Fourteen human specimens (L1-S1) were divided into cortical and pedicle screw treatment groups. Motion was captured for the following conditions: intact, bilateral posterior fixation (L3-L5), fixation with laminectomy at L3-L5, fixation with laminectomy and transforaminal lumbar interbody fusion at L3-L5 both prior to, and following, simulated in vivo fatigue. Following fatigue, screw pullout force was collected and \"effective shear stress\" [pullout force/screw surface area] (N/mm2) was calculated; comparisons and correlations were performed.
In flexion-extension and lateral bending, all operative constructs significantly reduced motion compared to intact (P < 0.05), regardless of pedicle or cortical screws; only posterior fixation with and without laminectomy significantly reduced motion in axial rotation (P < 0.05). Pedicle screws significantly increased average pullout strength (944.2 N vs. 690.2 N, P < 0.05), but not the \"effective shear stress\" (1.01 N/mm2 vs. 1.1 N/mm2, P > 0.05).
In a posterior laminectomy and fusion model, cortical screws provided equivalent stability to pedicle screw fixation, yet had significantly lower screw pullout force. No differences in \"effective shear stress\" warrant further investigation of the effect of screw length/diameter in the aforementioned screw trajectories.