The natural history of cranial dural arteriovenous fistulae with cortical venous reflux--the significance of venous ectasia.

BACKGROUND The quoted risk of hemorrhage from dural arteriovenous fistulae with cortical venous reflux varies widely, and the influence of angiographic grade on clinical course has not previously been reported. OBJECTIVE To assess the risk of hemorrhage and the influence of angiographic grade on this risk, compared with known predictors of hemorrhage such as presentation. METHODS Seventy-five fistulae with cortical venous reflux identified in our arteriovenous malformations clinic between 1992 and 2007 were followed up clinically, and their angiograms were reviewed. RESULTS There were 8 hemorrhages in 90 years of follow-up. The annual incidence of hemorrhage before any treatment was 13%, and 4.7% after partial treatment, giving an overall incidence of 8.9% before definitive treatment. Borden and Cognard grades were poor discriminators of risk for lesions with the exception of Cognard type IV lesions. These lesions, characterized by venous ectasia, had a 7-fold increase in the incidence of hemorrhage (3.5% no ectasia vs 27% with ectasia). Patients presenting with hemorrhage (20%) or nonhemorrhagic neurological deficit (22%) had a higher incidence of hemorrhage than those with a benign presentation (4.3%), but this may be directly linked to the presence of venous ectasia. CONCLUSION In this series untreated dural arteriovenous fistulae with cortical venous reflux had a 13% annual incidence of hemorrhage after diagnosis. There was a significant difference between those with and without venous ectasia. This should be confirmed by further studies, but probably defines a high-risk subgroup of patients that requires rapid intervention.