E-085 Hemorrhagic dural arteriovenous fistulas: the role of early endovascular treatment to prevent re-bleeding

Purpose Intracranial dural arteriovenous fistulas (dAVFs) constitute approximately 10%–15% of all intracranial vascular malformations. More rarely, these lesions can lead to intracranial hemorrhage. Early re-bleeding rates from unsecured, ruptured dAVFs have been reported as high as 50%. Therefore, early intervention is paramount. Our aim of this study was to investigate the angiographic and clinical outcomes of early endovascular intervention in this rare disease presentation. Methods A HIPPA compliant neuroendovascular procedural database was reviewed for all treated dAVFs from 1999 to 2017. Inclusion criteria for the study included: 1) hemorrhagic presentation and 2) early endovascular intervention defined as less than 48 hours from initial ictus. Clinical demographics, angiographic and procedural details (Anatomic location, Borden classification, route of embolization, material used during embolization, number of retreatments, and need for completion microsurgical ligation), as well as clinical and angiographic follow up outcomes (modified Rankin scale and occlusion status on last follow-up imaging) were recorded. Appropriate statistical analyses were performed on parametric and non-parametric variables as applicable (statistical significance, α=0.05). Results A total of 262 dAVFs were identified, of those only 18 patients met inclusion criteria. Overall, the majority of patients were male (12/18; 66%), with a median age of 56 years (IQR 46–68 years). Hemorrhagic presentations included isolated intraparenchymal hemorrhage (IPH) (8/18; 44%), isolated subarachnoid hemorrhage (SAH) (7/18; 39%), SAH with subdural hematoma/intraventricular hemorrhage (IVH) (2/18; 11%), and isolated IVH (1/18; 6%). Almost all the hemorrhagic dAVF were of the aggressive type (Borden type II/III, 17/18; 94%). Anatomically, the most common location was tentorial (8/18; 44%). The anterior fossa, torcula, superior sagittal sinus, and transverse/sigmoid sinus all proportionately were the next common location (2/18; 11%, each respective group). Foramen magnum and straight sinus were the least common locations (1/18; 6%, each respective group). Transarterial embolization was performed most commonly (15/18; 83%), followed by transvenous approach (3/18; 17%). nBCA glue or Onyx embolization was most commonly used (15/18; 83%) with coils used exclusively in all transvenous cases (3/18; 17%). At last clinical follow-up, two patients succumbed to complications related to initial hemorrhage with the majority of patients (14/16; 88%) achieving good clinical outcome (mRS=0–2). On imaging follow-up when residual dAVF or incomplete embolization was noted (5/12; 42%) all cases underwent endovascular retreatment. Overwhelmingly, endovascular re-treatment failed in the majority of the cases (3/5; 60%); requiring microsurgical ligation for completion therapy. In no instance was early or late re-bleeding observed on last clinical follow-up, median 5 months (IQR 2–19 months). Conclusion Early endovascular intervention for ruptured dAVFs is feasible and effectively secures these lesions with no re-bleeding. Attention must be paid to lesions, which are incompletely embolized/show residual filling, as microsurgical ligation is imperative for completion therapy. Disclosures T. Madaelil: None. J. Grossberg: None. M. Cawley: None. F. Tong: None. J. Dion: 5; C; Vice President, Scientific Affairs at MicroVention-Terumo. B. Howard: None.