Dural arteriovenous fistulas in cerebral venous thrombosis

Abstract Background and purpose To explore the prevalence, risk factors, time correlation, characteristics and clinical outcome of dural arteriovenous fistulas (dAVFs) in a cerebral venous thrombosis (CVT) population. Methods We included patients from the International CVT Consortium registries. Diagnosis of dAVF was confirmed centrally. We assessed the prevalence and risk factors for dAVF among consecutive CVT patients and investigated its impact on clinical outcome using logistic regression analysis. We defined poor outcome as modified Rankin Scale score 3–6 at last follow‐up. Results dAVF was confirmed in 29/1218 (2.4%) consecutive CVT patients. The median (interquartile range [IQR]) follow‐up time was 8 (5–23) months. Patients with dAVF were older (median [IQR] 53 [44–61] vs. 41 [29–53] years; p < 0.001), more frequently male (69% vs. 33%; p < 0.001), more often had chronic clinical CVT onset (>30 days: 39% vs. 7%; p < 0.001) and sigmoid sinus thrombosis (86% vs. 51%; p < 0.001), and less frequently had parenchymal lesions (31% vs. 55%; p = 0.013) at baseline imaging. Clinical outcome at last follow‐up did not differ between patients with and without dAVF. Additionally, five patients were confirmed with dAVF from non‐consecutive CVT cohorts. Among all patients with CVT and dAVF, 17/34 (50%) had multiple fistulas and 23/34 (68%) had cortical venous drainage. Of 34 patients with dAVF with 36 separate CVT events, 3/36 fistulas (8%) were diagnosed prior to, 20/36 (56%) simultaneously and 13/36 after (36%, median 115 [IQR 38–337] days) diagnosis of CVT. Conclusions Dural arteriovenous fistulas occur in at least 2% of CVT patients and are associated with chronic CVT onset, older age and male sex. Most CVT‐related dAVFs are detected simultaneously or subsequently to diagnosis of CVT.