICH patients with uremia undergoing maintenance dialysis received hemodialysis or peritoneal dialysis as well as the symptomatic treatment including lowering intracraninal pressure,control blood pressure and hemostasis.All patients were followed up for 6 months to assess Rankin ratings,and were divided into good prognosis group(Rankin score≤2)and poor prognosis group(Rankin score>3 to 6).The clinical data were compared between two groups.The influencing factors for prognosis were analyzed by multivariate logistic regression model,the risk models were constructed according to factor weights,and the value of risk models was evaluated by ROC.Results The patients were younger in good prognosis group((44.34±12.44)years)than in poor prognosis group((65.92±10.01)years)(P<0.05).The percentages of diabetes(11.54%),mean arterial pressure((120.53±11.76)mm Hg)and fasting plasma glucose((5.92±1.88)mmol/L)in good prognosis group were lower than those in poor prognosis group(25.00%,(129.27±14.89)mm Hg,(8.16±2.38)mmol/L)(P<0.05).The volume of ICH was lower in good prognosis group((13.20±6.79)mL)than that in poor prognosis group((31.06±11.56)mL)(P<0.05).The levels of hemoglobin and albumin were higher in good prognosis group((100.88±20.72),(38.97±3.96)g/L)than those in poor prognosis group((86.18±16.70),(34.12±8.31)g/L)(P<0.05).There were no significant differences in the sex ratio,percentages of cardiocerebral vascular diseases,ICH breaking into cerebral ventricles,location of ICH,dialysis age,platelet count,platelet to lymphocyte ratio,prothrombin time,activated partial prothrombin time,blood creatinine and urea nitrogen between two groups(P>0.05).Multivariate logistic regression analysis results showed age(OR=1.224,95 CI%:1.027-1.459,P=0.024),mean arterial pressure(OR=1.051,95 CI%:1.006-1.098,P=0.027),volume of ICH(OR=1.320,95 CI%:1.081-1.612,P=0.007)and albumin(OR=0.899,95 CI%:0.818-0.989,P=0.028)were the influencing factors for the prognosis of ICH patients with uremia undergoing maintenance dialysis.Risk models were constructed by logistic regression analysis.ROC showed the maximum value of the Youden index was 0.816,the optimal cut-off value was0.764,the AUC was 0.952,the sensitivity was 87.3%,and the specificity was 98.7%.Conclusion Age,mean arterial pressure,volume of ICH and albumin are the influencing factors for the prognosis of ICH patients with maintenance dialysis.The constructed risk model has a good evaluation performance and certain clinical application value.