Clinically relevant finite element technique based protocol to evaluate growing rods for early onset scoliosis correction

Objective The emergence of distraction-based growing rods has provided the means to reduce the progression of spinal deformity in early onset scoliosis (EOS). The current protocols for evaluating spinal implants (ie, ASTM-F1717 and ISO-12189) were developed for fusion/dynamic devices. These protocols do not feature long unsupported rod lengths subjected to distraction. Due to the unsuitability of the existing guidelines for the evaluation of growing rods, a new distraction-based finite element protocol is presented herein for the first time. Method A vertebrectomy (VO) model from current protocols was modified to accommodate multi-spinal segments (ie, MS model) in which springs with appropriate stiffness were simulated in between the plastic blocks. To assess the efficacy of the protocol, two different computational studies were conducted: (a) compression-bending (MS-CB) with no distraction, and (b) distraction followed by compression-bending (MS-D + CB). In each study, the model with no axial connector (rods-only) was modified to include a) 80-mm long tandem (LT) connectors, and b) side-by-side (SBS) connectors. Stiffness and yield loads were calculated as per ASTM-F1717 guidelines and compared with the corresponding VO models with no distraction. In the MS-D + CB models, distraction was applied at the top block, stretching the spring-block construct in a simulation of scoliosis surgery prior to locking the construct at the connector-rods' interface. Results MS-CB models predicted higher stiffness and yield loads, compared to the VO models. The locking mechanism produced pre-existing stresses on the rod-connector interface, which caused a shift in the location of high-stress regions to the distraction site. Distraction led to a decrease in the construct's stiffness and yield load. Discussion The proposed protocol enables the simulation of clinical parameters that are not feasible in the F1717 models and predicted stress patterns in the hardware consistent with observed clinical failures.