Comparative Analysis of Surgery, Endovascular, and Combined Approaches in the Management of spinal dural and epidural arteriovenous fistulas (sAVF). A 10-year systematic review and meta-analysis

Introduction: Spinal dural and epidural arteriovenous fistulas (sAVF) are the most common types of spinal shunts. Treating sAVF involves disrupting the arteriovenous shunt on the venous side. Surgery offers a curative approach with low recurrence rates, while embolization is less invasive with favorable long-term results but a higher rate of recurrences. This study aims to assess the relative superiority of these techniques, considering the improvements made in endovascular therapies in the last decade.

Methods: A systematic review of the PubMed, Embase and Cochrane databases was performed identifying relevant studies in English addressing the surgical and/or the endovascular treatment of the sAVF. The search terms used were: “arteriovenous spinal dural fistula”, “arteriovenous spinal epidural fistula”, “arteriovenous spinal fistula surgical”, “arteriovenous spinal epidural fistula endovascular”. The authors excluded technical notes, case reports, reviews and cadaveric studies, paediatric population, as well as intracranial and/or cranio-cervical junction studies. Data was extracted by hand and analysed in R version 4.3.1.

Results: The PRISMA protocol was used. 28 retrospective and 2 prospective clinical studies with a total of 1877 patients were included. 986 patients received surgical, 863 endovascular and 27 combined (endovascular + surgical) treatment. The pooled proportion of fistula total occlusions was 92% (CI 95% [85%,95%]) for surgery, 70% (CI 95% [62%,77%]) for endovascular and 84% (CI 95% [57%,95%]) for combined. Regarding the pooled proportion of fistula partial occlusions was 16% (CI 95% [10%,25%]) for surgery, 28% (CI 95% [22%,36%]) for endovascular and 17% (CI 95% [5%, 45%]) for combined. In patients who underwent surgery, the pooled proportion of adverse effects was 9% (CI 95% [7%,12%]), while for endovascular was 7% (CI 95% [4%,11%]). The pooled mean difference (reduction) of pre and post (1 year follow up) Aminoff and Logue’s scale was -2.35 (CI 95% [-3.72,-0.98]) for surgery and -1.11 (CI 95% [-3.75,1.53]) for endovascular.

Conclusion : Despite the remarkable improvement in the last decade made by the endovascular treatment, surgery still offers better clinical outcomes, especially in the long term, and a higher rate of complete occlusion of the shunt; however, a lower rate of complications is noted in favor of the endovascular treatment.