Recurrence Rates After Surgical or Endovascular Treatment of Spinal Dural Arteriovenous Fistulas

BACKGROUND: There is an increasing tendency to treat spinal dural arteriovenous fistulas (SDAVFs) endovascularly despite the lack of clear evidence favoring embolization over surgery. OBJECTIVE: To compare the initial failure and recurrence rates of primary treatment of SDAVFs by surgery and endovascular techniques. METHODS: A meta-analysis using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) standard was performed. All the English literature from 2004 onward was evaluated. From each article that compared the 2 treatment modalities, the odds ratio (OR) was calculated. Combined ORs were calculated with Review Manager 5.3 of The Cochrane Collaboration. RESULTS: A total of 35 studies harboring 1112 patients were assessed. Initial definitive fistula occlusion was observed in 588 of 609 surgical patients (96.6%; 95% confidence interval [CI], 94.8-97.8) vs 363 of 503 endovascularly treated patients (72.2%; 95% CI, 68.1-75.9; P < .001). The combined OR from 18 studies that assessed both treatment modalities (730 patients) was 6.15 (95% CI, 3.45-11.0) in favor of surgical treatment. Late recurrence (13 studies, 480 patients) revealed an OR of 3.15 (95% CI, 1.66-5.96; P < .001) in favor of surgery. In a subgroup, recurrence was reported in 10 of 22 patients (45%) treated with Onyx vs 8 of 35 (23%) treated with n-butyle-2-cyanoacrylate (OR, 2.51; 95% CI, 0.75-8.37; P = .13). CONCLUSION: Although hampered by inclusion of poor quality studies, this meta-analysis shows a definite advantage of primary surgical treatment of SDAVF over endovascular treatment in initial failure rate and late recurrences. The often-used argument that endovascular techniques have improved and therefore outweigh surgery is not supported by this meta-analysis.