改良ICH量表对入院清醒脑出血患者28天预后的预测价值

Modified ICH Scale Predictive Value on the 28 Day Prognosis on Clear Admission in Patients with Cerebral Hemorrhage

目的:评价改良ICH评分对入院时意识清醒的自发性脑出血患者28天预后的预测价值。方法:225例自发性脑出血患者为研究对象,根据入院时GCS评分筛选出意识清醒患者,记录性别、年龄、入院NIHSS评分等临床特征,以28d时预后结局为评价指标,对影响不同结局的相关因素进行单因素回归分析。根据筛选出的影响清醒脑出血患者预后的危险因素,建立改良ICH量表,计算其灵敏度、特异度、Youden指数以及最佳诊断点,以评估改良ICH量表的诊断效能。结果:患者入院时NIHSS评分、丘脑出血、随机血糖或有糖尿病史、合并症、再出血等与28天预后密切相关。改良ICH量表诊断死亡的ROC曲线下面积是0.880,最佳诊断点是≥4,总准确性92.4%;诊断预后良好(m RS≤2)的ROC曲线下面积是0.762,最佳诊断点是≤2,总准确性是68.0%。结论:改良ICH量表对入院清醒的自发性脑出血患者28天预后预测、病情评估有一定参考价值。

Objective： To evaluate modified ICH score predictive value for the 28 day prognosis in patients with spontaneous intracerebral hemorrhage in conscious admission.Methods：225 cases of spontaneous cerebral hemorrhage patients as the research object,according to admission GCS score were screened with the conscious patients,the clinical features of recorded the sex,age,admission NIHSS score.In 28 d outcomes as evaluation indexes,factors relevant to different outcomes of single factor regression analysis.According to the screening effect of awake brain hemorrhage risk factors for the prognosis of the patients,to develop the modified ICH scale calculated the sensitivity,specificity,Youden index and the best diagnostic point,to assess the amount of modified ICH diagnostic efficiency.Results：Patients on admission NIHSS score,hemorrhage of thalamus,random blood glucose or had a history of diabetes mellitus,complications,bleeding,and 28 days is closely related to the prognosis.The modified ICH scale area under the ROC curve for diagnosis of death was 0.880,optimal diagnostic point is larger than or equal to 4,total accuracy 92.4%; diagnosis prognosis is good（MRS is less than or equal to 2） under the ROC curve area is 0.762,optimal diagnostic point is less than or equal to 2,the total accuracy was 68.0%.Conclusion：The modified ICH volume table clear admission of spontaneous cerebral hemorrhage patients for 28 day prognosis prediction and assessment of the disease have a certain reference value.