Predictive variables for the presence of vascular malformations as the cause of basal ganglia hemorrhages

To evaluate potential bleeding sources and predictive variables for basal ganglia hemorrhage. Fifty-seven patients with basal ganglia hemorrhage admitted to our neurosurgical ICU between 2005 and 2016 were retrospectively reviewed. Univariate and multivariate logistic analyses were used to assess predictive variables for identifying the bleeding source and outcome. ROC curves were plotted for a cutoff value for age and hematoma volume in patients with a vascular pathology and patients without a vascular pathology. In 19 patients, a vascular pathology was found as a bleeding source for basal ganglia hemorrhage (33.3%; 95% CI 0.33 [0.21; 0.47]). Most of the arteriovenous malformations (AVMs) were small sized (61.1%) with deep venous drainage (94.4%). A single vein was found in 17 (77.8%) AVMs. Patients younger than 50 years were more likely to have a vascular pathology (AUC of 0.85 [95% CI 0.73; 0.98]; p = 0.001; cutoff value 46.5 years). Four (21.1%) patients older than 50 years suffered an AVM hemorrhage; 75% of them were located ventricular or thalamic. Hematoma volume in patients with AVM hemorrhage was predominantly less than 30 cm³ (AUC of 0.86 [95% CI 0.76; 0.96]; p = 0.001; cutoff value 12.6 cm³). Outcome in patients with a vascular pathology was more often favorable as in patients with a spontaneous hemorrhage (92.9% vs. 7.1%; p = 0.001). Young age and hematoma volume are significant predictors for presence of a bleeding source and outcome in basal ganglia hemorrhage. These criteria must be taken into account in the emergency diagnostics and therapy in order to achieve a rapid and sufficient result. Outcome in patients with AVM hemorrhage in basal ganglia is more often favorable.