Position of Interbody Spacer in Transforaminal Lumbar Interbody Fusion: Effect on 3-dimensional Stability and Sagittal Lumbar Contour

Abstract

Biomechanical study. To test 2 different intervertebral positions of a semilunar cage and their effects on 3-dimensional stability and segmental lordosis in a model of transforaminal lumbar interbody fusion (TLIF). In his original TLIF description, Harms recommended decortication of endplates, followed by placement of mesh cages in the middle-posterior intervertebral third. Subsequent studies presented conflicting recommendations: anterior placement of the spacer-cage for better load-sharing versus placement on the stronger posterolateral endplate regions. Six human lumbar spinal functional units were first tested intact. TLIF was performed using a semilunar poly-ether-ether-ketone cage randomly inserted in the anterior (TLIF-A) or posterior (TLIF-P) disc space. Pedicle screws and rods were added. Unconstrained pure moments in axial-torsion, lateral-bending (LB), and flexion-extension (FE) were applied under 0.05 Hz and +/-5 Nm sinusoidal waveform. Segmental motions were recorded. Range of motion (ROM) and neutral zone (NZ) were calculated. Pairwise comparisons were made using nonparametric Wilcoxon-matched pairs signed rank sum test with statistical significance set at P<0.05. TLIF-A and TLIF-P significantly decreased ROM (P<0.05) of the intact spinal functional unit, in FE and LB. In axial-torsion, decrease of ROM after TLIF procedures was not significant (P>0.05). Delta-ROM between TLIF-A and TLIF-P was not significant (P>0.05). TLIF-A and TLIF-P significantly decreased NZ in LB (P<0.05). In FE, TLIF-P significantly decreased NZ (P<0.05); TLIF-A showed a trend toward significance (P=0.09). Delta-NZ between TLIF-A and TLIF-P was not significant (P>0.05). Segmental lordosis of TLIF-A and TLIF-P on C-arm views showed angle differences within the range of measurement error of Cobb angles. Difference in ROM and NZ between anterior (TLIF-A) or posterior (TLIF-P) positions was not statistically significant. Similarly, both positions did not influence segmental lordosis.