Fusion rates for odontoid fractures after treatment by anterior odontoid screw versus posterior C1-C2 arthrodesis: a meta-analysis

Objective For odontoid fractures, surgical treatment approaches including anterior odontoid screw fixation approach and the posterior C1-C2 arthrodesis approach are generally adopted in practice. However, the choice of different surgical procedures remains controversial. In addition to surgical technique, the fusion rate is an important factor contributing to the clinical efficacy. Therefore, this study was aimed to investigate the discrepancy in fusion rate between these two surgical approaches through synthesizing the currently available evidence on the topic. Methods A computerized search of Ovid, Medline, Embase, and the Cochrane library up to December 2017 for literature on the complication rate during odontoid fracture treatment was conducted. Risk ratio (RR) with its 95% confidence interval (CI) was pooled to assess fusion rates after surgical treatments, including anterior odontoid screw fixation approach or posterior C1-2 arthrodesis procedure, for patients with odontoid fractures. Results Thirteen studies were enrolled in the meta-analysis. Results show that no significant difference was found in the overall fusion rate (RR = 0.96, 95% CI 0.90-1.01). There was no significant heterogeneity among the studies (p value = 0.60). As to age- and economic-level subgroups, there was no statistical evidence to suggest an association of the patient age and economy development level with the choice of surgical approach. However, it is shown that better fusion rates of patients (>= 60 years) in developed countries received a better fusion rates after posterior fixation compared with anterior group using the fixed-effect model (RR = 0.88, 95% CI 0.79-0.98). Conclusion Elderly patients (>= 60 years) underwent posterior C1-2 arthrodesis fixation shows higher fusion rates in developed countries comparing with patients who underwent anterior odontoid screw fixation. Overall, there is no significant discrepancy between these two surgical approaches. However, the conclusion should be verified by further study enrolling larger sample size.