Cerebral venous drainage in patients with dural arteriovenous fistulas: correlation with clinical presentation

OBJECTIVE Dural arteriovenous fistulas (DAVFs) are abnormal, acquired arteriovenous connections within the dural leaflets. Their associated symptoms may be mild or severe and are related to the patient's venous anatomy. With the hypothesis that the patient's venous anatomy determines the development of symptoms, the authors aimed to identify which venous anatomy elements are important in the development of major symptoms in patients with a DAVF. METHODS A multicenter study was performed based on the retrospective analysis of cerebral angiographies with systematic assessment of brain drainage pathways (including fistula drainage) in patients over 18 years of age with a single DAVF. The patients were divided into two groups: those with minor (group 1, n = 112) and those with major (group 2, n = 89) symptoms. Group 2 was subdivided into two groups: patients with hemorrhage (group 2a, n = 47) and patients with severe nonhemorrhagic symptoms (group 2b, n = 42). RESULTS The prevalence of stenosis in DAVF venous drainage and the identification of tiny anastomoses between venous territories were significantly higher in group 2 (32.6% and 19.1%, respectively) compared with group 1 (2.68% and 5.36%, respectively). Stenosis of DAVF venous drainage was significantly more frequent in group 2a than in group 2b (51.1% vs 11.9%, p < 0.001). Group 2b patients had increased prevalence of shared use of the cerebral main drainage pathway (85.0% vs 53.2%, p = 0.002), the absence of an alternative route (45.0% vs 17.0%, p = 0.004), and the presence of contrast stagnation (62.5% vs 29.8%, p = 0.002) compared with group 2a patients. In patients with high-grade fistulas, the group with major symptoms had increased prevalence of a single draining direction (31.3% vs 8.33%, p = 0.003), stenosis in the draining vein (35.0% vs 6.25%, p = 0.000), the absence of an alternative pathway for brain drainage (31.3% vs 12.5%, p = 0.017), and the presence of contrast stagnation (48.8% vs 22.9%, p = 0.004). CONCLUSIONS Major symptoms were observed when normal brain tissue venous drainage was impaired by competition with DAVF (predominance in group 2b) or when DAVF venous drainage had anatomical characteristics that hindered drainage, with consequent venous hypertension on the venous side of the DAVF (predominance in group 2a). The same findings were observed when comparing two groups of patients with high-grade lesions: those with major versus those with minor symptoms.