高血压脑出血预后不良风险预测模型的构建与验证

Construction and validation of a risk predictive model for poor prognosis in patients with hypertensive intracerebral hemorrhage

目的探讨高血压脑出血预后不良的相关因素,构建高血压脑出血预后不良风险预测模型,并验证该模型的预测效果。方法回顾性分析同济大学附属第十人民医院神经外科2016年1月—2020年12月收治的高血压脑出血患者531例,按照术后3个月mRS评分分为预后良好组(mRS≤2分)142例,和预后不佳组(mRS>2分)389例,对两组资料行单因素分析、回归模型筛选高血压脑出血预后不良的独立相关因素,并建立风险预测模型及内部验证。结果基底节出血、顶叶出血、枕叶出血、出血量>30 mL、偏瘫、意识障碍、头痛、头晕、恶心呕吐、糖尿病、红细胞分布宽度、D-二聚体水平均是高血压脑出血预后不良的影响因素,顶叶出血、出血量>30 mL、偏瘫、意识障碍、恶心呕吐、D-二聚体水平是高血压脑出血预后不良的独立相关因素;构建高血压脑出血预后不良预测模型,建模组AUC=0.8950,验证组AUC=0.8538,模型预测能力较好。结论该研究构建的风险预测模型一致性和预测效果良好,可为临床预测评估高血压脑出血患者预后不良风险提供借鉴,早期实行治疗干预,为临床治疗提供参考。

Objective To construct and validate a risk predictive model for poor prognosis in patients with hypertensive intracerebral hemorrhage.Methods Clinical data of 531 patients with hypertensive intracerebral hemorrhage admitted to the Department of Neurosurgery,Shanghai Tenth People's Hospital from January 2016 to December 2020,were retrospectively analyzed.According to the mRS score at 3 months after surgery,there were 142 cases with mRS≤2 points(good prognosis group),and 389 cases with mRS>2 points(poor prognosis group).The factors related to the poor prognosis of hypertensive intracerebral hemorrhage were screened with univariate analysis,then multivariate Logistic regression models for predicting poor prognosis were established and validated.Results Univariate regression analysis showed that basal ganglia hemorrhage,parietal hemorrhage,occipital hemorrhage,hemorrhage greater than 30 mL,hemiplegia,disturbance of consciousness,headache,dizziness,nausea and vomiting,diabetes,red blood cell distribution width,and D-dimer(DD)were factors affecting the prognosis of hypertensive intracerebral hemorrhage.Multivariate regression analysis showed that parietal hemorrhage,hemorrhage volume greater than 30 mL,hemiplegia,disturbance of consciousness,nausea and vomiting,and DD levels were independent factors of poor prognosis in patients with hypertensive intracerebral hemorrhage.A risk predictive model for poor prognosis of hypertensive intracerebral hemorrhage was constructed,the area under the ROC curve(AUC)was 0.8950 in the modeling group,and that in the validation group was 0.8538,with a good predictive ability.Conclusion The risk prediction model constructed in this study has good consistency and predictive effect,which can provide reference for clinical prediction and evaluation of the poor prognosis risk in patients with hypertensive intracerebral hemorrhage.