OBJECT: Some centers report a lower incidence of vertebral artery (VA) injury with C-2 pars screws compared with pedicle screws without sacrificing construct stability, despite biomechanical studies suggesting greater load failures with C-2 pedicle screws. The authors reviewed published series describing C-2 pars and pedicle screw implantation and atlantoaxial fusions and compared the incidence of VA injury, screw malposition, and successful atlantoaxial fusion with each screw type. METHODS: Online databases were searched for English-language articles between 1994 and April of 2011 describing the clinical and radiographic outcomes following posterior atlantoaxial fusion with C-1 lateral mass and either C-2 pars interarticularis or pedicle screws. Thirty-three studies describing 2975 C-2 pedicle screws and 11 studies describing 405 C-2 pars screws met inclusion criteria for the safety analysis. Seven studies describing 113 patients treated with C-2 pars screws and 20 studies describing 918 patients treated with C-2 pedicle screws met inclusion criteria for fusion analysis. Standard and formal meta-analysis techniques were used to compare outcomes. RESULTS: All studies provided Class III evidence. Ten instances of VA injury occurred with C-2 pedicle screws (0.3%) and no VA injury occurred with pars screws. The point estimate of VA injury for C-2 pedicle screws was 1.09% (95% CI 0.73%-1.63%) and was similar to that of C-2 pars screws (1.48%, 95% CI 0.62%-3.52%). Similarly, there was no statistically significant difference in the rate of clinically significant screw malpositions (1.14% [95% CI 0.77%-1.69%) vs 1.69% [95% CI 0.73%-3.84%]). Radiographically identified screw malposition occurred in a higher proportion of C-2 pedicle screws compared with C-2 pars screws (6.0% [95% CI 3.7%-9.6%] vs 4.0% [95% CI 2.0%-7.6%], p < 0.0001). Pseudarthrosis occurred in a greater proportion of patients treated with C-2 pars screws (5 [4.4%] of 113) compared with those treated with C-2 pedicle screws (2 [0.22%] of 900). Point estimates with 95% confidence intervals show a slightly higher rate of successful atlantoaxial fusion in the pedicle screw cohort (97.8% [CI 96.0%-98.8%] vs 93.5% [CI 86.6%-97.0%]; p < 0.0001). Q-testing ruled out heterogeneity between the study groups. CONCLUSIONS: With a thorough knowledge of axis anatomy, surgeons can place both C-2 pars and C-2 pedicle screws accurately with a small risk of VA injury or clinically significant malposition. There may be subtle trade-off of safety for rigidity when using axial pedicle instead of pars screws, and the decision to use either screw type must be made only after careful review of the preoperative CT imaging and must take into account the surgeon's expertise and the particular demands of the clinical scenario in any given case.
OBJECT: Some centers report a lower incidence of vertebral artery (VA) injury with C-2 pars screws compared with pedicle screws without sacrificing construct stability, despite biomechanical studies suggesting greater load failures with C-2 pedicle screws. The authors reviewed published series describing C-2 pars and pedicle screw implantation and atlantoaxial fusions and compared the incidence of VA injury, screw malposition, and successful atlantoaxial fusion with each screw type. METHODS: Online databases were searched for English-language articles between 1994 and April of 2011 describing the clinical and radiographic outcomes following posterior atlantoaxial fusion with C-1 lateral mass and either C-2 pars interarticularis or pedicle screws. Thirty-three studies describing 2975 C-2 pedicle screws and 11 studies describing 405 C-2 pars screws met inclusion criteria for the safety analysis. Seven studies describing 113 patients treated with C-2 pars screws and 20 studies describing 918 patients treated with C-2 pedicle screws met inclusion criteria for fusion analysis. Standard and formal meta-analysis techniques were used to compare outcomes. RESULTS: All studies provided Class III evidence. Ten instances of VA injury occurred with C-2 pedicle screws (0.3%) and no VA injury occurred with pars screws. The point estimate of VA injury for C-2 pedicle screws was 1.09% (95% CI 0.73%-1.63%) and was similar to that of C-2 pars screws (1.48%, 95% CI 0.62%-3.52%). Similarly, there was no statistically significant difference in the rate of clinically significant screw malpositions (1.14% [95% CI 0.77%-1.69%) vs 1.69% [95% CI 0.73%-3.84%]). Radiographically identified screw malposition occurred in a higher proportion of C-2 pedicle screws compared with C-2 pars screws (6.0% [95% CI 3.7%-9.6%] vs 4.0% [95% CI 2.0%-7.6%], p < 0.0001). Pseudarthrosis occurred in a greater proportion of patients treated with C-2 pars screws (5 [4.4%] of 113) compared with those treated with C-2 pedicle screws (2 [0.22%] of 900). Point estimates with 95% confidence intervals show a slightly higher rate of successful atlantoaxial fusion in the pedicle screw cohort (97.8% [CI 96.0%-98.8%] vs 93.5% [CI 86.6%-97.0%]; p < 0.0001). Q-testing ruled out heterogeneity between the study groups. CONCLUSIONS: With a thorough knowledge of axis anatomy, surgeons can place both C-2 pars and C-2 pedicle screws accurately with a small risk of VA injury or clinically significant malposition. There may be subtle trade-off of safety for rigidity when using axial pedicle instead of pars screws, and the decision to use either screw type must be made only after careful review of the preoperative CT imaging and must take into account the surgeon's expertise and the particular demands of the clinical scenario in any given case.