Factors influencing radiographic and clinical outcomes in adult scoliosis surgery: a study of 448 European patients.

INTRODUCTION/
PURPOSE: In adult scoliosis surgery (AS) delineation of risk factors contributing to failure is important to improve patient care. Treatment goals include deformity correction resulting in a balanced spine and horizontal lowest instrumented vertebra (LIV) in fusions not ending at S1. Therefore, the study objectives were to determine predictors for deformity correction, complications, revision surgery, and outcomes as well as to determine predictors of postoperative evolution of the LIV-take-off angle (LIV-TO) and symptomatic adjacent segment disease (ASD).
METHODS: The authors performed a retrospective analysis of 448 patients who had AS surgery. Patients' age averaged 51 years, BMI 26, and follow-up of 40 months. According to the SRS adult scoliosis classification, 51 % of patients had major lumbar curves, 24 % each with single thoracic or double major curves. 54 % of patients had stable vertebra at L5 and 34 % of patients had fusion to S1. The mean number of posterior fusion levels was eight and implant density 73 %. Among standard radiographic measures of deformity the LIV-TO was assessed on neutral and bending/traction-films (bLIV-TO). Clinical outcomes were assessed in 145 patients with degenerative-type AS using validated measures (ODI, COMI and SF-36). Prediction analysis was conducted with stepwise multiple regression analyses.
RESULTS: Preoperative thoracic curve (TC) was 53° and 33° at follow-up. Preoperative lumbar curve (LC) was 43° and 24° at follow-up. Curve flexibility was low (TC 34 %/LC 38 %). TC-correction (38 %) was predicted by preoperative TC (r = 0.9) and TC-flexibility (r = 0.8). LC-correction (50 %) was predicted by preoperative LC (r = 0.8), LC-flexibility (r = 0.8) and screw density (r = 0.7). Preoperative LIV-TO was 18.2° and at follow-up 9.4° (p < 0.01). 20 % of patients had a non-union (18 % at L5-S1). The risk for non-union at L5-S1 increased with age (p = 0.04), low screw density (p = 0.03), and postoperative sagittal imbalance [(T9-tilt (p = 0.01), C7-SVA (p = 0.01), LL (p = 0.01) and PI-LL mismatch (p = 0.01)]. 32 % of patients had revision surgery. Risk for revision was increased in fusions to S1 (p < 0.01), increased BMI (p < 0.01), sagittal imbalance (C7-SVA, p < 0.01), age (p = 0.02), and disc wedging distal to the LIV (p < 0.01). To a varying extent, clinical outcomes negatively correlated (p < 0.05) with revision, ASD, perioperative complications, age, low postoperative TC- and LC-correction, and sagittal and coronal imbalance at follow-up (C7-SVA, PT, and C7-CSVL). 59 patients had ASD, which correlated with preoperative and postoperative sagittal and coronal parameters of deformity. In a multivariate model, preoperative bLIV-TO (p < 0.01) and preoperative LIV-TO (p < 0.01) demonstrated the highest predictive strength for follow-up LIV-TO.
CONCLUSION: In the current study, the magnitude of deformity correction in the sagittal and coronal planes was shown to have significant impact on radiographic and clinical outcomes as well as revision rates. Findings indicate that risks for complications might be reduced by restoration of sagittal balance, appropriate deformity correction and advanced lumbosacral fixation. The use of preoperative LIV-TO and LIV-TO on bending/traction-films were shown to be useful for surgical planning, selection of the LIV and prediction of follow-up-TO, respectively. Parameters of sagittal balance rather than coronal deformity predicted ASD.