An in vitro flexibility experiment was performed to compare the biomechanical stability of asymmetrical lumbar pedicle screw fixation (longer hardware attached ipsilaterally to a 1-sided lesion), short and long fixation, and fixation with and without interconnection to the involved vertebra.

Seven human cadaveric specimens (T12–S1) were studied intact; after simulated unilateral lesions were created at L2–3 and L3-4, the segments were stabilized by 1) L2–4 unilateral fixation (L-3 excluded), 2) L2–4 bilateral fixation (L-3 included contralaterally), 3) L2–5 unilateral fixation (L-3 excluded), 4) L2–5 fixation ipsilateral (L-3 excluded) and L2–4 fixation contralateral (L-3 included), 5) L2–5 bilateral fixation (L-3 included contralaterally), and 6) L2–5 bilateral fixation (L-3 excluded). The testing order varied among specimens. Angular range of motion (ROM) and lax zone were recorded optically while loading to 6.0 Nm was created with nonconstraining pure moments.

Unilateral short fixation provided significantly worse stabilization than any other construct tested in all loading modes (p < 0.05, repeated-measures analysis of variance). There was a mean 56% reduction in ROM across the lesion after adding 1 additional level rostrally and caudally. Asymmetrical long/short stabilization provided similar stability to symmetrical long stabilization. Minimal additional stability was gained by including L-3 in the long bilateral fixation construct.

Unilateral fixation is inadequate for stabilizing a 2-level unilateral lesion. Bilateral fixation, whether symmetrical or asymmetrical, provides good stabilization for this injury. It is not important for stability to include the level of the lesion within the long construct contralaterally.