LAUSR

The purpose of this study is to evaluate if AR offers improved 3D kyphosis restoration during PSF for hypokyphosis in moderate AIS (< 70° coronal cobb), where the decision for AR is likely driven by sagittal concerns. A multicenter pediatric spine registry was queried for hypokyphotic (< 10°) Lenke 1–4 AIS patients aged < 20 years with > 2-year surgical follow-up. Coronal curves were limited to < 70°. A linear mixed model was created to predict 2-year 3D kyphosis by treatment and pre-op 3D kyphosis, while controlling for age, sex, thoracic coronal deformity and flexibility, osteotomy use, implant characteristics, surgery recency, and surgeon. 1384 patients were included with 53 (3.8%) undergoing PSF + AR. Mean preop 3D kyphosis was similar between PSF and PSF + AR groups (− 3.7° vs. − 0.5°; p = 0.08). PSF-AR had similar 2-year 3D kyphosis (23.0° [95% CI 20.5–25.4°] vs. 23.3° [22.9–23.6°]) and correction (26.7° [23.3–29.9°] vs. 23.7° [23.3–24.2°]) compared to PSF. When controlling for covariates, the models demonstrated no difference between approach (p = 0.058) or interaction of approach and preop 3D kyphosis (p = 0.31). Post-hoc power analysis showed an adequate sample size to detect a difference of 5° between approaches. PSF + AR had longer surgical times (324 vs. 266 min, p < 0.001) though no significant increase in overall complications (17% vs. 12.4%) compared to PSF alone. In AIS patients with coronal curve < 70° and 3D hypokyphosis of 10 to − 40°, treatment with PSF + AR did not improve 2-year sagittal correction more than PSF alone. Surgeon identity and surgery recency influenced post-operative kyphosis more than any other patient or surgical factor.