Intracranial dural arteriovenous fistulas: association with cerebral venous thrombosis, baseline aggressiveness, and clinical outcomes. A retrospective multicenter study on 263 consecutive patients and literature review

Abstract

Objective: The pathogenesis of intracranial dural arteriovenous fistulas (icDAVFs) is controversial. Cerebral vein thrombosis (CVT) and venous hypertension are recognized predisposing factors. This study aimed to evaluate the incidence of association between icDAVF and CVT and describe baseline aggressiveness and clinical outcomes for icDAVFs associated with CVT. The authors also performed a literature review of studies reporting icDAVF associated with CVT.

Methods: Two hundred sixty-three consecutive patients in two university hospitals with confirmed icDAVFs were included. A double-blind imaging review was performed to determine the presence or absence of CVT close or distant to the icDAVF. Location, type (using the Cognard classification), aggressiveness of the icDAVF, clinical presentation, treatment modality, and clinical and/or angiographic outcomes at 6 months were also collected. All prior brain imaging was analyzed to determine the natural history of onset of the icDAVF.

Results: Among the 263 included patients, 75 (28.5%) presented with a CVT concomitant to their icDAVF. For 18 (78.3%) of 23 patients with previous brain imaging available, CVT preceding the icDAVF was proven (6.8% of the overall population). Former/active smoking (OR 2.0, 95% CI 1.079-3.682, p = 0.022) and prothrombogenic status (active inflammation or cancer/coagulation trouble) were risk factors for CVT associated with icDAVF (OR 3.135, 95% CI 1.391-7.108, p = 0.003). One hundred eighty-seven patients (71.1%) had a baseline aggressive icDAVF, not linked to the presence of a CVT (p = 0.546). Of the overall population, 11 patients (4.2%) presented with spontaneous occlusion of their icDAVF at follow-up. Seven patients (2.7%) died during the follow-up period. Intracranial DAVF + CVT was not associated with a worse prognosis (modified Rankin Scale score at 3-6 months: 0 [interquartile range {IQR} 0-1] for icDAVF + CVT vs 0 [IQR 0-0] for icDAVF alone; p = 0.055).

Conclusions: This was one of the largest studies focused on the incidence of CVT associated with icDAVF. For 6.8% of the patients, a natural history of CVT leading to icDAVF was proven, corresponding to 78.3% of patients with previous imaging available. This work offers further insights into icDAVF pathophysiology, aiding in identifying high-risk CVT patients for long-term follow-up imaging. Annual imaging follow-up using noninvasive vascular imaging (CT or MR angiography) for a minimum of 3 years after the diagnosis of CVT should be considered in high-risk patients, i.e., smokers and those with prothrombogenic status.

Keywords: MR angiography; MRI; brain; cerebral vein thrombosis; digital subtraction angiography; dural arteriovenous fistula.