Background The prevalence of pial arterial supply to cranial dural arteriovenous fistulas (dAVF) and its implication in the management of these fistulas is not well characterized. We performed a retrospective study to characterize pial arterial supply to dural arteriovenous fistulas and the implications for treatment.

Methods Consecutive patients evaluated over a 12-year period were retrospectively reviewed. Angiograms were reviewed to characterize dAVF angioarchitecture and the presence of pial artery supply. Pial artery supply was categorized as dilated pre-existing dural branches and pure pial supply. We then studied the association between pial artery supply and clinical, angiographic, and treatment features.

Results A total of 201 patients were included of which 27 (13.4%) had pial artery supply. Of these, 11 had supply from dilated pre-existing dural branches, nine had pure pial supply,and seven had both. There was a higher rate of dAVF rupture in the pial supply group (30.8% vs 9.8%, P=0.003) and these fistulas had a higher rate of Borden 2 and 3 (88.9% vs 38.4%, P<0.0001). Fistulas with pial artery supply had similar rates of endovascular and gamma knife treatment, but were more likely to undergo surgery than those without pial supply (25.9% vs 10.4%, P=0.03). Major complication rates were similar between groups (0% vs 1.1%, P=0.55).

Conclusions More than 10% of dAVFs also have pial supply but this is not a contraindication to embolization. In our study pure pial supply was associated with a more aggressive fistula and was most common in tentorial dAVFs.