Comparison of pedicle screw loosening under uniaxial and multiaxial loading.

Abstract

Pedicle screw loosening is a clinically-relevant failure mode following spinal fusion procedures, resulting from damage and compaction to the surrounding bone with loading of the bone-implant interface. Current in vitro tests for investigating pedicle screw loosening commonly apply uniaxial loading (e.g. screw pullout or toggle testing). However, these methods do not replicate multiaxial spinal loading associated with activities of daily living. The goals of this study were to (1) develop and assess a novel test method for inducing pedicle screw loosening under multiaxial loading conditions, and (2) compare this approach to uniaxial toggle testing. Eight osteoporotic human lumbar vertebrae (L2-L4, 70.1 ± 5.5 years old) underwent testing to compare volumetric bone damage and screw displacement under uniaxial and multiaxial loading. Load magnitudes were incrementally increased in a step-wise fashion after 1000 and 2000 cycles, with the final load step continuing until the screw head displaced more than 6 mm. Post-test computed tomography (CT) imaging was used to quantify volumetric bone damage. Multiaxial loading accelerated bone damage leading to earlier implant loosening (p < 0.001). Failure predominantly occurred in the inferior direction (p = 0.06). Medial (p = 0.07) and lateral (p < 0.001) screw displacement were greater under multiaxial loading. CT imaging showed greater volumetric bone deformation with multiaxial loading (p = 0.01). This study demonstrated that physiologically-derived multiaxial loading accelerates volumetric bone damage and pedicle screw loosening. Further, the increased bone deformation in the medial and lateral directions suggests that spinal movements involving these loads may accelerate risk of screw loosening.