Finite element model of spinal hemiepiphysiodesis: effect of contact conditions, initial conditions, and growth

Growth modification is under investigation to treat pediatric spine deformities. A hemiepiphyseal staple construct has been shown to alter growth and create physeal structural gradients in an in vivo porcine model. A finite element model (FEM) of a motion segment with and without implant was developed based on preliminary experimental results of initial post-operative motion segment compressive properties. The nonlinear tangent stiffness determined from the model correlated well with the experiments for the native segment, whereas after addition of the implant, the model overestimated the stiffness. The current purpose was to determine the effect of implant-bone contact conditions and initial disc displacement conditions, include growth effects, and compare FEM and experimental results at each stage. A 3D FEM was developed from a CT scan of a porcine T7-T8 segment. The annulus was modeled as an incompressible anisotropic hyperelastic material, and the nucleus as an incompressible fluid. A CAD model of the implant was constructed. Load-displacement curves in compression were determined from a nonlinear analysis performed under different initial and bone-implant interface conditions. Contact conditions were a) perfect, b) friction of 0.1-0.3, or c) soft contact. Initial conditions were that implant insertion induced a) no change in stress or strain in the disc, b) a 2˚ angulation with a centrally located neutral axis and no residual stresses, and c) both stress and coronal plane displacement gradients. Growth modulation effects were added using a published linear relationship between compressive stress and growth rate. A 2 month PO time was simulated. Altering bone - implant surface contact conditions from perfect to either friction or soft contact decreased the stiffness, but all models remained stiffer than experimental results. An initial disc angulation without residual stress did not affect stiffness, whereas stiffness increased with an initial angle and compressive stress. The growth simulation results predicted 32% of control growth on the side ipsilateral to the implant and 81% on the contralateral side. Reductions in growth were similar in pattern, but overestimated, experimental histomorphometric changes.