Abstract TP446: Predictors of Hemorrhagic Severity in Arteriovenous Malformation Patients

Abstract

Introduction: One of the most frequent causes of non-traumatic intracerebral hemorrhage (ICH) in young adults is from a brain arteriovenous malformation (AVM) rupture. Understanding the potential spectrum of AVM outcomes is crucial in assessing treatment options. Although ICH mortality from ruptured brain AVMs is less frequent than for primary ICH, the severity of brain AVM hemorrhage is not well documented. We aimed to identify the clinical and anatomical variables associated with: (1) ICH volume and (2) intraventricular hemorrhage (IVH) presence to better prognosticate eventual clinical hemorrhage severity in brain AVM patients. Methods: ICH volume and presence of IVH (outcomes) were retrospectively assessed based on CT images for 150 patients admitted with ruptured AVMs and prospectively enrolled in our AVM database between 2000-2015. Parenchymal hemorrhage volumes were calculated using the ABC/2 method. Regression analysis was performed to assess the relationship between clinical and anatomical factors and ICH volume (proportional increase, PI) and presence of IVH (odds ratio, OR). Results: Older patients were less likely to have IVH (p=0.036). Exclusively deep venous drainage was associated with both smaller hemorrhage volumes (p<0.001) and an increased likelihood of IVH (p<0.001). Lobar AVM locations were associated with larger hemorrhage volumes (p<0.001). Venous stenosis, based on a 5-point grading scale (5 = complete occlusion), was associated with larger hemorrhage volumes (p=0.006) and a decreased likelihood of IVH (p=0.014). Exclusively deep venous drainage (PI=0.29, 95% CI: 0.19 - 0.44), lobar location (PI=2.18, 95% CI: 1.47 - 3.24), and venous stenosis (PI=1.34, 95% CI: 1.09 - 1.64) remained significant independent predictors of hemorrhage volume in a multi-predictor model (P<0.05). Similarly, exclusively deep venous drainage (OR=7.51, 95% CI: 2.69 – 21.00) and venous stenosis (OR=0.61, 95% CI: 0.41 – 0.90) remained significantly associated with IVH (P<0.05). Conclusions: With the lack of a widely adopted AVM natural history score, our study provides further information regarding potential hemorrhage severity in addition to the current knowledge of hemorrhagic risk to better scrutinize patients with unruptured AVMs.