Introduction: Currently, there is significant equipoise regarding the selection and placement of growing spinal instrumentation when treating patients with early-onset scoliosis (EOS). The primary purpose of this study was to compare complications following surgery in patients receiving rib-based versus spine-based proximal anchors as a part of posterior growing instrumentation in the management of EOS.
Methods: Retrospective cohort study. Inclusion criteria required: age 3-10 years old, diagnosis of EOS, treatment with a growing construct that utilized rib- or spine-based proximal anchors, and a major coronal curve larger than 40 degrees. The primary outcome analyzed was postoperative complications. Secondary outcomes included coronal major curve correction and patient reported outcomes measured by the Early-Onset Scoliosis 24-item Questionnaire (EOSQ-24). Subjects were categorized into rib- or spine-based proximal fixation groups for comparison.
Results: Of 104 patients included in the study, 76 (73.1%) were treated with rib-based constructs and 28 (26.9%) were treated with spine-based constructs. 24 (31.6%) patients with rib-based constructs and 9 (32.1%) patients with spine-based constructs experienced at least one implant related complication (p = 0.956). Rod fracture was observed more often in spine-based groups than rib-based groups for both patients with congenital/idiopathic EOS (rib: 0 (0%) vs. spine: 3 (13.6%), p = 0.009) and neuromuscular/syndromic EOS (rib: 0 (0%) vs. spine: 2 (33.3%), p = 0.002). Furthermore, surgical site infection was found to be more frequent in rib-based than spine-based groups for neuromuscular/syndromic patients (rib: (13) 27.15 vs. spine: (1) 4.5%, p = 0.029). The most commonly reported complication was device migration. In patients with rib-based constructs, 2 (12.5%) patients with ≥ 5 anchors and 13 (21.7%) patients with < 5 anchors experienced device migration (p = 0.413). In patients with spine-based constructs, 1 (11.1%) patient with ≥ 5 anchors and 4 (21.1%) patients with < 5 anchors experienced device migration (p = 0.064). Spine-based anchors had significantly higher% correction (42.0%) compared to rib-based anchors (20.6%) (p = 0.003) at the most recent follow-up. There were no significant differences in the change of patient reported outcomes as measured by the EOSQ-24 between patients who received rib or spine-based anchors.
Discussion: The number of patients with at least one implant related complication was similar between the rib- and spine-based groups. Having 5 or more proximal anchors appeared protective against proximal device migration; however, this result was not statistically significant. Spine-based anchors had better overall correction than rib-based anchors. There were no differences in the change in patient reported outcomes between spine- and rib-based cohorts.