Masè G, Zorzon M, Biasutti E, Tasca G, Vitrani B, Cazzato G
Department of Neurology, University of Trieste, Italy.
Acta Neurol Scand. 1995 Apr;91(4):306-9. doi: 10.1111/j.1600-0404.1995.tb07011.x.
We reviewed the medical records and CT scans of 138 patients hospitalized for spontaneous supratentorial hemorrhage and conservatively treated. Seventeen clinico-anamnestic variables and five CT desumed findings were collected. Using univariate analysis we found eight significant predictors of 30-day mortality: intraventricular spread (IVS) of blood, volume of the hemorrhage, Glasgow Coma Scale (GCS) score, midline shift, hyperglycemia, pupillary abnormalities, limb paresis and gaze deviation. With multivariate logistic regression analysis we found three independent predictors of 30-day mortality: IVS of blood, volume of the hemorrhage and GCS score. Using these three independent variables we developed an easy model which allows an immediate estimate of the probability of survival with a high degree of sensitivity and specificity.
我们回顾了138例因自发性幕上出血住院并接受保守治疗患者的病历和CT扫描结果。收集了17项临床-既往史变量和5项CT推测结果。通过单因素分析，我们发现了30天死亡率的8个显著预测因素：血液脑室扩散（IVS）、出血量、格拉斯哥昏迷量表（GCS）评分、中线移位、高血糖、瞳孔异常、肢体轻瘫和凝视偏斜。通过多因素逻辑回归分析，我们发现30天死亡率的3个独立预测因素：血液IVS、出血量和GCS评分。使用这3个独立变量，我们开发了一个简单的模型，该模型能够以高度的敏感性和特异性立即估计生存概率。
