Risk Factors Of Intracranial Hemorrhage In Patients With Dural Arteriovenous Fistula:Clinical And Angioarchitectural Analysis

Objective:To investigate the clinical and angioarchitectural characteristics, analyze the risk factors of intracranial hemorrhage in patients with dural arteriovenous fistula(DAVF), and provide a basis for treatment options of DAVF.Methods: The clinical data of 103 patients with DAVF who were admitted to the Department of Neurosurgery, Tongji Hospital, Huazhong University of Science and Technology from November 2009 to December 2015 were analyzed retrospectively. These patients were divided into hemorrhagic group and non-hemorrhagic group according to clinical presentations. The differences of potential risk factors between the two groups were analyzed, including sex, age, medical history(hypertension, hyperlipidemia, diabetes mellitus, alcohol consumption and smoking), angioarchitectural characteristics(the location of fistula, bilateral arterial feeders, arterial feeders from multiple systems, multiple or simple fistula, superficial cortical venous drainage, deep venous drainage, transverse sinus type and venous drainage pattern) and sinus thrombosis formation. Logistic regression analysis was performed to analyze the independent factors responsible for the intracranial hemorrhage in DAVF patients.Results: In univariate analysis, there were significant differences between the hemorrhagic group and non-hemorrhagic group on age(49.18±9.17 vs 43.73±12.43, p=0.02), sex(76.9% vs 56.3%, p= 0.034), smoking(48.7%vs23.4%, p=0.008), location of DAVF(p=0.001) [cavernous sinus(2.5% vs 31.2%, p<0.001), other location of fistula(59.2%vs29.7, p=0.003)], superficial cortical venous drainage(76.9% vs 43.8%, p=0.001), deep venous drainage(43.6% vs 18.8%, p=0.007) and arterial feeders from multiple systems(3.3% vs 54.7%, p=0.035), venous drainage pattern(p<0.001), the hemorrhage risk was higher in high-grade venous drainage pattern(Borden?type vs Borden ? type vs Borden ? type = 2.6% vs 28.2% vs 69.2%). However, there were no significant differences in the following variations: bilateral arterial feeders(43.6% vs 34.4%, p=0.35), multiple or simple fistula(56.4% vs 68.8%, p=0.205), transverse sinus type(p=0.129), sinus thrombosis formation(30.8% vs 29.7%,p=0.908). In multivariate logistic regression analysis, the venous drainage pattern was the only one found to be independently associated with hemorrhagic presentation. Especially, the location of fistula and age became the most risky factors for bleeding when inconsiderate the effect of venous drainage pattern. Large venous sinus(OR 12.507, 95%CI: 1.497-104.486 p=0.020) and other location of fistula(OR 24.842, 95%CI: 3.000-205.688 p=0.003) had a higher risk of hemorrhage presentation compared with DAVF in cavernous sinus. DAVF in other locations was more prone to bleed than in the large venous sinus(54.8% vs 37.5%). The result also showed that the patients in hemorrhage group were older than in non-hemorrhage group and the old age was a risk factor for intracranial hemorrhage(OR 1.049, 95%CI: 1.005-1.096, p=0.030).Conclusion: The venous drainage pattern is the most important risk factor for DAVF and the hemorrhage risk is higher in high-grade venous drainage pattern. The location and age become the most risky factors when inconsiderate the effect of venous drainage pattern. Sex, smoking, arterial feeders from multiple systems, superficial and deep venous drainage are confounding factors when evaluating the hemorrhagic risk. By contrast, hypertension, hyperlipidemia, diabetes mellitus, alcohol consumption, multiple or simple fistulas, bilateral arterial feeders, transverse sinus type and sinus thrombosis formation have no effect on the risk evaluation of DAVF.