What happens to the unfused upper thoracic curve after posterior spinal fusion for adolescent idiopathic scoliosis?

OBJECTIVE Spontaneous lumbar curve correction after selective thoracic fusion in surgery for adolescent idiopathic scoliosis (AIS) is well described. However, only a few articles have described the course of the uninstrumented upper thoracic (UT) curve after fusion, and the majority involve a hybrid construct. In this study, the authors sought to determine the outcomes and associated factors of uninstrumented UT curves in patients with AIS. METHODS The authors retrospectively reviewed a prospectively collected multicenter AIS registry for all consecutive patients with Lenke type 1-4 curves with a 2-year minimum follow-up. UT curves were considered uninstrumented if the upper instrumented vertebra (UIV) did not extend above 1 level from the lower end vertebra of the UT curve. The authors defined progression as > 5°, and divided patients into two cohorts: Those with improvement in the UT curve (IMP) and those without improvement in the UT curve (NO IMP). Radiographic, demographic, and Scoliosis Research Society (SRS)-22 survey outcome measures were compared using univariate analysis, and significant factors were compared using a multivariate regression model. RESULTS The study included 450 patients (370 females and 80 males). The UT curve self-corrected in 86% of patients (n = 385), there was no change in 14% (n = 65), and no patients worsened. Preoperatively, patients were similar with respect to Lenke classification (p = 0.44), age (p = 0.31), sex (p = 0.85), and Risser score (p = 0.14). The UT curves in the IMP group self-corrected from 24.7° ± 6.5° to 12.6° ± 5.9°, whereas in the NO IMP group UT curves remained the same, from 20.3° ± 5.8° to 18.5° ± 5.7°. In a multivariate analysis, preoperative main thoracic (MT) curve size (p = 0.004) and MT curve correction (p = 0.001) remained significant predictors of UT curve improvement. Greater correction of the MT curve and larger initial MT curve size were associated with greater likelihood of UT curve improvement. CONCLUSIONS Spontaneous UT curve correction occurred in the majority (86%) of unfused UT curves after MT curve correction in Lenke 1-4 curve types. The magnitude of preoperative MT curve size and postoperative MT curve correction were independent predictors of spontaneous UT curve correction.