Mechanical stiffness of segmental versus nonsegmental pedicle screw constructs: the effect of cross-links

A biomechanical study in porcine spines of the construct stiffness effects of segmental pedicle screws. Stiffness effects of supplementation of nonsegmental screw constructs with cross-links was also evaluated.

To assess the biomechanical differences between constructs using segmental versus nonsegmental pedicle screw-based instrumentation as well as the effect of cross-links.

An in vitro biomechanical comparison of segmental versus nonsegmental pedicle screw constructs with and without cross-links using porcine lumbar vertebrae was performed. Mechanical trade-offs of reducing the number of pedicle screws in a given construct and substituting a cross-link for a pair of screws are not well understood.

Three, 4, and 5-vertebral segments from 18 porcine spines were instrumented with segmental and nonsegmental pedicle screw constructs, and with nonsegmental screws augmented with cross-links. Unconstrained biomechanical testing in flexion, extension, and axial rotation with 6 degree-of-freedom motion tracking was performed. Statistical comparisons of stiffness data were conducted using 2-tailed paired t tests.

There was a statistically significant increase in stiffness between models with segmental pedicle screws compared to nonsegmental pedicle screws in 6 of the 9 mechanical tests. The remaining 3 tests approached but did not reach statistical significance (P = 0.087, 0.062, and 0.078). When cross-links were added to the nonsegmental models, differences in stiffness compared to segmental pedicle screws were largely eliminated, decreasing well below statistical significance in 8 of 9 tests. The highest difference in nonsegmental models with cross-links and segmental pedicle screw models was observed for the 5-vertebrae fusion models, for which axial rotation testing maintained statistically significant differences (P = 0.006), and flexion testing approached significance (P = 0.062).

Segmental pedicle screw constructs increased mechanical stiffness compared to nonsegmental constructs in our fusion models. Placement of a single cross-link with nonsegmental screws eliminated statistical differences for 3 and 4-vertebral level constructs, and may be a satisfactory alternative in this clinical setting. Caution in applying these results inlonger constructs is recommended, given persistent increased stiffness found for the segmental 5-vertebral level models.