The Effect of Proximal Anchor Choice during Distraction-based Surgeries for Patients with Nonidiopathic Early-onset Scoliosis: A Retrospective Multicenter Study

Background: It is unclear whether the type of proximal anchor affects the spine length achieved with distraction-based surgeries in patients with nonidiopathic early-onset scoliosis (EOS). Since distraction may produce kyphosis, spine length should be assessed in the sagittal plane using the sagittal spine length (SSL - curved arc length of the spine in the sagittal plane). Our purpose was to determine if the type of proximal anchor in distraction-based surgeries will affect final spine length. Methods: Patients with nonidiopathic EOS treated with distraction-based systems (minimum 5 y follow-up, 5 lengthenings) were identified from 2 EOS registries. Radiographic analysis preoperative, postimplant (L1), and after each lengthening (L2-L5, L6-L10, L11-L15) was performed with the primary outcome of T1-S1 SSL. Results: We identified 126 patients - 70 had rib-based implants (52 congenital, 9 syndromic, 9 neuromuscular) and 56 had spine-based implants (15 congenital, 29 syndromic, 12 neuromuscular) with preoperative age 4.6 years, scoliosis 75 degrees, and kyphosis 48 degrees. After initial correction (P<0.05), scoliosis remained constant [58 degrees (13 to 104 degrees) at L11-L15] and kyphosis increased over time [38 degrees (9 to 108 degrees) at L1 to 60 degrees (17 to 134 degrees) at L11-L15] (P<0.05). Preoperative SSL was higher in the spine-based group (29.6 cm) when compared with the rib-based group (25.2 cm) (P<0.05). This difference was maintained after initial implantation (spine-based: 32.2 cm vs. rib-based: 26.7 cm, P<0.05) and at final follow-up (spine-based: 37.0 cm vs. rib-based: 34.4 cm, P<0.05). As preoperative SSL differed between groups, maximum SSL gains per interval were also normalized to preoperative SSL. There was no statistically significant difference between groups at L1, L2-L5, and L6-L10. However, at L11-L15, the rib-based group achieved a more relative increase in spine length compared with the spine-based group (45% vs. 31%, P<0.05). Conclusion: At minimum 5 year follow-up, distraction-based surgeries increased spine length for patients with nonidiopathic EOS; regardless of proximal anchor type.