Surgery versus endovascular treatment for spinal dural arteriovenous fistulas: a multicenter experience and systematic literature review

Surgery and endovascular therapy are the primary treatment options for spinal dural arteriovenous fistula (SDAVF). Due to the absence of a consensus regarding which therapy yields a superior outcome, we conducted a comparative analysis of the surgical and endovascular treatment of SDAVF through a multicenter case series and a systematic literature review. Patients with SDAVF, surgically or endovascularly treated at four neurosurgical centers from January 2001 to December 2021, were included in this study. Level of SDAVF, primary treatment modality, baseline and post-procedural neurological status were collected. The primary outcomes were failure, complication rates, and a newly introduced parameter named as therapeutic delay. A systematic review of the literature was performed according to PRISMA-P guidelines. The systematic review identified 511 papers, of which 18 were eligible for analysis, for a total of 814 patients, predominantly male (72%) with a median age of 61 and mainly thoracic SDAVFs (65%). The failure rate was significantly higher for endovascular therapy (20%) compared to surgery (4%) (p < 0.01). Neurological complications were generally rare, with similar rates among the two groups (endovascular 2.9%; surgery 2.6%). Endovascular treatment showed a statistically significantly higher rate of persistent neurological complications than surgical treatment (2.9% versus 0.2%; p < 0.01). Both treatments showed similar rates of clinical improvement based on Aminoff Logue scale score. The multicenter, retrospective study involved 131 patients. The thoracic region was the most frequent location (58%), followed by lumbar (37%). Paraparesis (45%) and back pain (41%) were the most common presenting symptoms, followed by bladder dysfunction (34%) and sensory disturbances (21%). The mean clinical follow-up was 21 months, with all patients followed for at least 12 months. No statistically significant differences were found in demographic and clinical data, lesion characteristics, or outcomes between the two treatment groups. Median pre-treatment Aminoff-Logue score was 2.6, decreasing to 1.4 post-treatment with both treatments. The mean therapeutic delay for surgery and endovascular treatment showed no statistically significant difference. Surgical treatment demonstrated significantly lower failure rates (5% vs. 46%, p < 0.01). In the surgical group, 2 transient neurological (1 epidural hematoma, 1 CSF leak) and 3 non-neurological (3 wound infections) complications were recorded; while 2 permanent neurological (spinal infarcts), and 5 non-neurological (inguinal hematomas) were reported in the endovascular group. According to the literature review and this multicenter clinical series, surgical treatment has a significantly lower failure rate than endovascular treatment. Although the two treatments have similar complication rates, endovascular treatment seems to have a higher rate of persistent neurological complications.