Change in spinal height following correction of adolescent idiopathic scoliosis

Corrective surgery for adolescent idiopathic scoliosis (AIS) leads to vertical growth arrest of the instrumented spine. This might be offset by the immediate gain in spinal height (SH) as a result of correction of the curvature.

This study aimed to identify predictors of gain in SH following corrective surgery for AIS. We present a unique model to predict postoperative height prior to intervention, which could contribute to the preoperative counseling and consenting process.

This was a retrospective case series. All surgeries were performed by one of four substantive pediatric spinal surgeons within a single regional center over a 3.5-year period.

There were 104 patients who had instrumented posterior spinal fusion for AIS included. There were 93 females, and the age range was from 11 to 17 years. All patients had posterior instrumented fusion using rods and anchors (pedicle screws±hooks).

Postoperative SH was the primary outcome measure. The SH (C7-L5) and Cobb angles were measured from a pre- and postoperative standing X-ray of each patient.

Variables associated with patients (demographic and radiological) and the surgical constructs were analyzed for predictability of height gain. A model was derived including only significant predictors of substantive importance using hierarchical regression methods. Cross-validation procedures verified the adequacy of the model fit. Analysis was performed using IBM SPSS Statistics for Windows version 20.0 (IBM Corp. Armonk, NY, USA).

The major curve was thoracic in 90% of cases. The number of vertebrae fused ranged from 5 to 15. The average preoperative Cobb angle was 66°, with an average correction of 45°. The average change in SH was 4.66 cm (SD 2.13 cm). The model presented included preoperative height, preoperative Cobb angle, and number of vertebrae within the construct, with coefficients of 1.00 (95% CI: 0.90, 1.09), 0.067 (95% CI: 0.039, 0.095), and 0.26 (95% CI: 0.11, 0.41), respectively. This model had an adjusted-R(2) value of 0.83 and a R(2) for prediction of 0.79, and can be shown to have similar predictive capability as a model comprising a wider range of predictors.

The greatest postoperative height values following posterior spinal fusion for AIS could be expected from a patient with greater preoperative height and Cobb angle, and whose construct spans a large number of vertebrae.