1450. Instrumentation Methods and Bone Graft Options in Atlantoaxial Posterior Fusion: Meta-Analysis and Systematic Review Encompassing 79 Articles and 4113 patients

Authors: Wataru Ishida, MD; Seba Ramhmdani, MD; Risheng Xu; Alexander Perdomo-Pantoja, MD; YuanXuan Xia, BA; Ethan Cottrill, BS; Ali Bydon, MD; Timothy Witham, MD; Sheng-fu Lo, MD (Baltimore, MD)

Introduction: Posterior atlantoaxial fusion is an important armamentarium for neurosurgeons to treat various pathologies involving the craniovertebral junction. With regards to methods of instrumentation and bone graft options, there are several options including C1 lateral mass and C2 pedicle/pars screws (C1-2Ss), transarticular screws (TASs), autograft versus allograft, and bone morphogenetic protein-2 (rhBMP-2). Hence, we aimed to compare each method by conducting meta-analyses encompassing 79 articles and 4113 patients. Methods: A web-based literature search (1980-2017, Pubmed) was performed in compliance with the PRISMA guidelines, using the keywords: "C1-2 (or atlantoaxial) fusion (or fixation)." Inclusion criteria were: 1) English, clinical studies reporting instrumentation and bone graft options for posterior C1-2 fusion and 2) C1-2 fusion status, misplaced screws, and vascular injury as the primary/secondary outcomes were reported. Exclusion criteria were 1) anterior surgery and 2) operated levels not confined to C1-2. Forest plots and funnel plots were drawn, and meta-regression analyses were performed to identify influential factors on the primary outcomes, assuming the random effect. Results: There were 8 prospective, non-randomized studies (10.1%) and 71 retrospective studies. The overall pooled fusion rates of patients with autograft versus allograft were significantly different (93.4% VS 89.5%, p=0.04), whereas morselized versus structural, rhBMP-2, sublaminar wiring, and C1-2 versus TAS did not affect the overall pooled fusion rates. The pooled rates of misplaced screws were affected by the screw types (TAS versus C1-2s, 4.7% versus 1.9%, p=0.03) and intraoperative assist techniques (intraoperative CT versus others, 2.1% versus 3.8%, p=0.04), while the pooled vascular injury rates were similar regardless of the screw types (2.2% versus 1.9%, p=0.25) and intraoperative assist techniques. Conclusion: The majority of the studies were retrospective with low-quality clinical evidence, which warrant future prospective, randomized studies. Despite this limitation, it can be concluded that autograft with C1-2Ss and intraoperative CT might be the most efficacious and safest option.