Five or more proximal anchors and including upper end vertebra protects against reoperation in distraction-based growing rods

Study design: Retrospective multi-center enrollment. Objective: To examine the impact of patient and surgical factors on proximal complication and revision rates of early onset scoliosis patients using a multicenter database. Summary of background data: Proximal anchor pullout and junctional kyphosis are common causes necessitating revision surgery during growth friendly treatment of early onset scoliosis (EOS). Many options exist for proximal fixation and may impact the rate of these complications. Methods: Retrospective review of multicenter database of patients with growth friendly constructs for EOS. Inclusion criteria were patients with index instrumentation < 10 years of age and minimum of 2 year follow-up. Results: 353 patients met the inclusion criteria and had the following constructs: growing rods with spine anchors = 303; growing rods with rib anchors = 15 and VEPTR = 35. Mean age at index instrumentation was 6.0 years. Mean preoperative Cobb angle was 76° and mean kyphosis was 54°. Mean follow-up was 6.0 years. 21.8% of patients (77/353) experienced anchor pullout. Lower anchor pullout rates were associated with a higher numbers of proximal anchors (p = 0.003, r = − 0.157), and 5 or more anchors were associated with lower rates of anchor pullout (p = 0.014). Anchor type (rib hooks vs spine anchors vs rib cradle) did not impact rate of anchor pullout (p = 0.853). Kyphosis data was available for 198 patients. 23.2% (46/198) of these patients required proximal extension of their construct after index surgery. Initial instrumentation below the upper end vertebrae (UEV) of kyphosis was associated with higher rates of subsequent proximal revision; 28.9% (20/69) compared to 20.1% (26/129) for those instrumented at or above the UEV (p = 0.035). Preoperative kyphosis and change in thoracic kyphosis were not associated with anchor pullout (p = 0.436, p = 0.115) or proximal revision rates (p = 0.486, p = 0.401). Conclusion: Five or more anchors are associated with lower rates of anchor pullout. Proximal anchor placement at or above the UEV resulted in a significant decrease in rates of proximal extension of the construct.